HPC Systems Inc. operates in two segments: High Performance Computing (HPC; 67.8% of total revenue and 73.0% of total operating profit in FY06/21) and Configure to Order (CTO; 32.2%, 27.0%). In FY06/21, it had revenue of JPY5.8bn (+23.2% YoY) and operating profit of JPY677mn (+41.6% YoY).
In the HPC segment, the company provides one-stop services for high-performance computing (HPC). These include sales of HPC systems for scientific computing and simulation software; systems integration services (i.e., building scientific computing, artificial intelligence [AI], and deep-learning environments); services to parallelize and speed up simulation software programs; development and sales of computational chemistry and materials informatics software; contracted computing services and support for scientific R&D; and science cloud services (Software as a Service for drug discovery, materials R&D, etc.).
In the CTO segment, HPC Systems supplies customized industrial and edge computers that meet customer applications and challenges as determined through interviews. It proposes specifications for such computers, and develops, manufactures, and maintains them, ensuring long-term stable supply of products. Industrial and edge computers are used as controlling equipment in plant production facilities, manufacturing equipment, inspection equipment, control devices, transportation infrastructure, automated driving applications, and retail stores. The company customizes these products in accordance with customer requirements such as pricing, performance, quality levels, operation in harsh environments, and resistance to low or high temperatures, dust, water, and static electricity.
In the HPC segment, the company procures high-performance computer hardware from major electronics manufacturers, and configures these to match the operating requirements of its customers. It builds small to large computer clusters that connect workstations, servers, and storage devices through high-speed networking equipment, installs the necessary software, and optimizes these systems as supercomputers. The hardware and software that make up these systems differ from consumer PCs in that getting them to operate requires more than simply setting up the necessary connections and installing the appropriate software. The hardware, operating system (OS), middleware, software, and network must be configured and optimized to run multiple computational processes in parallel while reducing processing time, and this requires the expertise of a company like HPC Systems.
Scientific computing is used in a range of fields such as elementary particle and astronomical research, weather forecasting, global warming projections, predictions of disasters such as earthquakes and tsunamis, drug discovery, development of semiconductors and battery materials, and structural and aerodynamic analysis of vehicles and aircraft. HPC Systems excels in building systems for the life science and material science fields, and it mainly supplies these to universities, public research institutions, and R&D divisions of private companies. The process to develop new drugs or materials involves synthesizing various compounds as potential candidates for new drugs or materials, and verifying whether these deliver the desired effect. Supercomputers can simulate combinations of countless different materials, allowing researchers to predict the attributes of new materials before they produce them. This contributes to more effective and efficient development and verification.
Other companies that operate in the systems integration field are HPC Solutions Co., Ltd. (unlisted), HPC Tech Corporation (unlisted), and Visual Technology, Inc. (unlisted). However, HPC Systems does not actually compete with these companies for large projects or in the computational chemistry field. The target market for the HPC segment is worth JPY260.0bn (based on a 2021 company survey), and the segment recorded revenue of JPY4.0bn (+23.6% YoY) in FY06/21.
The HPC segment generates a mix of one-time revenue and recurring revenue, with the former accounting for over 90% of the total revenue. One-time revenue is determined by order value and the number of orders, and the lead time for orders is roughly three months. Order value is several to 10 million yen for small orders, goes up to JPY50mn for mid-sized orders, and reaches or exceeds JPY400mn for large orders. Recurring revenue is monthly revenue generated from the company’s science cloud services, and varies based on the type of services requested by customers.
The target market for the CTO segment is worth JPY178.0bn (based on a 2021 company survey), and the segment recorded revenue of JPY1.9bn (+22.8% YoY) in FY06/21. Using a high-mix/low-volume manufacturing approach, HPC Systems prepares custom designs for industrial computers in accordance with customer requirements, and procures and assembles the optimal parts to build these products. In contrast, its rivals mass-produce and sell general-purpose industrial computers. HPC Systems has also established an organization that enables it to supply its products to customers over the long term. The company’s peers include Contec Ltd. (TSE2: 6639) and Innotech Corporation (TSE1: 9880), but it do not actually compete with these companies, which mass-produce general-purpose industrial computers. HPC Systems has lower revenue than its peers, but it boasts higher profit margins.
In the CTO segment, HPC Systems mainly supplies computers for inspection equipment (e.g., inspection and monitoring of products in plants), Internet of Things (IoT) computers used in plants (e.g., network-based management of cameras and monitoring in plants), and high-end industrial computers for medical image processing (e.g., computed tomography [CT] scans and magnetic resonance imaging [MRI]). Customers often place repeat orders once they have integrated a computer supplied by the company into one of their products, and this trend is reflected in a repeat purchase rate of over 90%.
In both HPC and CTO segments, HPC Systems provides products and services tailored to customer needs (i.e., it procures hardware/software, assembles these, and ships the final products). Accordingly, variable (materials) costs accounted for 97.8% of cost of revenue in FY06/21 (cost of revenue ratio: 69.0%). Its SG&A ratio stood at 19.4%, mostly comprising fixed costs such as personnel expenses and depreciation.
Earnings trends
In FY06/21, revenue was JPY5.8bn (+23.3% YoY), operating profit JPY677mn (+41.6% YoY), recurring profit JPY666mn (+43.1% YoY), and net income JPY447mn (+45.4% YoY). The results reflected contributions from large orders received from universities and public research institutions. The SG&A expense ratio improved 3.3pp YoY as revenue growth outpaced the increase in SG&A expenses, and OPM improved 1.5pp YoY to 11.6%.
In FY06/22, the company forecasts revenue of JPY6.4bn (+10.2% YoY), operating profit of JPY745mn (+10.2% YoY), recurring profit of JPY741mn (+11.3% YoY), and net income of JPY496mn (+11.0% YoY). HPC Systems thinks the business environment is hard to read at present because it remains unclear when the COVID-19 pandemic will subside and give way to an economic recovery, and also due to impact from semiconductor shortages.
In August 2021, the company unveiled its medium-term management plan “Vision 2024” (FY06/22–FY06/24). As numerical targets, the plan calls for revenue of JPY10.0bn (three-year CAGR of 19.7%), operating profit of JPY1.3bn (23.8%), and ROE of 26% (28.1% in FY06/21). By segment, it looks for revenue of JPY7.0bn (three-year CAGR of 21.0%) and operating profit of
Strengths and weaknesses
Shared Research regards HPC System’s strengths as: (1) its expertise in supercomputers and scientific computing software for quantum chemical computing, molecular dynamics, and other fields in the HPC segment , (2) its ability to rapidly deliver cutting-edge products at affordable prices and secure orders mainly through customer requests in the HPC segment, and (3) its capacity to sell and produce high-performance industrial computers with custom designs tailored to the needs of its customers using a high-mix/low-volume manufacturing approach in the CTO segment. We see its weaknesses as (1) a susceptibility to direct impact of disruptions at suppliers of semiconductors and peripheral electronic parts, (2) a small operational scale that hinders the company’s ability to accept large projects in the HPC segment, and (3) difficulties in generating economies of scale in the CTO segment due to a high-mix/low-volume manufacturing structure. For details, see the Strengths and weaknesses section.
Key financial data
Income statement
FY03/14
FY06/14
FY06/15
FY06/16
FY06/17
FY06/18
FY06/19
FY06/20
FY06/21
FY06/22
(JPYmn)
Non-cons.
Non-cons.
Non-cons.
Non-cons.
Non-cons.
Non-cons.
Non-cons.
Non-cons.
Non-cons.
Est.
Revenue
2,666
370
2,587
2,896
3,901
4,053
5,396
4,725
5,828
6,425
YoY
-
-
11.9%
34.7%
3.9%
33.1%
-12.4%
23.3%
10.2%
Gross profit
1,030
1,190
1,398
1,551
1,806
YoY
15.5%
17.5%
10.9%
16.5%
Gross profit margin
26.4%
29.4%
25.9%
32.8%
31.0%
Operating profit
244
282
370
478
677
745
YoY
15.6%
30.9%
29.3%
41.6%
10.1%
Operating profit margin
6.3%
7.0%
6.8%
10.1%
11.6%
11.6%
Recurring profit
130
-33
114
75
254
292
367
465
666
741
YoY
-
-
-34.1%
238.3%
14.8%
25.8%
26.8%
43.1%
11.3%
Recurring profit margin
4.9%
-
4.4%
2.6%
6.5%
7.2%
6.8%
9.8%
11.4%
11.5%
Net income
68
-30
42
39
163
190
219
307
447
496
YoY
-
-
-8.0%
317.9%
16.5%
15.6%
40.1%
45.4%
10.9%
Net margin
2.6%
-
1.6%
1.3%
4.2%
4.7%
4.1%
6.5%
7.7%
7.7%
Per-share data (split- and reverse split-adjusted; JPY)
Shares issued (year-end; '000)
8
8
8
8
8
8
8
4,119
4,234
EPS
16.86
-7.46
10.50
9.65
40.34
46.99
54.33
75.37
107.87
119.52
EPS (fully diluted)
-
-
-
-
-
-
-
71.06
104.06
Dividend per share
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Book value per share
105.48
98.02
108.81
118.77
159.11
206.10
260.43
353.00
416.90
Balance sheet (JPYmn)
Cash and cash equivalents
794
576
938
1,399
1,731
Total current assets
1,585
1,877
2,054
2,424
3,007
Tangible fixed assets
93
98
92
101
123
Investments and other assets
25
22
83
92
125
Intangible assets
34
25
49
37
27
Total assets
1,737
2,022
2,277
2,654
3,282
Accounts payable
3
2
1
3
4
Short-term debt
546
603
463
624
589
Total current liabilities
940
1,073
1,095
1,145
1,286
Long-term debt
154
117
128
54
263
Total fixed liabilities
154
117
128
54
263
Total liabilities
1,094
1,189
1,224
1,199
1,549
Total net assets
643
833
1,053
1,455
1,732
Total interest-bearing debt
700
719
592
678
853
Cash flow statement (JPYmn)
Cash flows from operating activities
200
-191
537
342
426
Cash flows from investing activities
-42
-47
-48
-61
-100
Cash flows from financing activities
254
19
-126
181
4
Financial ratios
ROA (RP-based)
9.3%
6.1%
17.4%
15.5%
17.1%
18.9%
22.4%
ROE
10.1%
8.5%
29.0%
25.7%
23.3%
24.5%
28.1%
Equity ratio
33.5%
34.5%
40.5%
37.0%
41.2%
46.2%
54.8%
52.8%
Source: Shared Research based on company data
Note: Figures may differ from company data due to differences in rounding methods.
Recent updates
Capital and business alliance with QunaSys Inc.
2022-03-29
HPC Systems Inc. announced an agreement to form a capital and business alliance
with QunaSys Inc. (unlisted), with a view to accelerating business development in computational quantum chemistry.
In July 2020, HPC Systems entered into a business alliance with QunaSys, to pursue development in the computational quantum chemistry field. The aim of this business alliance was to fuse HPC Systems' advanced technologies and solutions in computational chemistry with the cutting-edge technologies of QunaSys, a developer of quantum computing applications, to develop the world’s first computational chemistry solution for quantum computers and enhance both companies' enterprise value in the process.
The two companies have been engaged in R&D and business development as part of this alliance. As these activities have generated synergies in both technology and business, HPC Systems has agreed to enter into a capital and business alliance with QunaSys by way of acting as an underwriter for the latter's Series B third-party allotment of new shares (investment amount and stake undisclosed).
The parties believe that this capital and business alliance will further strengthen their collaborative relationship and accelerate business growth toward the practical application of quantum computing, expected within a few years.
Source: Shared Research based on company data
Note: Figures may differ from company data due to differences in rounding methods.
Cumulative Q3 FY06/22 results (out May 13, 2022)
Revenue: JPY5.0mn (-1.2% YoY)
Operating profit: JPY638mn (-9.6% YoY)
Recurring profit: JPY621mn (-10.7% YoY)
Net income: JPY432mn (-9.8% YoY)
The company adopted the new Accounting Standard for Revenue Recognition starting Q1 FY06/22. As a result of the change in accounting policy, revenue, operating profit, ordinary profit, and net income for cumulative Q3 decreased by JPY42mn.
The company made no change to its full-year forecast for FY06/22. The progress versus the full-year forecast was 77.7% for revenue (versus 86.7% in cumulative Q3 FY06/21 vs. full-year result), 85.7% for operating profit (104.3%), 83.9% for recurring profit (104.6%), and 87.0% for net income (107.1%) .
In cumulative Q3 FY06/22, revenue was JPY5.0bn (-1.2% YoY), breaking down into JPY3.6bn for the HPC segment (-2.2% YoY) and JPY1.4bn for the CTO segment (+1.4% YoY). The cost of revenue ratio was 68.8%, down 0.9pp YoY due to the recording of revenue from high-margin projects, which offset higher material prices. The SG&A ratio was 18.4%, up 2.1pp YoY on higher personnel expenses stemming from aggressive hiring in line with the medium-term management plan. Operating profit was JPY638mn (-9.6% YoY) and the OPM was 12.8% (-1.2pp YoY).
HPC segment
Revenue: JPY3.6bn (-2.2% YoY)
Operating profit: JPY488mn (-12.7% YoY)
Orders declined YoY in Q3 (three months). Orders from universities and government offices were down YoY due to the absence of extraordinary demand seen in the previous year. In the private sector, orders increased in the computational chemistry solutions (CCS) and computer aided engineering solution (CAES) areas, but decreased in the data science solution (DSS) area.
The company recorded revenue in Q3 from large orders which it received in Q4 FY06/21. Revenue from the private sector increased 59% YoY, backed by strong performance of DDS projects for AI and deep learning development platforms and CCS projects for materials development and drug discovery research. On the other hand, revenue from universities and government offices declined 48% YoY in reaction to the extraordinary demand seen in the previous fiscal year.
The company was able to sell its science cloud system to overseas customers for the first time in Q1. HPC Systems has positioned overseas expansion as a key initiative in its medium-term plan (See "(5) Advance into overseas markets and promote related operation” in the Medium- to long-term outlook section).
OPM declined 1.7pp YoY to 13.6%. This was attributable to an increase in SG&A expenses on a higher headcount (another key policy of the medium-term plan, "(2) Cultivate human capital and recruit human resources").
CTO segment
Revenue: JPY1.4bn (+1.4% YoY)
Operating profit: JPY150mn (+2.4% YoY)
Orders were up 1.5x YoY in Q3 (three months). Orders from repeat customers increased for SF machine vision projects for semiconductor-related equipment (wafer inspection, etc.) and Edge IoT projects. Orders from new customers were robust for Edge IoT and Edge AI projects.
Revenue from repeat customers increased YoY. Major contributions came from Edge IoT projects for medical equipment and amusement equipment, as well as from SF machine vision projects for semiconductor-related equipment (wafer inspection, etc.). Revenue from new customers declined due to the absence of large orders received in the previous year.
OPM improved 0.1pp YoY to 10.7%. Although personnel expenses rose on a higher headcount (a key policy of the medium-term plan) and depreciation expenses increased due to the upgrade of testing and inspection equipment, these were more than offset by the recording of revenue from high-margin projects.
For details on previous quarterly and annual results, please refer to the Historical results section.
Company forecast for FY06/22
Earnings
FY06/20
FY06/21
FY06/22
(JPYmn)
1H
2H
FY
1H
2H
FY
1H
2H Est.
FY Est.
Revenue
2,360
2,366
4,725
2,326
3,502
5,828
2,237
4,188
6,425
YoY
-
-
-12.4%
-1.4%
48.0%
23.3%
-3.8%
19.6%
10.2%
Cost of revenue
1,556
1,619
3,175
1,594
2,428
4,022
1,508
YoY
-
-
-20.6%
2.5%
50.0%
26.7%
-5.4%
Cost of revenue ratio
65.9%
68.4%
67.2%
68.5%
69.3%
69.0%
67.4%
Gross profit
804
747
1,551
732
1,074
1,806
728
YoY
-
-
10.9%
-9.0%
43.9%
16.5%
-0.4%
Gross profit margin
34.1%
31.6%
32.8%
31.5%
30.7%
31.0%
32.6%
SG&A expenses
553
520
1,073
533
596
1,129
591
YoY
-
-
4.3%
-3.7%
14.7%
5.2%
10.8%
SG&A ratio
23.4%
22.0%
22.7%
22.9%
17.0%
19.4%
26.4%
Operating profit
251
227
478
198
478
677
138
607
745
YoY
-
-
29.3%
-20.8%
110.5%
41.6%
-30.6%
27.0%
10.1%
Operating profit margin
10.6%
9.6%
10.1%
8.5%
13.7%
11.6%
6.2%
14.5%
11.6%
Recurring profit
238
227
465
201
465
666
128
613
741
YoY
-
-
26.8%
-15.8%
104.8%
43.1%
-36.1%
31.7%
11.3%
Recurring profit margin
10.1%
9.6%
9.8%
8.6%
13.3%
11.4%
5.7%
14.6%
11.5%
Net income
161
146
307
137
310
447
85
411
496
YoY
-
-
40.1%
-15.3%
112.5%
45.4%
-38.0%
32.5%
10.9%
Net margin
6.8%
6.2%
6.5%
5.9%
8.9%
7.7%
3.8%
9.8%
7.7%
Source: Shared Research based on company data
Note: Figures may differ from company data due to differences in rounding methods.
For FY06/22, HPC Systems forecasts revenue
of JPY6.4bn (+10.2% YoY), operating profit of JPY745mn (+10.1% YoY), recurring
profit of JPY741mn (+11.3% YoY), and net income of JPY496mn (+10.9% YoY).
The company has compiled conservative
forecasts for FY06/22 as it expects shortages of semiconductors and peripheral
electronic components to cause disruptions in its supply chain.
HPC segment
HPC Systems forecasts revenue of JPY4.3bn
(+9.4% YoY) in FY06/22. Following a rise in revenue from universities and
public research institutions driven by the government’s supplementary
budget in FY06/21, the company expects demand to revert to normal levels in
FY06/22. Demand from private companies is rebounding across the board, and demand
for drug discovery, materials development, and AI-related applications has been
particularly strong. The company also secured a large order related to machine
learning in Q4 FY06/21, and plans to record related revenue in FY06/22.
CTO segment
The company forecasts revenue of JPY2.1bn
(+12.0% YoY) in FY06/22. It expects a recovery in demand for medical equipment
and amusement equipment applications after a slump in demand attributable to
the COVID-19 pandemic in FY06/21. In addition, demand for machine vision and
inspection equipment applications also remains strong.
Machine vision: A form of computer-based image processing used in areas such as automated inspections, process control, and robot coordination. Specific tasks that can be achieved with machine vision include counting the number of objects on a conveyor belt, scanning serial numbers, and detecting damage on the surface of objects.
Company forecasts versus results
The table below shows how HPC Systems’ results have compared with its forecasts. The company revised up its FY06/21 forecast when it announced Q3 FY06/21 results.
Company forecasts versus results
Results vs. Initial Est.
FY06/20
FY06/21
(JPYmn)
Revenue
Initial Est.
5,785
5,200
Results
4,725
5,828
Results vs. Initial Est.
-18.3%
12.1%
Operating profit
Initial Est.
469
505
Results
478
677
Results vs. Initial Est.
1.9%
34.0%
Recurring profit
Initial Est.
457
495
Results
465
666
Results vs. Initial Est.
1.8%
34.5%
Net income
Initial Est.
309
330
Results
307
447
Results vs. Initial Est.
-0.5%
35.5%
Source: Shared Research based on company data
Note: Figures may differ from company data due to differences in rounding methods.
Medium- to long-term outlook
HPC Systems aspires to be a “global visionary company of the 21st century that specializes in the field of computer science” with the capacity to endure through several eras and continue operating beyond 2030. As a medium-term goal, it aims to become the leader in a global niche market by 2030. It has formulated the medium-term management plan “Vision 2024” (FY06/22–FY06/24) using a backcasting approach, and unveiled the plan in August 2021.
As numerical targets, the plan calls for FY06/24 revenue of JPY10.0bn (three-year CAGR of 19.7%), operating profit of JPY1.3bn (23.8%), and ROE of 26% (28.1% in FY06/21). The breakdown by segment is revenue of JPY7.0bn (three-year CAGR of 21.0%) and operating profit of JPY960mn (24.9%) in the HPC segment, and revenue of JPY3.0bn (17.0%) and operating profit of JPY320mn (20.9%) in the CTO segment. In addition, the plan looks to keep ROE above 25% through FY06/24. It outlines specific priority measures for each segment, which are discussed in detail below.
Numerical targets included in medium-term plan “Vision 2024” (FY06/22–FY06/24)
FY06/21
FY06/24
3-year CAGR
(JPYmn)
Results
Targets
Revenue
5,828
10,000
19.7%
HPC business
3,953
7,000
21.0%
CTO business
1,875
3,000
17.0%
Operating profit
677
1,280
23.7%
Operating profit margin
11.6%
12.8%
HPC business
494
960
24.8%
Operating profit margin
12.5%
13.7%
CTO business
183
320
20.5%
Operating profit margin
9.8%
10.7%
ROE
28.1%
26%
Source: Company data
Based on its own survey, HPC Systems estimates
the target market for the HPC segment will expand at a three-year CAGR of 13.0%
from JPY260.0bn in 2021 to JPY380.0bn in 2024, and aims for above-market growth
of 21.0% in the segment. Similarly, it estimates the target market for the CTO
segment will grow at a CAGR of 11.1% from JPY187.0bn in 2021 to JPY257.0bn in 2024,
and aims for above-market growth of 17.0% in the segment.
Based on its own survey, HPC Systems estimates the target market for the HPC segment will expand at a three-year CAGR of 13.0% from JPY260.0bn in 2021 to JPY380.0bn in 2024, and aims for above-market growth of 21.0% in the segment. Similarly, it estimates the target market for the CTO segment will grow at a CAGR of 11.1% from JPY187.0bn in 2021 to JPY257.0bn in 2024, and aims for above-market growth of 17.0% in the segment.
As priority measures, the company will
(1) evolve its core businesses, (2) cultivate human capital and recruit human
resources, (3) create new businesses, (4) strengthen management foundations,
and (5) advance into overseas markets and promote related operations.
(1) Evolve core businesses
Priority measures in the HPC segment
(1) Develop practical applications for computational chemistry and materials informatics
The company aims to increase revenue in this field from JPY400mn in FY06/21 to JPY1.0bn in FY06/24. In addition to the computational chemistry software it already provides, it plans to promote the adoption and expansion of materials informatics and the Global Reaction Route Mapping (GRRM) program it has jointly developed with Hokkaido University.
Computational science: A field of study that builds numerical and quantitative evaluation models to resolve scientific problems with computers.
Materials informatics: A technology that utilizes artificial intelligence (AI) applications such as machine learning and deep learning to build statistical models from databases that map the relationships between the attributes and chemical structure of materials. By predicting the attributes of materials based on their chemical structure, it facilitates the discovery of chemical structures that possess a desired set of attributes, and thus contributes to enhanced efficiency in new materials development.
(2) Standardize high-performance computing (HPC) systems integration
Integrate heterogenous hardware environments and diverse software in HPC-AI platforms while considering future strategies. As a KPI, expand the new adoption rate for HPC-AI platforms to 50%.
Heterogeneous: A condition in which equipment, devices, and software with different roles and uses are combined into one system, and function on the basis of mutual coordination and division of roles.
(3) Diversify science cloud services
Under its science cloud services, HPC Systems provides various types of scientific computing software to customers over the cloud. The company will aim to convert 20 or more new customers into repeat customers for its cloud services.
Priority measures in the CTO segment
HPC Systems aims to expand revenue from repeat customers in the CTO segment by 130% from FY06/21 to FY06/24. It believes it will need to develop growth drivers and strengthen its business foundations to achieve this target.
The company envisions the following growth drivers: (1) cultivating industry knowledge, (2) accelerating the development of customers through combination technologies, and (3) applying network technologies. It will aim to understand the challenges their customers face on the ground as well as their core needs, and supply the optimal products customers require through its proprietary combination technologies.
To strengthen its business foundations, HPC Systems plans to (1) upgrade its plants, (2) forge ahead with digital transformation (DX: adopting of digital technologies to optimize operations) and reform its business processes, and (3) improve its production technology and quality-cost-delivery (QCD). To upgrade its plants, the company looks for an 50% increase its production capacity by 2024 by making more efficient use of its current space. It will also consider expanding existing plants and building new facilities. To forge ahead with DX and reform its business processes, the company will promote the sharing and standardization of operations, diversify materials procurement channels, and promote initiatives related to chemSHERPA.
chemSHERPA: An information transmission and sharing scheme under which information on chemical materials contained in products is adequately managed across the entire supply chain from upstream to downstream companies, and effectively and efficiently transmitted.
(2) Cultivate human capital and recruit human resources
The company plans to increase its human
resources by 20–50% from FY06/21. To this end, it will actively recruit personnel in
specialized fields, provide internal training, and expand skill enhancement
programs.
(3) Create new businesses
The company looks to create new
businesses that can become fresh earnings drivers. Specifically, it plans to leverage
synergies between its core businesses to create businesses in the fields of
machine learning and AI. At the same time, it will work to expand and promote
business alliances.
(4) Strengthen management foundations
The company plans to promote DX, develop
business tools, and adopt robotic process automation (RPA) to strengthen its
management foundations. It will also develop and operate marketing tools.
Robotic process automation (RPA): Software bots that use rules-based engines or AI to automate deskwork ordinarily performed by white collar workers.
(5) Advance into overseas markets and promote related operations
On the overseas front, the accompany plans to (1) move into the European and US markets by forming alliances with partners in the computational chemistry field, (2) reinforce its base in Vietnam, (3) strengthen the functions of its base in Taiwan, and (4) cultivate the Chinese market. In the HPC segment, it has handled orders for the Vietnamese government and Vietnam National University in collaboration with a major trading house. By FY06/24, it plans to increase overseas revenue to JPY500mn–JPY1.0bn.
Market environment
In March 2021, the Japanese Cabinet approved and set in motion the 6th Science, Technology, and Innovation Basic Plan that maps out the science and technology policy of the government for the five years from FY2021. The government budgets JPY30tn in R&D spending over the five-year period, and hopes this will spur private investment. On the whole, it targets public-private investment of JPY120tn.
The plan calls for a “societal vision predicated on innovation (societal change)” as its first theme, and promotes digitalization and the implementation of data-driven strategies in the government, as well as the preparation and development of next-generation infrastructure and technology such as supercomputers and quantum technology.
The government will establish a JPY10tn university fund, and establish a research system for universities that can compete on the global stage.
Business
Business overview
As its mission, HPC Systems seeks to “provide researchers and developers with the tools to research and develop products.”
The company operates in two segments: High Performance Computing (HPC; 67.8% of total revenue and 73.0% of total operating profit in FY06/21) and Configure to Order (CTO; 32.2%, 27.0%).
Revenue by product offerings
FY06/17
FY06/18
FY06/19
FY06/20
FY06/21
(JPYmn)
Revenue
3,901
4,053
5,396
4,725
5,828
YoY
3.9%
33.1%
-12.4%
23.3%
HPC business
2,738
2,648
3,787
3,198
3,953
YoY
-3.3%
43.0%
-15.5%
23.6%
% of total
70.2%
65.3%
70.2%
67.7%
67.8%
CTO business
1,163
1,405
1,609
1,527
1,875
YoY
20.8%
14.6%
-5.1%
22.8%
% of total
29.8%
34.7%
29.8%
32.3%
32.2%
Operating profit
244
282
370
478
677
YoY
15.6%
30.9%
29.3%
41.6%
Operating profit margin
6.3%
7.0%
6.8%
10.1%
11.6%
HPC business
152
139
211
299
494
YoY
-8.6%
52.2%
41.3%
65.3%
Operating profit margin
5.5%
5.2%
5.6%
9.3%
12.5%
% of total
62.2%
49.2%
57.2%
62.5%
73.0%
CTO business
92
143
158
179
183
YoY
55.3%
10.3%
13.2%
2.1%
Operating profit margin
7.9%
10.2%
9.8%
11.7%
9.8%
% of total
37.8%
50.8%
42.8%
37.5%
27.0%
Source: Shared Research based on company data
Note: Figures may differ from company data due to differences in rounding methods.
HPC segment
In this segment, HPC Systems develops, manufactures, and sells hardware and software for scientific computers. Its customers are mainly universities and R&D divisions of private companies. Scientific computers such as supercomputers and workstations differ from their consumer-oriented counterparts in that they deliver vastly higher performance. The aforementioned R&D divisions rely on computational science to develop new materials, discover new drugs, and conduct simulations of fluid structures, which are an essential part of product manufacturing. The company provides the environments (integrated hardware, operating system [OS], middleware, software, and network, and related optimization) to facilitate these processes.
In this segment, HPC Systems develops, manufactures, and sells hardware and software for scientific computers. Its customers are mainly universities and R&D divisions of private companies. Scientific computers such as supercomputers and workstations differ from their consumer-oriented counterparts in that they deliver vastly higher performance. The aforementioned R&D divisions rely on computational science to develop new materials, discover new drugs, and conduct simulations of fluid structures, which are an essential part of product manufacturing. The company provides the environments (integrated hardware, operating system [OS], middleware, software, and network, and related optimization) to facilitate these processes.
Core customers
The company has supplied products and services to private companies operating in the pharmaceuticals, chemicals, materials, automotive, electrical and electronic equipment, and heavy industry sectors. It also counts many public research institutions and universities among its customers.
From FY06/19 to FY06/21, the HPC segment accounted for 69% of total revenue. By strategic (customer) area, computer science solutions (CSS: fundamental research into science and technology) for universities and public research institutions made up 29%, data science solutions (DSS: machine learning and AI R&D) 21%, computer-aided engineering solutions (CAES: simulation data analysis during design phase) for automotive, general electrical machinery, optical equipment, precision equipment, and information and communication equipment companies 9%, computer chemistry solutions (CCS: drug discovery and materials development) for comprehensive chemicals, derivatives, electronic materials, drug discovery, and food companies 8%, and HPC infrastructure (HPCI) 2%. Over the same period, the CTO segment generated 31% of total revenue. By strategic (customer) area, Edge IoT for medical and amusement equipment accounted for 15%, smart factory solutions for manufacturing equipment and plant facilities (visual inspections and semiconductors) for 12%, and edge AI for infrastructure monitoring and security equipment for 4%.
Systems delivered to customers
Private companies
[Automotive]
[Electronics, electric equipment, home appliances]
[Chemistry and materials]
[Manufacturing, cosmetics, food]
・Toyota Motor
・Panasonic
・Sumitomo Chemical
・Takeda Pharmaceutical
・Honda Motor
・Sony
・Mitsubishi Chemical
・Astellas Pharma
・Nissan Motor
・Hitachi
・AGC
・Otsuka Pharmaceutical
・Mazda Motor
・Mitsubishi Electric
・Mitsubishi Gas Chemical
・Kao
・Mitsubishi Motors
・Kyocera
・Nippon Shokubai
・Chugai Pharmaceutical
・Hino Motors
・Fuji Xerox
・Nissan Chemical
・Yakult
・Yamaha Motor
・Nikon
・Kuraray
・Roche Group
・Daimler AG
・Omron
・Daicel
・Mitsubishi Tanabe Pharma
・TRI-AD
・Toshiba
・Kaneka
・Shionogi
・Advanced Smart Mobility
・Casio
・Toray Industries
・Ajinomoto
・J-QuAD Dynamics
・Canon
・Asahi Kasei
・Kyowa Kirin
・TOTO
・Toyobo
・Sumitomo Dainippon Pharma
[Communication, Internet]
・Tokyo Electron
・Mitsui Chemicals
・SoftBank
・HOYA
・Taiyo Yuden
[Heavy Industry]
・NTT
・Daikin Industries
・Teijin
・ENEOS
・KDDI
・Konica Minolta
・Showa Denko
・Kawasaki Heavy Industries
・NTT Docomo
・Nabtesco
・Tokyo Ohka Kogyo
・Mitsubishi Heavy Industries
・Yahoo Japan
・NGK Insulators
・Mitsui E&S Holdings
・CyberAgent
[AI]
・NGK Spark Plug
・IHI
・GMO Internet
・AI inside
・Tosoh
・Hitachi Zosen
・SenseTime Japan
・Fujifilm Holdings
・Nippon Steel
[Construction]
・Magne-Max Capital Management
・Shimizu
・Cybercore
・Taisei
・DEEPCORE
・Chiyoda
・PKSHA Technology
・Kajima
・Morpho
・Takenaka
Public research institutions and universities
[Public research institutions]
[Universities]
・Institute of Physical and Chemical Research (RIKEN)
・The University of Tokyo
・Keio University
・Iwate University
・Japan Aerospace Exploration Agency
・Kyoto University
・Chuo University
・University of Tsukuba
・Technical Research and Development Institute, Ministry of Defense
・Osaka University
・Waseda University
・Shizuoka University
・National Institute of Advanced Industrial Science and Technology
・Tohoku University
・Sophia University
・Rikkyo University
・High Energy Accelerator Research Organization
・Kyushu University
・Tokushima University
・Kanazawa University
・Japan Atomic Energy Agency
・Hokkaido University
・Chukyo University
・Tokyo City University
・Japan Meteorological Agency
・Nagoya University
・Toyama Prefectural University
・Vietnam National University
・Japan Agency for Marine-Earth Science and Technology
・Yamagata University
・Tokyo University of Science
・Stanford University
・Research Institute of Innovative Technology for the Earth
・Tokyo Institute of Technology
・Saga University
・Princeton University
・National Institutes of Natural Sciences
・Japan Advanced Institute of Science and Technology
・Texas Advanced Computing Center
Source: Shared Research based on company data
Some examples of supercomputer tenders of universities or public research institutions that have been awarded to the company are shown below.
Tenders of universities and other public institutions awarded to the company
Awarding institution
Project
Bidding date
Awardee
Contract price (JPYmn)
Tokyo University of Agriculture and Technology
Deep-learning computing system for analysis of large brainwaves (full project)
Jan. 2021
HPC Systems Inc.
13
Tokyo University of Agriculture and Technology
Computing system for large-scale deep learning
Jan. 2021
HPC Systems Inc.
22
Tokyo University of Agriculture and Technology
Computing system for analysis of large-scale, complex data sets (full project)
Jan. 2021
HPC Systems Inc.
13
The University of Tokyo
General-purpose computing platform system
Oct 2020
HPC Systems Inc.
98
National Institute of Information and Communications Technology
Procurement of server farm to support physical downsizing for the NICTER system
Sep 2020
HPC Systems Inc.
68
The University of Tokyo
CPU/GPU computing system for quantum algorithm verification equipment
Sep 2020
HPC Systems Inc.
35
Hokkaido University
High-performance computing system
May 2020
HPC Systems Inc.
86
Hokkaido University
High-performance computing system
Dec 2019
HPC Systems Inc.
99
Source: Shared Research based on data compiled by HPC Wire Japan
Examples of systems deployed by the company at customers are shown below.
Systems deployed by the company
Customer
Case study
International Center for Elementary Particle Physics, The University of Tokyo
Exploration of supercomputer application potential in data analysis with PRIMEHPC FX700
Shizuoka University
ChemPark cloud service used for chemical computing task in classrooms
Tokyo University of Agriculture and Technology
Reconfiguration of continuous physical spaces from complex, large datasets with signal information processing and fundamental machine-learning technologies
Tokyo City University
Potential of COCOA contact-tracing app for COVID-19
Yahoo Japan Corporation (TI Office of CTO Technology Strategy Division)
Immersion-cooled supercomputer kukai, the world’s second-best energy-saving supercomputer
Nara Institute of Science and Technology
Promotion of paradigm shift to support the cultivation of a new material science
Tokai University, Tokai Formula Club
Construction of a formula one car
Nihon University
Design and development of ultra-high-speed and ultra-high-density next-generation magnetic recording
Rhelixa Co., Ltd.
Proprietary computational algorithm to decode the genome information
Tokyo University of Technology
Web application FEHA that provides simple learning environment for programming
Hoshi University
Development of practical FMO drug discovery technology through industry-academia-government collaboration / Use of FMO drug discovery consortium and supercomputer k
True Cell Simulations
Acceleration of basic research, drug discovery, and medical research with 4D cell simulation
Rikkyo University
Development of new molecular asymmetric catalyst based on a molecular-level understanding
Yokohama Soei University
Simulation-based analysis of structural flaws in crystals
Kyushu University
Realization of organic EL device with internal emission quantum efficiency of 100%
Maxis One Ltd.
Contribution to advancements in crystal device industry with proprietary technology
Hiroshima University
Advancing the field of chemistry by elucidating reaction behavior
The University of Tokyo
To elucidate the basic principles of life
Ibaraki University
Growing crops in a computer to assess the impact of global warming
Utsunomiya University
Dynamics of interdependent balance state and corresponding control
Kogakuin University Shinjuku Campus
Movement of molecular quantum dot cell automaton and theoretical analysis
Waseda University
Analysis of chemical reaction at solid-liquid interface with theoretical calculations
Kogakuin University Shinjuku Campus
Supporting the development of high-performance sports equipment with material mechanics
Kyushu University
Establishing numerical wind prediction technology that supports wind power generation
Using computational chemistry to elucidate the structure of organic compounds
Tottori University
Projecting changes in water environments via field observations and numerical calculations
Nagaoka University of Technology
Attempting direct detection of gravitational waves
Hachinohe Institute of Technology
Developing chemical reaction process to use microwave irradiation as a heat source
Utsunomiya University
Theoretically projecting turbulence phenomena in various settings
Iwate University
Reconstructing and regenerating Iwate with thermal fluid analysis technology
Chukyo University
Contributing to the design of new devices with large-scale calculation and spectroscopy technology
Yokohama City University
Simultaneous development of theory and application
Nabtesco Corporation
Cluster system that supports original motion technology
Source: Shared Research based on company website
Competitive landscape
Suppliers of supercomputers include major companies such as Fujitsu Limited (TSE1: 6702), NEC Corporation (TSE1: 6701), and Hewlett Packard Enterprise (NYSE: HPE). To deliver the operating environments required by its customers, HPC Systems builds computer clusters that connect workstations, servers, and storage devices through high-speed networking equipment, installs the necessary software, and optimizes the computer clusters as supercomputers. The hardware and software that make up these systems differ from consumer counterparts in that getting them to operate requires more than simply establishing the necessary connections and installing the appropriate software. The hardware, operating system (OS), middleware, software, and network must be configured and optimized to run multiple computational processes in parallel while reducing the processing time, and this requires the expertise of a company like HPC Systems.
The company identifies the following as its main areas of business.
System overview by price bracket
Source: Shared Research based on company data
Other companies that operate in the systems integration field are HPC Solutions Inc. (unlisted, consolidated subsidiary of Argo Graphics Inc. [TSE1: 7595]), HPC Tech Corporation (unlisted, consolidated subsidiary of AKIBA Holdings Co., Ltd. [TSE JASDAQ: 6840]), and Visual Technology, Inc. (unlisted). However, HPC Systems does not actually compete with these companies in the computational chemistry field. Scientific computing is used in a wide range of fields such as elementary particle and astronomical research, weather forecasting, global warming projections, predictions of disasters such as earthquakes and tsunamis, drug discovery, development of semiconductors and battery materials, and structural and aerodynamic analysis of vehicles and aircraft. HPC Systems excels in building scientific computing systems for the life science and material science fields, and its customers are mainly universities, public research institutions, and R&D divisions of private companies.
Hardware
HPC Systems procures and delivers hardware
and software in accordance with the needs of its customers. It mainly uses
products supplied by Super Micro Computer, Inc. (NASDAQ: SMCI), Hewlett-Packard
Company, GIGA-BYTE Technology Co., Ltd. (TWSE: 2376, Taiwan), DataDirect
Networks, Inc. (unlisted, US), and Graphcore Limited (unlisted, UK). It meticulously
selects hardware in function of customer budgets and desired research
environments. However, the company has particularly strong ties with Super
Micro Computer, and therefore sources a lot of its hardware from this supplier.
Workstation
Rack-mounted server
Storage system
Source: Company materials
Super Micro Computer
HPC Systems has concluded a distributor agreement with Super Micro Computer, which has become a core supplier of servers for the company. From FY09/18 to FY09/21, Mr. Tau Leng, currently the Senior Vice President of Technology at Super Micro Computer, served as an outside director at HPC Systems. Super Micro Computer’s plant in Taiwan automatically configures the specifications (CPU, memory, GPU, OS, software, etc.) for orders placed by the company using an install system developed by HPC Systems. As a result, the company does not have to open the products to install parts or software after they have arrived in Japan. Thanks to these arrangements and the discounts it receives on bulk orders, HPC Systems can rapidly provide products manufactured by Super Micro Computer at affordable prices.
Incidentally, according to a survey conducted by IDC Japan in March 2021, Fujitsu Ltd. led the domestic server market in 2020 with a share of 27.4%, followed by NEC Corporation at 17.8%, Hewlett-Packard Japan, Ltd. at 12.7%, Dell Technologies Japan Inc. at 11.2%, Hitachi, Ltd. at 6.0%, other companies at 13.6%, and ODM production at 11.3%. Super Micro Computer was included under ODM production in the survey. On its website, Super Micro Computer states it was the world’s fifth largest supplier of servers by shipments in 2019.
Super Micro Computer works with several distributors in Japan, HPC Systems being one of them.
Software
We now turn to the software offered by HPC Systems for scientific computing. The company handles a broad range of software packages in the field of quantum chemistry, proposing solutions tailored to the needs of its customers. One example of such a software package is Gaussian. Developed by Gaussian, Inc., this quantum chemistry computing application uses cutting-edge technology to model the electronic structure of various compounds. It is used by many researchers in the fields of chemistry, biochemistry, and physics, and has become the de-facto standard for research in these areas. In Japan, Gaussian is sold by HPC Systems, Hulinks Inc. (unlisted, consolidated subsidiary of Argo Graphics Inc. [TSE1: 7595]), Conflex Corporation (unlisted), and Molsis Inc. (unlisted).
In the quantum chemistry and molecular dynamics fields, HPC Systems sells software developed by its own engineers, many of whom hold relevant PhD or master’s degrees. Having used standard software packages such as Gaussian and AMBER during their academic training, these engineers have a firm grasp of the functions that are either missing or difficult to use in these products, and develop software that makes up for these flaws. For example, the company released the Reaction Plus Pro software in June 2018. This software relies on Gaussian for computational processes but extends its functionality.
In July 2020, HPC Systems concluded a joint research agreement with the Faculty of Science at Hokkaido University. The aim of the agreement is to jointly research global reaction route mapping (GRRM), and commercialize related outcomes in the science cloud services offered by the company.
In February 2021, the company
successfully developed a groundbreaking molecular structure generation
algorithm as part of its materials informatics program development, and applied
for a corresponding patent.
Materials informatics: A new methodology in the materials development field that leverages big data analysis and AI technologies to reduce the costs and accelerate the pace of materials development. In the past, the molecular design process in materials development entailed synthesizing materials deemed to express a desired set of attributes. This approach is still commonly used in R&D today, but its drawback is that the development of new materials depends on the experience and intuition of researchers. In recent years, materials informatics has been promoted as a new methodology that takes advantage of machine learning technology underpinned by massive amounts of data to predict the attributes of new materials in a short time without having to rely on simulations.
Examples of scientific computing software packages supplied by the company
Source: Shared Research based on company data
Science cloud services
The company launched its science cloud
services in June 2018. Because they are delivered over the Internet, these
services allow customers to perform scientific computing on PCs, removing the
need to purchase expensive computers or software. The science cloud services have
been mainly adopted by chemical manufacturers to date.
In April 2021, the company also started
providing a paid science cloud service that demonstrates computing processes running
on the supercomputer Fugaku.
Overview of science cloud services
Entry plan
Standard plan
Enterprise plan
User base
Elementary
Intermediate and above
Daily use of HPC
Available computational chemistry software
Gaussian
Gaussian
Provided upon customer request
Reaction plus Expresso
Reaction plus Pro2
NWChem
GRRM-Neo11
Computer-aided engineering (CAE: structure, fluids, etc.) also available
LAMMPS
Quantum ESPRESSO
GROMACS, etc.
Pricing
From JPY98,000/month
From JPY98,000/month
Estimate provided
Usage terms
Starting from 30 min limit per computing session
Unlimited computing time
Unlimited computing time
Source: Company website
Process from order to delivery
It takes roughly three months for HPC
Systems to deliver orders. Upon receiving an order from a customer, the company
first orders the necessary parts, which take six weeks to two months to arrive.
Thereafter, it installs the operating system (OS) and applications, and
confirms the system operates as expected. Transactions with overseas hardware
and software vendors such as Super Micro Computer are settled in dollars, so
HPC Systems is susceptible to forex impact. However, because the lead time for
orders is about three months, the company can often pass on forex-driven price
increases to customers, so it does not shoulder major forex risk. The largest concern
for the company as of September 2021 is the global shortage of semiconductors,
as a result of which deliveries of parts could conceivably suffer delays.
Order format
While it also bids for tenders issued by
universities and public research institutions, HPC Systems mainly secures
orders via direct customer inquiries. Customers typically refer to the product
lineup and benchmarks (performance indicators) on the company’s website
while they are researching the optimal hardware and software for their R&D
objectives, and subsequently reach out to the company. Upon receiving a request,
the company’s sales and technical staff explain and propose products to the
customer. Customers are chiefly concerned with the benchmarks of the systems
they intend to purchase, and the company lists verified numerical benchmarks on
its website. HPC Systems also sells products through third parties under agency
agreements, but the revenue contributions from such arrangements are small
because the company does not sell general-purpose products. In April 2020, the
company formed a capital and business alliance with AS ONE Corporation (TSE1:
7476), and started offering its products and services on the AXEL e-commerce
platform operated by AS ONE in October 2020.
Benchmarks (numerical performance indicators)
HPC Systems measures benchmarks for systems that use the latest hardware (including next-generation processors), runs various types of optimized simulation software, and publishes these on its website. It uses benchmark software to measure and evaluate the quantitative processing capacity of the hardware and software employed. Customers can browse these benchmarks on its website and inquire about them before placing an order. In FY06/21, the company budgeted JPY76mn for capital investment, and allocated JPY58mn toward investments in servers and other hardware used to obtain benchmarks.
System, software, and application engineers
At end-June 2021, HPC Systems employed 92 people, of which 26 engineers engaged in R&D and verification technology. Its engineer team includes nine individuals with PhD degrees, six with master’s degrees, and 15 postdoctoral researchers in fields such as computational chemistry, organic synthesis, theoretical chemistry, machine learning, flow simulation, and physics. The company’s engineers are all experienced users of the Gaussian software package that has become the de-facto standard software for quantum chemistry. Having used the software during their academic training, they have a firm grasp of its shortcomings (such as missing functions), and use this information to develop the many proprietary software packages that are sold by the company.
HPC Technology Development Center
The company is headquartered in Minato-ku, Tokyo, but it operates its HPC Technology Development Center in Chuo-ku, Tokyo. From the center, its systems and application engineers provide sales support, develop proprietary software, and conduct technical research for its science cloud services.
Overseas operations
In March 2020, HPC Systems established
Intelligent Integration Company Limited, a wholly owned subsidiary, in Vietnam.
In the past, the company has handled orders for the Vietnamese government and Vietnam
National University in the HPC segment. Its customers in the CTO segment have
production bases in Vietnam, so the company expects to win related orders in
the future. It also plans to expand its operations in Asian markets other than
Vietnam.
HPC Systems has yet to handle orders in the US or Europe, but it is planning a move into these markets. Specifically, it intends to launch businesses that take advantage of its science cloud services.
Technology trend (1) Supercomputer Fugaku
The supercomputer Fugaku developed by Fujitsu Ltd. and RIKEN is the world’s most powerful supercomputer. As of September 2021, nine joint research projects were underway to expand the use and improve the usability of Fugaku, and HPC Systems was selected for one of these project. Since April 2021, HPC Systems has offered a paid cloud-based service that demonstrates its computational chemistry simulation software running on the supercomputer Fugaku.
Source: Fujitsu website
Supercomputer Fugaku: The Nikkei has reported the supercomputer Fugaku cost JPY130.0bn to develop (of which JPY111.1bn funded by the Japanese government according to the Mainichi Shimbun). The system was launched on a trial basis in 2020, and entered full-scale operation in March 2021. It topped the list of the world’s 500 most powerful computers in June 2020 and November 2020. It succeeds the supercomputer K (or “Kei” in Japanese), which was ranked the world’s most powerful computer in 2011.
Technology trend (2): Immersion-cooled servers
The kukai server delivered by HPC Systems to Yahoo Japan Corporation uses a liquid immersion cooling system, and has been ranked as the world’s second-best energy-saving supercomputer by GREEN500, a biannual ranking of the most energy-efficient supercomputers in the world. Servers are generally air-cooled, which means fans are used to divert the heat generated by their internal parts. The immersion-cooled server delivered by the company is entirely submerged in a thermally conductive dielectric liquid or coolant contained within a water tank, which contributes to a reduction in power consumption. This is a field poised for growth in the future.
Source: Company website
Technology trend (3): Quantum computers
Quantum computers are next-generation
computers that can perform calculations at far greater speeds than
supercomputers. The technology is still in the development phase, and it is said
to be two decades away from commercialization. Quantum computers perform
calculations based on technological principles that are drastically different
from classical computers. In November 2020, HPC Systems formed a business
alliance with QunaSys Inc. (unlisted). The company aims to fuse its advanced technologies and solutions in computational
chemistry with the cutting-edge technologies of QunaSys, a developer of quantum
computer applications, to develop the world’s first
computational chemistry solution for quantum computers.
Orders, revenue, and operating profit
The company’s revenue remains in an overall growth trend from a medium-term perspective, although it has declined in some years. The main factors driving fluctuations in revenue are the acquisition of large orders and the timing of revenue recognition.
In FY06/17 and FY06/19, HPC Systems handled major projects for Yahoo Japan Corporation worth over JPY400mn. OPM has improved mainly due to (1) a reduction in SG&A expenses (attributable to a relatively steady workforce) amid revenue expansion, and (2) the absence of impact from major projects with low margins.
HPC segment: Orders, revenue, and operating profit
FY06/17
FY06/18
FY06/19
FY06/20
FY06/21
(JPYmn)
Orders
3,858
3,389
5,127
YoY
-
-12.1%
51.3%
Revenue
2,738
2,648
3,787
3,198
3,953
YoY
-
-3.3%
43.0%
-15.5%
23.6%
Operating profit
152
139
211
299
494
YoY
-
-8.6%
52.2%
41.3%
65.3%
Operating profit margin
5.5%
5.2%
5.6%
9.3%
12.5%
Segment assets
1,048
1,353
1,461
1,839
YoY
-
29.1%
8.0%
25.8%
Segment ROA
-
17.6%
21.2%
29.9%
Source: Shared Research based on company data
Note: Figures may differ from company data due to differences in rounding methods.
CTO segment
In the CTO segment, HPC Systems develops, manufactures, and sells industrial computers that are configured to order. It operates a plant in Sosa City, Chiba Prefecture, from where it supplies products in accordance with customer needs using a high-mix/flexible-volume approach. Its product lineup includes inspection equipment computers that inspect and monitor products manufactured in plants, IoT computers used in plants to manage and monitor networks through on-site cameras, and high-spec industrial computers used for medical image processing (computed tomography [CT] scans and magnetic resonance imaging [MRI]). The company procures parts, cases that have been designed internally, and software, and subsequently develops and assembles these into systems that are supplied to customers under the HPC Systems brand. It says its repeat purchase rate exceeds 90%.
Competitors
The company’s rivals mainly mass-produce and sell general-purpose industrial computers. Conversely, HPC Systems employs high-spec designs customized to meet the requirements of its customers, sources the optimal parts to assemble its systems, while opting for a high-mix/flexible-volume approach toward manufacturing. It has established an organization to ensure long-term supply of its products to customers. Other companies that manufacture industrial computers include Contec Ltd. (TSE2: 6639) and Innotech Corporation (TSE1: 9880), but these do not qualify as genuine competitors as they mass-produce industrial computers as general-purpose products. HPC Systems has lower revenue than its competitors, but it boasts higher profit margins.
Order format
According to the company, exhibitions are one of the main channels through which it secures orders. While showcasing its products at such venues, it meets a large number of customers in search of cutting-edge technology. The company also frequently teams up in sales with several semiconductor trading houses, which act as agents. The rationale behind this approach is that customers that purchase semiconductors wholesale from trading houses also likely require industrial computers. Companies that serve as sales agents for HPC Systems include Ryoyo Electro Corporation (TSE1: 8068), Narasaki Sangyo Co., Ltd. (TSE2: 8085), and Daitron Co., Ltd. (TSE1: 7609).
New and ongoing projects
In the CTO segment, the company supplies industrial computers that are customized to meet the specifications requested by its customers. Because these computers are ultimately integrated in the products of its customers, the company receives ongoing orders as long as its customers continue to sell their products. As a result, the company concentrates on securing orders that will continue in subsequent years, rather than ad-hoc orders. It says it handled roughly equal shares of new and ongoing projects in FY06/21. GPM is generally lower for new projects, according to the company.
Sosa plant
The company operates its own plant for the CTO segment in Sosa City, Chiba Prefecture. At this plant, it assembles, manufactures, inspects, and ships products ordered by customers. The plant switched fully to renewable energy from September 2021.
Orders, revenue, and operating profit
Revenue in the CTO segment varies based on the computer equipment sold to customers. Revenue tends to accumulate over the medium term as the company increases the number of repeat customers, but new order conditions can cause swings in revenue. Because margins for repeat customers are generally higher than for new customers, an increase in the share of new customers tends to depress overall profitability.
Revenue and operating profit in CTO segment
FY06/17
FY06/18
FY06/19
FY06/20
FY06/21
(JPYmn)
Revenue
1,163
1,405
1,609
1,527
1,875
YoY
-
20.8%
14.6%
-5.1%
22.8%
Operating profit
92
143
158
179
183
YoY
-
55.3%
10.3%
13.2%
2.1%
Operating profit margin
7.9%
10.2%
9.8%
11.7%
9.8%
Segment assets
930
866
1,147
1,404
YoY
-
-6.9%
32.5%
22.4%
Segment ROA
-
17.6%
17.8%
14.3%
Source: Shared Research based on company data
Note: Figures may differ from company data due to differences in rounding methods.
Earnings structure
The company provides hardware and
software tailored to the needs of its customers in its HPC and CTO segments. As
this is its core business, its major cost items are materials costs under cost
of revenue and personnel expenses under SG&A expenses.
Cost of revenue
In FY06/21, materials costs accounted for 97.8% of manufacturing costs, labor costs for 1.2%, and other costs for 1.0%. The cost ratio tends to increase for tenders won from universities or public research institutions, and decline for projects handled for private companies. In the CTO segment, the cost ratio is typically high for new projects and low for ongoing projects.
Breakdown of cost of revenue
FY06/15
FY06/16
FY06/17
FY06/18
FY06/19
FY06/20
FY06/21
(JPYmn)
Cost of revenue
2,871
2,863
3,998
3,175
4,022
Cost of revenue ratio
73.6%
70.6%
74.1%
67.2%
69.0%
Manufacturing cots
3,039
3,917
3,078
4,052
Materials costs
2,970
3,848
2,997
3,961
Labor costs
43
43
49
50
Other costs
25
26
31
41
Materials costs (% of total)
97.8%
98.2%
97.4%
97.8%
Labor costs (% of total)
1.4%
1.1%
1.6%
1.2%
Other costs (% of total)
0.8%
0.7%
1.0%
1.0%
Work-in-process inventory at the beginning of year
66
169
164
82
Total
3,104
4,086
3,242
4,134
Work-in-process inventory at the end of year
169
164
82
106
Transfer to other accounts
46
34
37
58
Cost of manufacturing
2,890
3,887
3,122
3,970
Source: Shared Research based on company data
Note: Figures may differ from company data due to differences in rounding methods.
SG&A expenses
In FY06/21, personnel expenses made up 45.9% of SG&A expenses, depreciation 4.6%, and other expenses 49.5%.
Breakdown of SG&A expenses
FY06/15
FY06/16
FY06/17
FY06/18
FY06/19
FY06/20
FY06/21
(JPYmn)
SG&A expenses
786
908
1,028
1,073
1,129
SG&A ratio
20.1%
22.4%
19.1%
22.7%
19.4%
Salaries and allowances
359
358
365
386
Provision for employee bonuses
59
86
69
96
Provision for directors' bonuses
11
31
19
36
Depreciation
48
51
44
52
Other
430
502
575
559
No. of employees
62
63
74
81
84
85
92
No. of temporary employees (average)
11
12
11
12
14
15
14
Total
73
75
85
93
98
100
106
Personnel expenses per employee
5.32
5.43
5.08
5.52
Average age
43.8
44.2
44.5
Average years of service
7.8
8.5
8.2
Average annual salary
5.48
5.73
5.63
Source: Shared Research based on company data
Note: Figures may differ from company data due to differences in rounding methods.
Note: Personnel expenses per employee include labor expenses under cost of revenue.
Seasonality in earnings
As the company supplies its products and
services to universities, public research institutions, and major companies, it
tends to receive the bulk of its orders at the end of March. As a result, its
revenue and profit are skewed toward January–March. In
FY06/21, the quarterly shares of revenue and operating profit were 14.4% and
1.3% for Q1, 25.5% and 28.0% for Q2, 46.8% and 75.0% for Q3, and 13.3% and
-4.3% for Q4.
Global supercomputer and high performance-computing
markets
According to Hyperion Research, the
global supercomputer market was worth USD5.9bn in 2020. The launch of Fujitsu’s Fugaku
supercomputer created a new category for cutting-edge supercomputers. Hyperion
Research sees the market growing at a CAGR of 10.8% in the five years from 2019
to 2024. It valued the peripheral high performance-computing (HPC) market at
USD27.8bn in 2020, and expects this market to expand at a CAGR of 6.6% in the
five years from 2019 to 2024.
Global supercomputer market
(USDmn)
2019
2020
2021
2022
2023
2024
Cutting-edge
0
1,065
1,000
2,150
2,200
2,450
Large (USD3mn or more)
3,937
3,670
3,748
3,901
4,380
4,605
Entry-level (USD500,000–3mn)
1,287
1,204
1,214
1,347
1,515
1,678
Total
5,224
5,939
5,962
7,398
8,095
8,733
Source: Hyperion Research, June 2021
Global HPC market
(USDmn)
2019
2020
2021
2022
2023
2024
Servers
13,595
13,744
13,712
16,162
17,670
18,846
Storage systems
5,379
5,520
5,586
6,661
7,465
8,021
Middleware
1,599
1,618
1,636
1,943
2,138
2,294
Applications
4,647
4,682
4,629
5,371
5,774
6,049
Services
2,218
2,186
2,123
2,416
2,548
2,617
Total
27,438
27,750
27,686
32,553
35,595
37,827
Source: Hyperion Research, March 2021
Japanese supercomputer and HPC markets
The sizes of the Japanese supercomputer and HPC markets are unknown. The top 20 supercomputers by performance operating in Japan are shown in the table below.
Supercomputers operating in Japan (December 2020)
Organization
Department
System
Model
Peak performance
Launched
(TFLOPS)
1
Institute of Physical and Chemical Research (RIKEN)
Center for Computational Science
FUGAKU
FUGAKU
537,212
2020
2
National Institute of Advanced Industrial Science and Technology
AI Bridging Cloud Infrastructure
ABCI
FUJITSU PRIMERGY
34,200
2018
3
Joint Center for Advanced High Performance Computing
Oakforest PACS
FUJITSU
25,001
2016
4
Japan Aerospace Exploration Agency
Engineering Digital Innovation center
JSS3/TOKI-SORA
PRIMEHPC FX1000
19,464
2020
5
Japan Meteorological Agency
Forecast Department
Supercomputer
Cray XC50
18,000
2018
6
Japan Atomic Energy Agency
Center for Computational Science and e-Systems
Supercomputer System
HPE SGI 8600
12,600
2020
7
Tokyo Institute of Technology
Global Scientific Information and Computing Center
TSUBAME 3.0
HPE SGI ICE XA
12,150
2017
8
National Institute of Natural Sciences
National Institute for Fusion Science
Plasma Simulator System
NEC SX-Aurora TSUBASA
10,510
2020
9
Kyushu University
Research Institute for Information Technology
Supercomputer System ITO
FUJITSU PRIMERGY CX
10,430
2017
10
Nagoya University
Information Technology Center
Furo Type I
FUJITSU FX1000
7,785
2020
11
Nagoya University
Information Technology Center
Furo Type II
FUJITSU PRIMERGY CX2570 M5
7,519
2020
12
The University of Tokyo
Institute for Solid State Physics
System B
Dell PowerEdge C6525/R940
6,881
2020
13
Kyoto University
Academic Center for Computing and Media Studies
System A Camphor 2
Cray XC40
5,480
2016
14
SAKURA internet Inc.
Ishikari Data Center
HPE Apollo 6500
5,365
2019
15
National Institute for Materials Science
Materials Numerical Simulator
HPE SGI 8600
4,971
2020
16
National Institutes for Quantum Science and Technology
Rokkasho Nuclear Fusion Laboratory
JFRS-1
HPE Cray XC50
4,190
2018
17
Preferred Networks, Inc.
MN-3
3,137
2020
18
Hokkaido University
Information Technology Center
Grand Chariot Subsystem A
FUJITSU PRIMERGY SERVER
3,080
2018
19
Tohoku University
Center for Computational Materials Science, Institute for Materials Research
MASAMUNE-IMR
Cray XC50-LC / CS-Strom 500GT
3,030
2018
20
Japan Meteorological Agency
Meteorological Research Institute
Supercomputer System
FUJITSU PRIMERGY CX2550M4
2,816
2020
Source: Shared research based on data compiled by HPC Wire Japan
Note: Tera Floating-Point Operations Per Second (TFLOPS) is a unit of computer processing performance equal to one trillion (1012) floating-point operations per second.
Based on a survey conducted by HPC Systems, the target market for its HPC segment was worth JPY260.0bn in 2021, and the company sees this market expanding at a CAGR of 13.0% to JPY380.0bn in 2024.
Science, Technology, and Innovation Basic Plan
In March 2021, the Japanese Cabinet approved and set in motion the 6th Science, Technology, and Innovation Basic Plan that maps out the science and technology policy of the government for the five years from FY2021. The government budgets JPY30tn in R&D spending over the five-year period, and hopes this will spur private investment. On the whole, it targets public-private investment of JPY120tn. The plan calls for a “societal vision predicated on innovation (societal change)” as its first theme, and promotes digitalization and the implementation of data-driven strategies in the government, as well as the preparation and development of next-generation infrastructure and technology such as supercomputers and quantum technology. In addition, the government will establish a JPY10tn university fund, and set up a research system for universities that can compete on the global stage.
Global industrial computer and CTO markets
IHS Markit valued the global industrial computer market at USD4.1bn in 2020, and estimates this market expanded at a CAGR of 4.7% in the seven years from 2013 to 2020. In 2021 and 2022, it forecasts YoY growth of 2.3% and 3.6%, respectively.
Global industrial computer market
(USDmn)
2013
2014
2015
2016
2017
2018
2019
2020
2021F
2022F
Global industrial computer market
2,937
3,207
3,343
3,538
3,501
3,890
4,009
4,050
4,143
4,291
YoY
9.2%
4.2%
5.8%
-1.0%
11.1%
3.1%
1.0%
2.3%
3.6%
Source: IHS Markit
Japanese industrial computer and CTO markets
According to “Industrial PCs
and their Expansion in the Edge Computing Field: Conditions and Outlook
(FY2019)” published by the MIC Research Institute, the industrial computer
market was worth JPY109.3bn in FY2018. MIC Research Institute expects the
market to expand at a CAGR of 5.7% from FY2018 to FY2022, and reach JPY136.1bn
in FY2022.
Based on its own survey, HPC Systems values the target market for its CTO segment at JPY187.0bn in 2021, and it sees this market expanding at a CAGR of 11.1% to JPY257.0bn in 2024.
Strengths and weaknesses
Strengths
Expertise in supercomputers and scientific computing software for quantum chemical computing, molecular dynamics, and other fields in the HPC segment
HPC Systems uses supercomputers to create operating environments for scientific computing, and this requires specialized expertise according to the company. Its systems consist of interconnected supercomputers, workstations, and servers. These differ from consumer PCs in that getting them to operate requires more than simply establishing the necessary connections and installing the appropriate software. The hardware, operating system (OS), middleware, software, and network that make up these systems must be optimized to ensure compatibility, enable parallel processing of multiple computational tasks, and reduce overall processing time. HPC Systems has accumulated extensive expertise in computing systems for the life science and material science fields. It is one of only a handful of Japanese companies that offer the Gaussian software package that has become the de-facto industry standard for quantum chemistry computing. It also develops and sells proprietary extended functionality for the software.
Ability to rapidly deliver cutting-edge products at affordable prices and secure orders mainly through customer inquiries in the HPC segment
Excluding tenders invitations by universities and public research institutions, the company mainly secures orders through direct customer inquiries. Customers typically refer to the product lineup and benchmarks (performance indicators) on the company’s website while they are researching the optimal hardware and software for their R&D objectives, and subsequently reach out to the company. Upon receiving an inquiry, the company’s sales staff provide explanations and propose products. Customers are primarily concerned with benchmarks and pricing, and HPC Systems lists benchmarks that it has independently verified on its website. The company has developed strong ties with Super Micro Computer and serves as a distributor for the latter. Taking advantage of this relationship, it has built an organization that facilitates rapid delivery of high-spec products at affordable prices, and Shared Research understands it has gained the recognition of its customers in this respect.
Capacity to produce and sell high-spec industrial computers with custom designs using a high-mix/low-volume manufacturing approach in the CTO segment
HPC Systems produces and supplies industrial computers with high specifications and custom designs in its CTO segment with a focus on high-mix/low-volume manufacturing. This approach differs from its competitors, which mainly mass-produce and sell low-end, general-purpose products. In addition, the company has established an organization to ensure long-term supply of its products to customers, whereas its rivals update their product models every few years. As a result, its repeat purchase rate exceeds 90%. Although it generates lower revenue than its competitors, the company boasts higher margins.
Weaknesses
Susceptible to direct impact from disruptions at suppliers of semiconductors and peripheral electronic parts
HPC Systems meticulously selects and procures hardware in function of the budgets and desired research environments of its customers. However, semiconductors have been in short supply since the autumn of 2020, and related shortages affect the hardware procured by the company. That said, the company did not suffer lost orders or delayed deliveries in FY06/21, and thus managed to avert direct impact. While it hedges against potential adverse effects by placing orders in advance, there is a risk the company may not be able to procure sufficient volume, in which case its earnings could take a direct hit.
Small operational scale hinders the company’s ability to accept large projects in the HPC segment
The order value in the HPC segment ranges from several million yen to over JPY400mn per project, but the order value for large projects in the HPC market can exceed JPY1.0bn. Shared Research understands that such projects are almost exclusively handled by major electronics manufacturers. HPC Systems’ workforce and revenue scale make it difficult for the company to pursue such orders even though it possesses the necessary technical expertise.
Difficulties in generating economies of scale in the CTO segment due to a production structure geared toward high-mix/low-volume manufacturing
HPC Systems ensures long-term stable supply of high-end computers in its CTO segment, and has achieved a repeat purchase rate of over 90% as a result. It designs industrial computers tailored to customer preferences, procures the optimal parts, and assembles these using a high-mix/low-volume manufacturing approach. Its custom designs and manufacturing process are two major differentiators for the company. However, this also means its order growth is dependent on increases in product sales at existing customers or, in the absence of this, the acquisition of new customers. In other words, the current production structure of the CTO segment is not conducive to economies of scale. In contrast, its rivals in this space mass-produce and sell general-purpose industrial computers.
Historical results and financial statements
Income statement
Income statement
FY06/17
FY06/18
FY06/19
FY06/20
FY06/21
(JPYmn)
Non-cons.
Non-cons.
Non-cons.
Non-cons.
Non-cons.
Revenue
3,901
4,053
5,396
4,725
5,828
YoY
34.7%
3.9%
33.1%
-12.4%
23.3%
Cost of revenue
2,871
2,863
3,998
3,175
4,022
Cost of revenue ratio
73.6%
70.6%
74.1%
67.2%
69.0%
Gross profit
1,030
1,190
1,398
1,551
1,806
YoY
-
15.5%
17.5%
10.9%
16.5%
Gross profit margin
26.4%
29.4%
25.9%
32.8%
31.0%
SG&A expenses
786
908
1,028
1,073
1,129
SG&A ratio
20.1%
22.4%
19.1%
22.7%
19.4%
Operating profit
244
282
370
478
677
YoY
-
15.6%
30.9%
29.3%
41.6%
Operating profit margin
6.3%
7.0%
6.8%
10.1%
11.6%
Non-operating income
15
14
3
2
1
Non-operating expenses
5
5
6
14
12
Recurring profit
254
292
367
465
666
YoY
238.3%
14.8%
25.8%
26.8%
43.1%
Recurring profit margin
6.5%
7.2%
6.8%
9.8%
11.4%
Extraordinary gains
7
14
0
0
1
Extraordinary losses
1
-
-
1
0
Income taxes
98
116
148
158
219
Implied tax rate
37.5%
37.9%
40.2%
33.9%
32.9%
Net income
163
190
219
307
447
YoY
317.9%
16.5%
15.6%
40.1%
45.4%
Net margin
4.2%
4.7%
4.1%
6.5%
7.7%
Source: Shared Research based on company data
Note: Figures may differ from company data due to differences in rounding methods.
In FY06/21, revenue and recurring profit increased to record highs of JPY5.8bn and JPY666mn, respectively. In the four years from FY06/17 to FY06/21, revenue expanded at a CAGR of 10.6% and recurring profit at a CAGR of 27.2%.
In FY06/17 and FY06/19, the company recorded revenue for large
projects worth over JPY400mn. As a result, revenue either expanded at a slower
pace or declined in the subsequent years—namely, FY06/18 and FY06/20.
The cost of revenue ratio is around 70%, and typically rises in
tandem with higher revenue ratios of large projects, tenders for universities
or public research institutions in the HPC segment, or a higher revenue ratio
of new projects in the CTO segment.
The SG&A ratio is around 20%. The main cost items under SG&A expenses are fixed costs such as personnel expenses and depreciation. However, sales growth requires personnel increases, and therefore inevitably drives up SG&A expenses.
There are no noteworthy items related to non-operating income and extraordinary expenses.
Executive summary
Business overview
HPC Systems Inc. operates in two segments: High Performance Computing (HPC; 67.8% of total revenue and 73.0% of total operating profit in FY06/21) and Configure to Order (CTO; 32.2%, 27.0%). In FY06/21, it had revenue of JPY5.8bn (+23.2% YoY) and operating profit of JPY677mn (+41.6% YoY).
In the HPC segment, the company provides one-stop services for high-performance computing (HPC). These include sales of HPC systems for scientific computing and simulation software; systems integration services (i.e., building scientific computing, artificial intelligence [AI], and deep-learning environments); services to parallelize and speed up simulation software programs; development and sales of computational chemistry and materials informatics software; contracted computing services and support for scientific R&D; and science cloud services (Software as a Service for drug discovery, materials R&D, etc.).
In the CTO segment, HPC Systems supplies customized industrial and edge computers that meet customer applications and challenges as determined through interviews. It proposes specifications for such computers, and develops, manufactures, and maintains them, ensuring long-term stable supply of products. Industrial and edge computers are used as controlling equipment in plant production facilities, manufacturing equipment, inspection equipment, control devices, transportation infrastructure, automated driving applications, and retail stores. The company customizes these products in accordance with customer requirements such as pricing, performance, quality levels, operation in harsh environments, and resistance to low or high temperatures, dust, water, and static electricity.
In the HPC segment, the company procures high-performance computer hardware from major electronics manufacturers, and configures these to match the operating requirements of its customers. It builds small to large computer clusters that connect workstations, servers, and storage devices through high-speed networking equipment, installs the necessary software, and optimizes these systems as supercomputers. The hardware and software that make up these systems differ from consumer PCs in that getting them to operate requires more than simply setting up the necessary connections and installing the appropriate software. The hardware, operating system (OS), middleware, software, and network must be configured and optimized to run multiple computational processes in parallel while reducing processing time, and this requires the expertise of a company like HPC Systems.
Scientific computing is used in a range of fields such as elementary particle and astronomical research, weather forecasting, global warming projections, predictions of disasters such as earthquakes and tsunamis, drug discovery, development of semiconductors and battery materials, and structural and aerodynamic analysis of vehicles and aircraft. HPC Systems excels in building systems for the life science and material science fields, and it mainly supplies these to universities, public research institutions, and R&D divisions of private companies. The process to develop new drugs or materials involves synthesizing various compounds as potential candidates for new drugs or materials, and verifying whether these deliver the desired effect. Supercomputers can simulate combinations of countless different materials, allowing researchers to predict the attributes of new materials before they produce them. This contributes to more effective and efficient development and verification.
Other companies that operate in the systems integration field are HPC Solutions Co., Ltd. (unlisted), HPC Tech Corporation (unlisted), and Visual Technology, Inc. (unlisted). However, HPC Systems does not actually compete with these companies for large projects or in the computational chemistry field. The target market for the HPC segment is worth JPY260.0bn (based on a 2021 company survey), and the segment recorded revenue of JPY4.0bn (+23.6% YoY) in FY06/21.
The HPC segment generates a mix of one-time revenue and recurring revenue, with the former accounting for over 90% of the total revenue. One-time revenue is determined by order value and the number of orders, and the lead time for orders is roughly three months. Order value is several to 10 million yen for small orders, goes up to JPY50mn for mid-sized orders, and reaches or exceeds JPY400mn for large orders. Recurring revenue is monthly revenue generated from the company’s science cloud services, and varies based on the type of services requested by customers.
The target market for the CTO segment is worth JPY178.0bn (based on a 2021 company survey), and the segment recorded revenue of JPY1.9bn (+22.8% YoY) in FY06/21. Using a high-mix/low-volume manufacturing approach, HPC Systems prepares custom designs for industrial computers in accordance with customer requirements, and procures and assembles the optimal parts to build these products. In contrast, its rivals mass-produce and sell general-purpose industrial computers. HPC Systems has also established an organization that enables it to supply its products to customers over the long term. The company’s peers include Contec Ltd. (TSE2: 6639) and Innotech Corporation (TSE1: 9880), but it do not actually compete with these companies, which mass-produce general-purpose industrial computers. HPC Systems has lower revenue than its peers, but it boasts higher profit margins.
In the CTO segment, HPC Systems mainly supplies computers for inspection equipment (e.g., inspection and monitoring of products in plants), Internet of Things (IoT) computers used in plants (e.g., network-based management of cameras and monitoring in plants), and high-end industrial computers for medical image processing (e.g., computed tomography [CT] scans and magnetic resonance imaging [MRI]). Customers often place repeat orders once they have integrated a computer supplied by the company into one of their products, and this trend is reflected in a repeat purchase rate of over 90%.
In both HPC and CTO segments, HPC Systems provides products and services tailored to customer needs (i.e., it procures hardware/software, assembles these, and ships the final products). Accordingly, variable (materials) costs accounted for 97.8% of cost of revenue in FY06/21 (cost of revenue ratio: 69.0%). Its SG&A ratio stood at 19.4%, mostly comprising fixed costs such as personnel expenses and depreciation.
Earnings trends
In FY06/21, revenue was JPY5.8bn (+23.3% YoY), operating profit JPY677mn (+41.6% YoY), recurring profit JPY666mn (+43.1% YoY), and net income JPY447mn (+45.4% YoY). The results reflected contributions from large orders received from universities and public research institutions. The SG&A expense ratio improved 3.3pp YoY as revenue growth outpaced the increase in SG&A expenses, and OPM improved 1.5pp YoY to 11.6%.
In FY06/22, the company forecasts revenue of JPY6.4bn (+10.2% YoY), operating profit of JPY745mn (+10.2% YoY), recurring profit of JPY741mn (+11.3% YoY), and net income of JPY496mn (+11.0% YoY). HPC Systems thinks the business environment is hard to read at present because it remains unclear when the COVID-19 pandemic will subside and give way to an economic recovery, and also due to impact from semiconductor shortages.
In August 2021, the company unveiled its medium-term management plan “Vision 2024” (FY06/22–FY06/24). As numerical targets, the plan calls for revenue of JPY10.0bn (three-year CAGR of 19.7%), operating profit of JPY1.3bn (23.8%), and ROE of 26% (28.1% in FY06/21). By segment, it looks for revenue of JPY7.0bn (three-year CAGR of 21.0%) and operating profit of
Strengths and weaknesses
Shared Research regards HPC System’s strengths as: (1) its expertise in supercomputers and scientific computing software for quantum chemical computing, molecular dynamics, and other fields in the HPC segment , (2) its ability to rapidly deliver cutting-edge products at affordable prices and secure orders mainly through customer requests in the HPC segment, and (3) its capacity to sell and produce high-performance industrial computers with custom designs tailored to the needs of its customers using a high-mix/low-volume manufacturing approach in the CTO segment. We see its weaknesses as (1) a susceptibility to direct impact of disruptions at suppliers of semiconductors and peripheral electronic parts, (2) a small operational scale that hinders the company’s ability to accept large projects in the HPC segment, and (3) difficulties in generating economies of scale in the CTO segment due to a high-mix/low-volume manufacturing structure. For details, see the Strengths and weaknesses section.
Key financial data
Note: Figures may differ from company data due to differences in rounding methods.
Recent updates
Capital and business alliance with QunaSys Inc.
HPC Systems Inc. announced an agreement to form a capital and business alliance with QunaSys Inc. (unlisted), with a view to accelerating business development in computational quantum chemistry.
In July 2020, HPC Systems entered into a business alliance with QunaSys, to pursue development in the computational quantum chemistry field. The aim of this business alliance was to fuse HPC Systems' advanced technologies and solutions in computational chemistry with the cutting-edge technologies of QunaSys, a developer of quantum computing applications, to develop the world’s first computational chemistry solution for quantum computers and enhance both companies' enterprise value in the process.
The two companies have been engaged in R&D and business development as part of this alliance. As these activities have generated synergies in both technology and business, HPC Systems has agreed to enter into a capital and business alliance with QunaSys by way of acting as an underwriter for the latter's Series B third-party allotment of new shares (investment amount and stake undisclosed).
The parties believe that this capital and business alliance will further strengthen their collaborative relationship and accelerate business growth toward the practical application of quantum computing, expected within a few years.
Trends and outlook
Quarterly trends and results
Note: Figures may differ from company data due to differences in rounding methods.
Cumulative Q3 FY06/22 results (out May 13, 2022)
The company made no change to its full-year forecast for FY06/22. The progress versus the full-year forecast was 77.7% for revenue (versus 86.7% in cumulative Q3 FY06/21 vs. full-year result), 85.7% for operating profit (104.3%), 83.9% for recurring profit (104.6%), and 87.0% for net income (107.1%) .
In cumulative Q3 FY06/22, revenue was JPY5.0bn (-1.2% YoY), breaking down into JPY3.6bn for the HPC segment (-2.2% YoY) and JPY1.4bn for the CTO segment (+1.4% YoY). The cost of revenue ratio was 68.8%, down 0.9pp YoY due to the recording of revenue from high-margin projects, which offset higher material prices. The SG&A ratio was 18.4%, up 2.1pp YoY on higher personnel expenses stemming from aggressive hiring in line with the medium-term management plan. Operating profit was JPY638mn (-9.6% YoY) and the OPM was 12.8% (-1.2pp YoY).
HPC segment
Orders declined YoY in Q3 (three months). Orders from universities and government offices were down YoY due to the absence of extraordinary demand seen in the previous year. In the private sector, orders increased in the computational chemistry solutions (CCS) and computer aided engineering solution (CAES) areas, but decreased in the data science solution (DSS) area.
The company recorded revenue in Q3 from large orders which it received in Q4 FY06/21. Revenue from the private sector increased 59% YoY, backed by strong performance of DDS projects for AI and deep learning development platforms and CCS projects for materials development and drug discovery research. On the other hand, revenue from universities and government offices declined 48% YoY in reaction to the extraordinary demand seen in the previous fiscal year.
The company was able to sell its science cloud system to overseas customers for the first time in Q1. HPC Systems has positioned overseas expansion as a key initiative in its medium-term plan (See "(5) Advance into overseas markets and promote related operation” in the Medium- to long-term outlook section).
OPM declined 1.7pp YoY to 13.6%. This was attributable to an increase in SG&A expenses on a higher headcount (another key policy of the medium-term plan, "(2) Cultivate human capital and recruit human resources").
CTO segment
Orders were up 1.5x YoY in Q3 (three months). Orders from repeat customers increased for SF machine vision projects for semiconductor-related equipment (wafer inspection, etc.) and Edge IoT projects. Orders from new customers were robust for Edge IoT and Edge AI projects.
Revenue from repeat customers increased YoY. Major contributions came from Edge IoT projects for medical equipment and amusement equipment, as well as from SF machine vision projects for semiconductor-related equipment (wafer inspection, etc.). Revenue from new customers declined due to the absence of large orders received in the previous year.
OPM improved 0.1pp YoY to 10.7%. Although personnel expenses rose on a higher headcount (a key policy of the medium-term plan) and depreciation expenses increased due to the upgrade of testing and inspection equipment, these were more than offset by the recording of revenue from high-margin projects.
For details on previous quarterly and annual results, please refer to the Historical results section.
Company forecast for FY06/22
Note: Figures may differ from company data due to differences in rounding methods.
For FY06/22, HPC Systems forecasts revenue of JPY6.4bn (+10.2% YoY), operating profit of JPY745mn (+10.1% YoY), recurring profit of JPY741mn (+11.3% YoY), and net income of JPY496mn (+10.9% YoY).
The company has compiled conservative forecasts for FY06/22 as it expects shortages of semiconductors and peripheral electronic components to cause disruptions in its supply chain.
HPC segment
HPC Systems forecasts revenue of JPY4.3bn (+9.4% YoY) in FY06/22. Following a rise in revenue from universities and public research institutions driven by the government’s supplementary budget in FY06/21, the company expects demand to revert to normal levels in FY06/22. Demand from private companies is rebounding across the board, and demand for drug discovery, materials development, and AI-related applications has been particularly strong. The company also secured a large order related to machine learning in Q4 FY06/21, and plans to record related revenue in FY06/22.
CTO segment
The company forecasts revenue of JPY2.1bn (+12.0% YoY) in FY06/22. It expects a recovery in demand for medical equipment and amusement equipment applications after a slump in demand attributable to the COVID-19 pandemic in FY06/21. In addition, demand for machine vision and inspection equipment applications also remains strong.
Company forecasts versus results
The table below shows how HPC Systems’ results have compared with its forecasts. The company revised up its FY06/21 forecast when it announced Q3 FY06/21 results.
Note: Figures may differ from company data due to differences in rounding methods.
Medium- to long-term outlook
HPC Systems aspires to be a “global visionary company of the 21st century that specializes in the field of computer science” with the capacity to endure through several eras and continue operating beyond 2030. As a medium-term goal, it aims to become the leader in a global niche market by 2030. It has formulated the medium-term management plan “Vision 2024” (FY06/22–FY06/24) using a backcasting approach, and unveiled the plan in August 2021.
As numerical targets, the plan calls for FY06/24 revenue of JPY10.0bn (three-year CAGR of 19.7%), operating profit of JPY1.3bn (23.8%), and ROE of 26% (28.1% in FY06/21). The breakdown by segment is revenue of JPY7.0bn (three-year CAGR of 21.0%) and operating profit of JPY960mn (24.9%) in the HPC segment, and revenue of JPY3.0bn (17.0%) and operating profit of JPY320mn (20.9%) in the CTO segment. In addition, the plan looks to keep ROE above 25% through FY06/24. It outlines specific priority measures for each segment, which are discussed in detail below.
Based on its own survey, HPC Systems estimates the target market for the HPC segment will expand at a three-year CAGR of 13.0% from JPY260.0bn in 2021 to JPY380.0bn in 2024, and aims for above-market growth of 21.0% in the segment. Similarly, it estimates the target market for the CTO segment will grow at a CAGR of 11.1% from JPY187.0bn in 2021 to JPY257.0bn in 2024, and aims for above-market growth of 17.0% in the segment.
Based on its own survey, HPC Systems estimates the target market for the HPC segment will expand at a three-year CAGR of 13.0% from JPY260.0bn in 2021 to JPY380.0bn in 2024, and aims for above-market growth of 21.0% in the segment. Similarly, it estimates the target market for the CTO segment will grow at a CAGR of 11.1% from JPY187.0bn in 2021 to JPY257.0bn in 2024, and aims for above-market growth of 17.0% in the segment.
As priority measures, the company will (1) evolve its core businesses, (2) cultivate human capital and recruit human resources, (3) create new businesses, (4) strengthen management foundations, and (5) advance into overseas markets and promote related operations.
(1) Evolve core businesses
Priority measures in the HPC segment
(1) Develop practical applications for computational chemistry and materials informatics
The company aims to increase revenue in this field from JPY400mn in FY06/21 to JPY1.0bn in FY06/24. In addition to the computational chemistry software it already provides, it plans to promote the adoption and expansion of materials informatics and the Global Reaction Route Mapping (GRRM) program it has jointly developed with Hokkaido University.
(2) Standardize high-performance computing (HPC) systems integration
Integrate heterogenous hardware environments and diverse software in HPC-AI platforms while considering future strategies. As a KPI, expand the new adoption rate for HPC-AI platforms to 50%.
(3) Diversify science cloud services
Under its science cloud services, HPC Systems provides various types of scientific computing software to customers over the cloud. The company will aim to convert 20 or more new customers into repeat customers for its cloud services.
Priority measures in the CTO segment
HPC Systems aims to expand revenue from repeat customers in the CTO segment by 130% from FY06/21 to FY06/24. It believes it will need to develop growth drivers and strengthen its business foundations to achieve this target.
The company envisions the following growth drivers: (1) cultivating industry knowledge, (2) accelerating the development of customers through combination technologies, and (3) applying network technologies. It will aim to understand the challenges their customers face on the ground as well as their core needs, and supply the optimal products customers require through its proprietary combination technologies.
To strengthen its business foundations, HPC Systems plans to (1) upgrade its plants, (2) forge ahead with digital transformation (DX: adopting of digital technologies to optimize operations) and reform its business processes, and (3) improve its production technology and quality-cost-delivery (QCD). To upgrade its plants, the company looks for an 50% increase its production capacity by 2024 by making more efficient use of its current space. It will also consider expanding existing plants and building new facilities. To forge ahead with DX and reform its business processes, the company will promote the sharing and standardization of operations, diversify materials procurement channels, and promote initiatives related to chemSHERPA.
(2) Cultivate human capital and recruit human resources
The company plans to increase its human resources by 20–50% from FY06/21. To this end, it will actively recruit personnel in specialized fields, provide internal training, and expand skill enhancement programs.
(3) Create new businesses
The company looks to create new businesses that can become fresh earnings drivers. Specifically, it plans to leverage synergies between its core businesses to create businesses in the fields of machine learning and AI. At the same time, it will work to expand and promote business alliances.
(4) Strengthen management foundations
The company plans to promote DX, develop business tools, and adopt robotic process automation (RPA) to strengthen its management foundations. It will also develop and operate marketing tools.
(5) Advance into overseas markets and promote related operations
On the overseas front, the accompany plans to (1) move into the European and US markets by forming alliances with partners in the computational chemistry field, (2) reinforce its base in Vietnam, (3) strengthen the functions of its base in Taiwan, and (4) cultivate the Chinese market. In the HPC segment, it has handled orders for the Vietnamese government and Vietnam National University in collaboration with a major trading house. By FY06/24, it plans to increase overseas revenue to JPY500mn–JPY1.0bn.
Market environment
In March 2021, the Japanese Cabinet approved and set in motion the 6th Science, Technology, and Innovation Basic Plan that maps out the science and technology policy of the government for the five years from FY2021. The government budgets JPY30tn in R&D spending over the five-year period, and hopes this will spur private investment. On the whole, it targets public-private investment of JPY120tn.
The plan calls for a “societal vision predicated on innovation (societal change)” as its first theme, and promotes digitalization and the implementation of data-driven strategies in the government, as well as the preparation and development of next-generation infrastructure and technology such as supercomputers and quantum technology.
The government will establish a JPY10tn university fund, and establish a research system for universities that can compete on the global stage.
Business
Business overview
As its mission, HPC Systems seeks to “provide researchers and developers with the tools to research and develop products.”
The company operates in two segments: High Performance Computing (HPC; 67.8% of total revenue and 73.0% of total operating profit in FY06/21) and Configure to Order (CTO; 32.2%, 27.0%).
Note: Figures may differ from company data due to differences in rounding methods.
HPC segment
In this segment, HPC Systems develops, manufactures, and sells hardware and software for scientific computers. Its customers are mainly universities and R&D divisions of private companies. Scientific computers such as supercomputers and workstations differ from their consumer-oriented counterparts in that they deliver vastly higher performance. The aforementioned R&D divisions rely on computational science to develop new materials, discover new drugs, and conduct simulations of fluid structures, which are an essential part of product manufacturing. The company provides the environments (integrated hardware, operating system [OS], middleware, software, and network, and related optimization) to facilitate these processes.
In this segment, HPC Systems develops, manufactures, and sells hardware and software for scientific computers. Its customers are mainly universities and R&D divisions of private companies. Scientific computers such as supercomputers and workstations differ from their consumer-oriented counterparts in that they deliver vastly higher performance. The aforementioned R&D divisions rely on computational science to develop new materials, discover new drugs, and conduct simulations of fluid structures, which are an essential part of product manufacturing. The company provides the environments (integrated hardware, operating system [OS], middleware, software, and network, and related optimization) to facilitate these processes.
Core customers
The company has supplied products and services to private companies operating in the pharmaceuticals, chemicals, materials, automotive, electrical and electronic equipment, and heavy industry sectors. It also counts many public research institutions and universities among its customers.
From FY06/19 to FY06/21, the HPC segment accounted for 69% of total revenue. By strategic (customer) area, computer science solutions (CSS: fundamental research into science and technology) for universities and public research institutions made up 29%, data science solutions (DSS: machine learning and AI R&D) 21%, computer-aided engineering solutions (CAES: simulation data analysis during design phase) for automotive, general electrical machinery, optical equipment, precision equipment, and information and communication equipment companies 9%, computer chemistry solutions (CCS: drug discovery and materials development) for comprehensive chemicals, derivatives, electronic materials, drug discovery, and food companies 8%, and HPC infrastructure (HPCI) 2%. Over the same period, the CTO segment generated 31% of total revenue. By strategic (customer) area, Edge IoT for medical and amusement equipment accounted for 15%, smart factory solutions for manufacturing equipment and plant facilities (visual inspections and semiconductors) for 12%, and edge AI for infrastructure monitoring and security equipment for 4%.
Some examples of supercomputer tenders of universities or public research institutions that have been awarded to the company are shown below.
Examples of systems deployed by the company at customers are shown below.
Competitive landscape
Suppliers of supercomputers include major companies such as Fujitsu Limited (TSE1: 6702), NEC Corporation (TSE1: 6701), and Hewlett Packard Enterprise (NYSE: HPE). To deliver the operating environments required by its customers, HPC Systems builds computer clusters that connect workstations, servers, and storage devices through high-speed networking equipment, installs the necessary software, and optimizes the computer clusters as supercomputers. The hardware and software that make up these systems differ from consumer counterparts in that getting them to operate requires more than simply establishing the necessary connections and installing the appropriate software. The hardware, operating system (OS), middleware, software, and network must be configured and optimized to run multiple computational processes in parallel while reducing the processing time, and this requires the expertise of a company like HPC Systems.
The company identifies the following as its main areas of business.
Other companies that operate in the systems integration field are HPC Solutions Inc. (unlisted, consolidated subsidiary of Argo Graphics Inc. [TSE1: 7595]), HPC Tech Corporation (unlisted, consolidated subsidiary of AKIBA Holdings Co., Ltd. [TSE JASDAQ: 6840]), and Visual Technology, Inc. (unlisted). However, HPC Systems does not actually compete with these companies in the computational chemistry field. Scientific computing is used in a wide range of fields such as elementary particle and astronomical research, weather forecasting, global warming projections, predictions of disasters such as earthquakes and tsunamis, drug discovery, development of semiconductors and battery materials, and structural and aerodynamic analysis of vehicles and aircraft. HPC Systems excels in building scientific computing systems for the life science and material science fields, and its customers are mainly universities, public research institutions, and R&D divisions of private companies.
Hardware
HPC Systems procures and delivers hardware and software in accordance with the needs of its customers. It mainly uses products supplied by Super Micro Computer, Inc. (NASDAQ: SMCI), Hewlett-Packard Company, GIGA-BYTE Technology Co., Ltd. (TWSE: 2376, Taiwan), DataDirect Networks, Inc. (unlisted, US), and Graphcore Limited (unlisted, UK). It meticulously selects hardware in function of customer budgets and desired research environments. However, the company has particularly strong ties with Super Micro Computer, and therefore sources a lot of its hardware from this supplier.
Super Micro Computer
HPC Systems has concluded a distributor agreement with Super Micro Computer, which has become a core supplier of servers for the company. From FY09/18 to FY09/21, Mr. Tau Leng, currently the Senior Vice President of Technology at Super Micro Computer, served as an outside director at HPC Systems. Super Micro Computer’s plant in Taiwan automatically configures the specifications (CPU, memory, GPU, OS, software, etc.) for orders placed by the company using an install system developed by HPC Systems. As a result, the company does not have to open the products to install parts or software after they have arrived in Japan. Thanks to these arrangements and the discounts it receives on bulk orders, HPC Systems can rapidly provide products manufactured by Super Micro Computer at affordable prices.
Incidentally, according to a survey conducted by IDC Japan in March 2021, Fujitsu Ltd. led the domestic server market in 2020 with a share of 27.4%, followed by NEC Corporation at 17.8%, Hewlett-Packard Japan, Ltd. at 12.7%, Dell Technologies Japan Inc. at 11.2%, Hitachi, Ltd. at 6.0%, other companies at 13.6%, and ODM production at 11.3%. Super Micro Computer was included under ODM production in the survey. On its website, Super Micro Computer states it was the world’s fifth largest supplier of servers by shipments in 2019.
Super Micro Computer works with several distributors in Japan, HPC Systems being one of them.
Software
We now turn to the software offered by HPC Systems for scientific computing. The company handles a broad range of software packages in the field of quantum chemistry, proposing solutions tailored to the needs of its customers. One example of such a software package is Gaussian. Developed by Gaussian, Inc., this quantum chemistry computing application uses cutting-edge technology to model the electronic structure of various compounds. It is used by many researchers in the fields of chemistry, biochemistry, and physics, and has become the de-facto standard for research in these areas. In Japan, Gaussian is sold by HPC Systems, Hulinks Inc. (unlisted, consolidated subsidiary of Argo Graphics Inc. [TSE1: 7595]), Conflex Corporation (unlisted), and Molsis Inc. (unlisted).
In the quantum chemistry and molecular dynamics fields, HPC Systems sells software developed by its own engineers, many of whom hold relevant PhD or master’s degrees. Having used standard software packages such as Gaussian and AMBER during their academic training, these engineers have a firm grasp of the functions that are either missing or difficult to use in these products, and develop software that makes up for these flaws. For example, the company released the Reaction Plus Pro software in June 2018. This software relies on Gaussian for computational processes but extends its functionality.
In July 2020, HPC Systems concluded a joint research agreement with the Faculty of Science at Hokkaido University. The aim of the agreement is to jointly research global reaction route mapping (GRRM), and commercialize related outcomes in the science cloud services offered by the company.
In February 2021, the company successfully developed a groundbreaking molecular structure generation algorithm as part of its materials informatics program development, and applied for a corresponding patent.
Science cloud services
The company launched its science cloud services in June 2018. Because they are delivered over the Internet, these services allow customers to perform scientific computing on PCs, removing the need to purchase expensive computers or software. The science cloud services have been mainly adopted by chemical manufacturers to date.
In April 2021, the company also started providing a paid science cloud service that demonstrates computing processes running on the supercomputer Fugaku.
Process from order to delivery
It takes roughly three months for HPC Systems to deliver orders. Upon receiving an order from a customer, the company first orders the necessary parts, which take six weeks to two months to arrive. Thereafter, it installs the operating system (OS) and applications, and confirms the system operates as expected. Transactions with overseas hardware and software vendors such as Super Micro Computer are settled in dollars, so HPC Systems is susceptible to forex impact. However, because the lead time for orders is about three months, the company can often pass on forex-driven price increases to customers, so it does not shoulder major forex risk. The largest concern for the company as of September 2021 is the global shortage of semiconductors, as a result of which deliveries of parts could conceivably suffer delays.
Order format
While it also bids for tenders issued by universities and public research institutions, HPC Systems mainly secures orders via direct customer inquiries. Customers typically refer to the product lineup and benchmarks (performance indicators) on the company’s website while they are researching the optimal hardware and software for their R&D objectives, and subsequently reach out to the company. Upon receiving a request, the company’s sales and technical staff explain and propose products to the customer. Customers are chiefly concerned with the benchmarks of the systems they intend to purchase, and the company lists verified numerical benchmarks on its website. HPC Systems also sells products through third parties under agency agreements, but the revenue contributions from such arrangements are small because the company does not sell general-purpose products. In April 2020, the company formed a capital and business alliance with AS ONE Corporation (TSE1: 7476), and started offering its products and services on the AXEL e-commerce platform operated by AS ONE in October 2020.
Benchmarks (numerical performance indicators)
HPC Systems measures benchmarks for systems that use the latest hardware (including next-generation processors), runs various types of optimized simulation software, and publishes these on its website. It uses benchmark software to measure and evaluate the quantitative processing capacity of the hardware and software employed. Customers can browse these benchmarks on its website and inquire about them before placing an order. In FY06/21, the company budgeted JPY76mn for capital investment, and allocated JPY58mn toward investments in servers and other hardware used to obtain benchmarks.
System, software, and application engineers
At end-June 2021, HPC Systems employed 92 people, of which 26 engineers engaged in R&D and verification technology. Its engineer team includes nine individuals with PhD degrees, six with master’s degrees, and 15 postdoctoral researchers in fields such as computational chemistry, organic synthesis, theoretical chemistry, machine learning, flow simulation, and physics. The company’s engineers are all experienced users of the Gaussian software package that has become the de-facto standard software for quantum chemistry. Having used the software during their academic training, they have a firm grasp of its shortcomings (such as missing functions), and use this information to develop the many proprietary software packages that are sold by the company.
HPC Technology Development Center
The company is headquartered in Minato-ku, Tokyo, but it operates its HPC Technology Development Center in Chuo-ku, Tokyo. From the center, its systems and application engineers provide sales support, develop proprietary software, and conduct technical research for its science cloud services.
Overseas operations
In March 2020, HPC Systems established Intelligent Integration Company Limited, a wholly owned subsidiary, in Vietnam. In the past, the company has handled orders for the Vietnamese government and Vietnam National University in the HPC segment. Its customers in the CTO segment have production bases in Vietnam, so the company expects to win related orders in the future. It also plans to expand its operations in Asian markets other than Vietnam.
HPC Systems has yet to handle orders in the US or Europe, but it is planning a move into these markets. Specifically, it intends to launch businesses that take advantage of its science cloud services.
Technology trend (1) Supercomputer Fugaku
The supercomputer Fugaku developed by Fujitsu Ltd. and RIKEN is the world’s most powerful supercomputer. As of September 2021, nine joint research projects were underway to expand the use and improve the usability of Fugaku, and HPC Systems was selected for one of these project. Since April 2021, HPC Systems has offered a paid cloud-based service that demonstrates its computational chemistry simulation software running on the supercomputer Fugaku.
Technology trend (2): Immersion-cooled servers
The kukai server delivered by HPC Systems to Yahoo Japan Corporation uses a liquid immersion cooling system, and has been ranked as the world’s second-best energy-saving supercomputer by GREEN500, a biannual ranking of the most energy-efficient supercomputers in the world. Servers are generally air-cooled, which means fans are used to divert the heat generated by their internal parts. The immersion-cooled server delivered by the company is entirely submerged in a thermally conductive dielectric liquid or coolant contained within a water tank, which contributes to a reduction in power consumption. This is a field poised for growth in the future.
Technology trend (3): Quantum computers
Quantum computers are next-generation computers that can perform calculations at far greater speeds than supercomputers. The technology is still in the development phase, and it is said to be two decades away from commercialization. Quantum computers perform calculations based on technological principles that are drastically different from classical computers. In November 2020, HPC Systems formed a business alliance with QunaSys Inc. (unlisted). The company aims to fuse its advanced technologies and solutions in computational chemistry with the cutting-edge technologies of QunaSys, a developer of quantum computer applications, to develop the world’s first computational chemistry solution for quantum computers.
Orders, revenue, and operating profit
The company’s revenue remains in an overall growth trend from a medium-term perspective, although it has declined in some years. The main factors driving fluctuations in revenue are the acquisition of large orders and the timing of revenue recognition.
In FY06/17 and FY06/19, HPC Systems handled major projects for Yahoo Japan Corporation worth over JPY400mn. OPM has improved mainly due to (1) a reduction in SG&A expenses (attributable to a relatively steady workforce) amid revenue expansion, and (2) the absence of impact from major projects with low margins.
Note: Figures may differ from company data due to differences in rounding methods.
CTO segment
In the CTO segment, HPC Systems develops, manufactures, and sells industrial computers that are configured to order. It operates a plant in Sosa City, Chiba Prefecture, from where it supplies products in accordance with customer needs using a high-mix/flexible-volume approach. Its product lineup includes inspection equipment computers that inspect and monitor products manufactured in plants, IoT computers used in plants to manage and monitor networks through on-site cameras, and high-spec industrial computers used for medical image processing (computed tomography [CT] scans and magnetic resonance imaging [MRI]). The company procures parts, cases that have been designed internally, and software, and subsequently develops and assembles these into systems that are supplied to customers under the HPC Systems brand. It says its repeat purchase rate exceeds 90%.
Competitors
The company’s rivals mainly mass-produce and sell general-purpose industrial computers. Conversely, HPC Systems employs high-spec designs customized to meet the requirements of its customers, sources the optimal parts to assemble its systems, while opting for a high-mix/flexible-volume approach toward manufacturing. It has established an organization to ensure long-term supply of its products to customers. Other companies that manufacture industrial computers include Contec Ltd. (TSE2: 6639) and Innotech Corporation (TSE1: 9880), but these do not qualify as genuine competitors as they mass-produce industrial computers as general-purpose products. HPC Systems has lower revenue than its competitors, but it boasts higher profit margins.
Order format
According to the company, exhibitions are one of the main channels through which it secures orders. While showcasing its products at such venues, it meets a large number of customers in search of cutting-edge technology. The company also frequently teams up in sales with several semiconductor trading houses, which act as agents. The rationale behind this approach is that customers that purchase semiconductors wholesale from trading houses also likely require industrial computers. Companies that serve as sales agents for HPC Systems include Ryoyo Electro Corporation (TSE1: 8068), Narasaki Sangyo Co., Ltd. (TSE2: 8085), and Daitron Co., Ltd. (TSE1: 7609).
New and ongoing projects
In the CTO segment, the company supplies industrial computers that are customized to meet the specifications requested by its customers. Because these computers are ultimately integrated in the products of its customers, the company receives ongoing orders as long as its customers continue to sell their products. As a result, the company concentrates on securing orders that will continue in subsequent years, rather than ad-hoc orders. It says it handled roughly equal shares of new and ongoing projects in FY06/21. GPM is generally lower for new projects, according to the company.
Sosa plant
The company operates its own plant for the CTO segment in Sosa City, Chiba Prefecture. At this plant, it assembles, manufactures, inspects, and ships products ordered by customers. The plant switched fully to renewable energy from September 2021.
Orders, revenue, and operating profit
Revenue in the CTO segment varies based on the computer equipment sold to customers. Revenue tends to accumulate over the medium term as the company increases the number of repeat customers, but new order conditions can cause swings in revenue. Because margins for repeat customers are generally higher than for new customers, an increase in the share of new customers tends to depress overall profitability.
Note: Figures may differ from company data due to differences in rounding methods.
Earnings structure
The company provides hardware and software tailored to the needs of its customers in its HPC and CTO segments. As this is its core business, its major cost items are materials costs under cost of revenue and personnel expenses under SG&A expenses.
Cost of revenue
In FY06/21, materials costs accounted for 97.8% of manufacturing costs, labor costs for 1.2%, and other costs for 1.0%. The cost ratio tends to increase for tenders won from universities or public research institutions, and decline for projects handled for private companies. In the CTO segment, the cost ratio is typically high for new projects and low for ongoing projects.
Note: Figures may differ from company data due to differences in rounding methods.
SG&A expenses
In FY06/21, personnel expenses made up 45.9% of SG&A expenses, depreciation 4.6%, and other expenses 49.5%.
Note: Figures may differ from company data due to differences in rounding methods.
Note: Personnel expenses per employee include labor expenses under cost of revenue.
Seasonality in earnings
As the company supplies its products and services to universities, public research institutions, and major companies, it tends to receive the bulk of its orders at the end of March. As a result, its revenue and profit are skewed toward January–March. In FY06/21, the quarterly shares of revenue and operating profit were 14.4% and 1.3% for Q1, 25.5% and 28.0% for Q2, 46.8% and 75.0% for Q3, and 13.3% and -4.3% for Q4.
Market and value chain
Market overview
Global supercomputer and high performance-computing markets
According to Hyperion Research, the global supercomputer market was worth USD5.9bn in 2020. The launch of Fujitsu’s Fugaku supercomputer created a new category for cutting-edge supercomputers. Hyperion Research sees the market growing at a CAGR of 10.8% in the five years from 2019 to 2024. It valued the peripheral high performance-computing (HPC) market at USD27.8bn in 2020, and expects this market to expand at a CAGR of 6.6% in the five years from 2019 to 2024.
Japanese supercomputer and HPC markets
The sizes of the Japanese supercomputer and HPC markets are unknown. The top 20 supercomputers by performance operating in Japan are shown in the table below.
Note: Tera Floating-Point Operations Per Second (TFLOPS) is a unit of computer processing performance equal to one trillion (1012) floating-point operations per second.
Based on a survey conducted by HPC Systems, the target market for its HPC segment was worth JPY260.0bn in 2021, and the company sees this market expanding at a CAGR of 13.0% to JPY380.0bn in 2024.
Science, Technology, and Innovation Basic Plan
In March 2021, the Japanese Cabinet approved and set in motion the 6th Science, Technology, and Innovation Basic Plan that maps out the science and technology policy of the government for the five years from FY2021. The government budgets JPY30tn in R&D spending over the five-year period, and hopes this will spur private investment. On the whole, it targets public-private investment of JPY120tn. The plan calls for a “societal vision predicated on innovation (societal change)” as its first theme, and promotes digitalization and the implementation of data-driven strategies in the government, as well as the preparation and development of next-generation infrastructure and technology such as supercomputers and quantum technology. In addition, the government will establish a JPY10tn university fund, and set up a research system for universities that can compete on the global stage.
Global industrial computer and CTO markets
IHS Markit valued the global industrial computer market at USD4.1bn in 2020, and estimates this market expanded at a CAGR of 4.7% in the seven years from 2013 to 2020. In 2021 and 2022, it forecasts YoY growth of 2.3% and 3.6%, respectively.
Japanese industrial computer and CTO markets
According to “Industrial PCs and their Expansion in the Edge Computing Field: Conditions and Outlook (FY2019)” published by the MIC Research Institute, the industrial computer market was worth JPY109.3bn in FY2018. MIC Research Institute expects the market to expand at a CAGR of 5.7% from FY2018 to FY2022, and reach JPY136.1bn in FY2022.
Based on its own survey, HPC Systems values the target market for its CTO segment at JPY187.0bn in 2021, and it sees this market expanding at a CAGR of 11.1% to JPY257.0bn in 2024.
Strengths and weaknesses
Strengths
Expertise in supercomputers and scientific computing software for quantum chemical computing, molecular dynamics, and other fields in the HPC segment
HPC Systems uses supercomputers to create operating environments for scientific computing, and this requires specialized expertise according to the company. Its systems consist of interconnected supercomputers, workstations, and servers. These differ from consumer PCs in that getting them to operate requires more than simply establishing the necessary connections and installing the appropriate software. The hardware, operating system (OS), middleware, software, and network that make up these systems must be optimized to ensure compatibility, enable parallel processing of multiple computational tasks, and reduce overall processing time. HPC Systems has accumulated extensive expertise in computing systems for the life science and material science fields. It is one of only a handful of Japanese companies that offer the Gaussian software package that has become the de-facto industry standard for quantum chemistry computing. It also develops and sells proprietary extended functionality for the software.
Ability to rapidly deliver cutting-edge products at affordable prices and secure orders mainly through customer inquiries in the HPC segment
Excluding tenders invitations by universities and public research institutions, the company mainly secures orders through direct customer inquiries. Customers typically refer to the product lineup and benchmarks (performance indicators) on the company’s website while they are researching the optimal hardware and software for their R&D objectives, and subsequently reach out to the company. Upon receiving an inquiry, the company’s sales staff provide explanations and propose products. Customers are primarily concerned with benchmarks and pricing, and HPC Systems lists benchmarks that it has independently verified on its website. The company has developed strong ties with Super Micro Computer and serves as a distributor for the latter. Taking advantage of this relationship, it has built an organization that facilitates rapid delivery of high-spec products at affordable prices, and Shared Research understands it has gained the recognition of its customers in this respect.
Capacity to produce and sell high-spec industrial computers with custom designs using a high-mix/low-volume manufacturing approach in the CTO segment
HPC Systems produces and supplies industrial computers with high specifications and custom designs in its CTO segment with a focus on high-mix/low-volume manufacturing. This approach differs from its competitors, which mainly mass-produce and sell low-end, general-purpose products. In addition, the company has established an organization to ensure long-term supply of its products to customers, whereas its rivals update their product models every few years. As a result, its repeat purchase rate exceeds 90%. Although it generates lower revenue than its competitors, the company boasts higher margins.
Weaknesses
Susceptible to direct impact from disruptions at suppliers of semiconductors and peripheral electronic parts
HPC Systems meticulously selects and procures hardware in function of the budgets and desired research environments of its customers. However, semiconductors have been in short supply since the autumn of 2020, and related shortages affect the hardware procured by the company. That said, the company did not suffer lost orders or delayed deliveries in FY06/21, and thus managed to avert direct impact. While it hedges against potential adverse effects by placing orders in advance, there is a risk the company may not be able to procure sufficient volume, in which case its earnings could take a direct hit.
Small operational scale hinders the company’s ability to accept large projects in the HPC segment
The order value in the HPC segment ranges from several million yen to over JPY400mn per project, but the order value for large projects in the HPC market can exceed JPY1.0bn. Shared Research understands that such projects are almost exclusively handled by major electronics manufacturers. HPC Systems’ workforce and revenue scale make it difficult for the company to pursue such orders even though it possesses the necessary technical expertise.
Difficulties in generating economies of scale in the CTO segment due to a production structure geared toward high-mix/low-volume manufacturing
HPC Systems ensures long-term stable supply of high-end computers in its CTO segment, and has achieved a repeat purchase rate of over 90% as a result. It designs industrial computers tailored to customer preferences, procures the optimal parts, and assembles these using a high-mix/low-volume manufacturing approach. Its custom designs and manufacturing process are two major differentiators for the company. However, this also means its order growth is dependent on increases in product sales at existing customers or, in the absence of this, the acquisition of new customers. In other words, the current production structure of the CTO segment is not conducive to economies of scale. In contrast, its rivals in this space mass-produce and sell general-purpose industrial computers.
Historical results and financial statements
Income statement
Note: Figures may differ from company data due to differences in rounding methods.
In FY06/21, revenue and recurring profit increased to record highs of JPY5.8bn and JPY666mn, respectively. In the four years from FY06/17 to FY06/21, revenue expanded at a CAGR of 10.6% and recurring profit at a CAGR of 27.2%.
In FY06/17 and FY06/19, the company recorded revenue for large projects worth over JPY400mn. As a result, revenue either expanded at a slower pace or declined in the subsequent years—namely, FY06/18 and FY06/20.
The cost of revenue ratio is around 70%, and typically rises in tandem with higher revenue ratios of large projects, tenders for universities or public research institutions in the HPC segment, or a higher revenue ratio of new projects in the CTO segment.
The SG&A ratio is around 20%. The main cost items under SG&A expenses are fixed costs such as personnel expenses and depreciation. However, sales growth requires personnel increases, and therefore inevitably drives up SG&A expenses.
There are no noteworthy items related to non-operating income and extraordinary expenses.
Balance sheet