High Performance Computing Market Size, Share & Forecast to 2035

The High Performance Computing (HPC) Market covers the supercomputers, compute clusters, storage, interconnect, software, services and cloud capacity that research centers, government laboratories and enterprises buy to run simulation, modeling, AI-enhanced science and large-scale data analytics. DC Market Insights values the global High Performance Computing Market at USD 48.80 billion in 2025 and forecasts USD 115.78 billion by 2035, a CAGR of 9.02%. The 2025 value is built from units × price: DC Market Insights estimates 6 leadership-class systems × USD 431.70 million, 255 large supercomputers × USD 22.37 million, 4,850 divisional systems × USD 1.27 million and 51,500 departmental and workgroup systems × USD 0.16 million, which gives USD 22.60 billion of compute servers; storage, networking, software and services add USD 15.27 billion, and cloud HPC services add USD 10.94 billion. AI-enhanced HPC is the fastest-growing workload segment at 16.53% a year because national laboratories now buy systems built around tens of thousands of graphics processing units (GPUs), and the Middle East and Africa is the fastest-growing region at 10.49% a year as sovereign compute programs start from a small base.

Executive Summary

The Global High Performance Computing Market size was valued at USD 29.16 billion in 2020, reached USD 48.80 billion in 2025, and is anticipated to reach USD 115.78 billion by 2035, at a CAGR of 9.02% during the forecast period.

Report attribute	Details
Historical Period	2020-2024
Base Year	2025
Forecast Period	2026-2035
High Performance Computing Market Size 2025	USD 48.80 Billion
High Performance Computing Market, CAGR	9.02%
High Performance Computing Market Size 2035	USD 115.78 Billion

The market grew at 10.84% a year between 2020 and 2025, and 2025 added 11.55% over the USD 43.75 billion of 2024 as DC Market Insights estimates GPU-dense systems spread from leadership centers to university and industrial clusters. DC Market Insights forecasts growth of 9.02% a year from 2025 to 2035. On-premises spending grows more slowly, at 7.39% a year from USD 37.86 billion to USD 77.26 billion, while cloud HPC services rise from 22.4% of the market in 2025 to 33.3% by 2035.

Two forces set the pace. First, leadership systems are getting larger and dearer: El Capitan cost USD 600 million and draws 30 MW. Second, the buyer base is widening, as AI-enhanced HPC moves into life sciences, manufacturing and financial services, where workloads were once run on small departmental clusters.

Market Scope: What Does the High Performance Computing Market Cover?

The High Performance Computing Market covers spending on systems and services used for compute-intensive technical and scientific work, valued at first sale in current US dollars, with 2025 as the base year and forecasts to 2035. DC Market Insights sizes it at USD 48.80 billion in 2025.

Included. Compute servers sold as HPC systems, from leadership-class supercomputers to departmental and workgroup clusters; parallel file systems and other HPC storage; HPC interconnects and cluster networking bought with or for these systems; cluster management, scheduling, compilers and simulation software licenses; installation, integration, support and managed services; and cloud HPC services, meaning HPC-configured instances, clusters and capacity rented from public cloud providers for technical workloads. AI-enhanced HPC systems at research centers, government laboratories and enterprises, such as GPU systems used for AI-driven science, are included.

Excluded. Hyperscale AI training and inference clusters built to serve commercial AI models and cloud AI services, which are covered in the AI Infrastructure Market report; general-purpose enterprise servers; data center buildings, power and cooling plant; facility electricity; and staff costs of the operating centers.

How the segments fit together. The offering, workload and end-user splits each re-cut the same USD 48.80 billion. The system-class split covers only on-premises compute servers, worth USD 22.60 billion in 2025, and divides them by the price band of the system.

Market Drivers: What Drives the High Performance Computing Market?

Three drivers explain most of the growth in the High Performance Computing Market: national laboratories buying GPU systems at a new scale, government programs that fund successive flagship machines, and AI methods that pull new industries into HPC. Together they lift the market from USD 48.80 billion in 2025 to USD 115.78 billion by 2035.

Leadership Systems Now Cost USD 600 Million and Carry 100,000 GPUs

The top of the market has moved up a price band. El Capitan at Lawrence Livermore National Laboratory is a USD 600 million system that uses 30 MW of power across 87 racks. The US Department of Energy (DOE) and NVIDIA Corporation (NVIDIA) have announced Solstice for Argonne National Laboratory with 100,000 NVIDIA Blackwell GPUs, more than twice the accelerators in El Capitan, alongside a smaller 10,000-GPU system called Equinox expected in 2026. DC Market Insights estimates the average leadership-class system at USD 431.70 million in 2025, rising to USD 693.40 million by 2035, and expects leadership-class spending to grow at 11.40% a year, the fastest of the four system classes.

Governments Fund Successive Flagship Machines

Public budgets set the floor under the High Performance Computing Market. DOE announced nine new supercomputers across Argonne, Oak Ridge and Los Alamos with Hewlett Packard Enterprise (HPE), NVIDIA, Advanced Micro Devices (AMD) and Oracle in October 2025. In Europe, JUPITER, the first European exascale system, cost EUR 500 million for the system and an expected six years of operation, funded 50% by the European Union and 25% each by the German federal research ministry and North Rhine-Westphalia. The United Kingdom confirmed up to GBP 750 million for a new national supercomputer at the University of Edinburgh in June 2025, and Japan’s Institute of Physical and Chemical Research (RIKEN) is developing FugakuNEXT with Fujitsu and NVIDIA for operation around 2030. Government and defense buyers account for USD 11.03 billion, or 22.6%, of 2025 spending.

AI-Enhanced HPC Grows at 16.53% a Year

AI methods are changing what HPC systems are bought for. DC Market Insights estimates AI-enhanced HPC, meaning AI training, inference and surrogate models run on HPC systems alongside simulation, at USD 11.76 billion in 2025, or 24.1% of the market, rising to USD 54.30 billion by 2035, when it overtakes classical simulation and modeling. The TOP500 list shows the hardware shift: 276 of the 500 systems on the June 2026 list use accelerators, up from 255 six months earlier.

Market Trends: How Is High Performance Computing Changing?

The High Performance Computing Market is moving toward exascale systems outside the United States, direct liquid cooling as a default, and HPC capacity rented from the cloud, which reaches USD 38.51 billion by 2035.

Five Systems Now Exceed One Exaflop

The June 2026 TOP500 list has five systems above 1 exaflop/s on the High-Performance Linpack (HPL) benchmark: LineShine at 2.198, El Capitan at 1.809, Frontier at 1.353, Aurora at 1.012 and the JUPITER Booster at 1.000. LineShine, the new No. 1, is installed at the National Supercomputing Centre in Shenzhen, China, and ran HPL on 13,789,440 cores. The combined performance of all 500 systems reached 18.73 exaflop/s, and the entry level rose to 2.66 petaflop/s.

Efficiency Becomes a Purchase Criterion

Power is now a design limit, so buyers compare systems on performance per watt as well as peak speed. The No. 1 system on the June 2026 Green500 list, KAIROS at CALMIP and the University of Toulouse in France, delivers 73.282 gigaflops per watt. The Doudna system for the National Energy Research Scientific Computing Center (NERSC) at Berkeley Lab, due in 2026, uses Dell ORv3 direct liquid-cooled servers with the NVIDIA Vera Rubin platform and is designed to deliver more than ten times the performance of Perlmutter. Cooling choices sit in the facility budget rather than in this market, and the Data Center Liquid Cooling Market report sizes them.

Cloud HPC Rises to 33.3% of Spending

Cloud HPC services hold 22.4% of the High Performance Computing Market in 2025, and DC Market Insights forecasts their share at 33.3% by 2035 as the segment grows 13.41% a year. Eagle, a Microsoft Azure system in the United States, ranks No. 7 on the June 2026 TOP500 list, which shows that cloud providers can deliver leadership-scale capacity. Cloud suits bursty engineering workloads and AI experiments; steady, sensitive and classified work stays on premises.

Market Challenges: What Holds the High Performance Computing Market Back?

Power and facility limits, export controls and the long budget cycles of public buyers are the three constraints DC Market Insights expects to slow the High Performance Computing Market, which still grows 9.02% a year.

Power Draw Reaches 30 MW per System

A leadership system now draws as much power as a small data center. El Capitan uses 30 MW drawn from the local grid. Sites that cannot add grid capacity or liquid cooling loops must delay procurements or split them, which is one reason DC Market Insights models leadership-class units at only 6 in 2025 and 11 in 2035 even as spending per system rises.

Export Controls Split the Supply Chain

Trade rules now shape who can buy which processors. In March 2025, the US Bureau of Industry and Security added 12 entities, 11 in China and 1 in Taiwan, to the Entity List for their involvement in developing Chinese exascale supercomputers. LineShine, the new No. 1, uses the custom LingKun platform, the proprietary LingQi interconnect and Kylin OS. DC Market Insights models China’s market at USD 5.60 billion in 2025, growing 9.61% a year, slower than India and South Korea.

Public Budgets Move in Multi-Year Steps

Government and academic buyers account for 40.5% of 2025 spending in the DC Market Insights model, and their purchases follow appropriation cycles rather than demand. Some announced systems depend on future funding: Los Alamos National Laboratory said its project will be supported through anticipated future funding, contingent upon the availability of appropriations. The timing of any single leadership machine can move market value by several hundred million dollars in a year.

Market Opportunities: Where Are the New Revenue Pools?

The two largest new revenue pools in the High Performance Computing Market are life sciences and healthcare, growing at 11.83% a year, and AI-enhanced HPC across all industries, which adds USD 42.54 billion of annual spending between 2025 and 2035.

Life Sciences Triples Its HPC Spend to USD 18.06 Billion

Life sciences and healthcare buyers spent USD 5.91 billion on HPC in 2025, and DC Market Insights expects USD 18.06 billion by 2035. Protein structure prediction, molecular dynamics and genomics pipelines combine simulation with AI models, and 36.0% of the segment’s 2025 spending is already AI-enhanced, the highest share of any end-user.

Mid-Size Centers Upgrade to GPU Systems

Large supercomputers, the band below leadership systems, are worth USD 5.70 billion in 2025 and grow at 9.59% a year to USD 14.25 billion by 2035. The EuroHPC Joint Undertaking reported two new systems on the June 2026 list, DAEDALUS at 85.69 petaflops and Arrhenius at 66.82 petaflops, and oversees 19 AI factories and 13 AI Factory Antennas across Europe. Programs of this kind turn national HPC centers into suppliers of AI capacity to companies, which widens the buyer base for vendors and integrators.

Pricing: How Much Does an HPC System Cost?

In the High Performance Computing Market, an average leadership-class system costs about USD 431.70 million in 2025, a large supercomputer about USD 22.37 million, a divisional system about USD 1.27 million and a departmental or workgroup cluster about USD 0.16 million. These are DC Market Insights estimates of the compute server price per system, excluding storage, software, services and facilities.

Average system price (USD million)	2020	2025	2030	2035
Leadership-Class Systems	312.40	431.70	558.90	693.40
Large Supercomputers	16.53	22.37	28.84	35.62
Divisional Systems	1.05	1.27	1.52	1.81
Departmental and Workgroup Systems	0.13	0.16	0.18	0.20

Public records anchor the top line: El Capitan cost USD 600 million, and JUPITER’s EUR 500 million covers the system plus six years of operation. Prices rise in every class because each system generation packs more accelerators and memory per node; performance per dollar improves much faster than price per system.

Standards and Regulation: Which Rules Shape High Performance Computing?

Three sets of rules shape the High Performance Computing Market: benchmark standards that buyers write into tenders, export controls on advanced computing, and energy reporting for the facilities that host the systems.

HPL, TOP500 and Green500. The TOP500 list ranks systems by their HPL benchmark result, and the Green500 list ranks them by energy efficiency. The June 2026 Green500 leader reached 73.282 gigaflops per watt.

US export controls. The US Bureau of Industry and Security uses the Entity List to restrict sales to organizations involved in supercomputer development; its March 2025 rule, effective March 25, 2025, added 12 entities tied to Chinese exascale supercomputers.

EU data center reporting. Commission Delegated Regulation (EU) 2024/1364 requires operators of data centers with installed IT power demand of at least 500 kW to report energy and sustainability indicators by 15 September 2024, then by 15 May 2025 and every year thereafter.

Market Segmentation: Which Segment Leads the High Performance Computing Market?

Compute servers lead the High Performance Computing Market with 46.3% of 2025 value, simulation and modeling leads workloads with 63.4%, and government and defense leads end-users with 22.6%. AI-enhanced HPC is the fastest-growing segment at 16.53% a year, against 7.93% for high-performance data analytics and 4.57% for simulation and modeling, because new systems are bought for AI-driven science as much as for classical simulation.

By Offering: Compute Servers Lead with 46.3%

Offering	2025 (USD billion)	Share 2025	2035 (USD billion)	CAGR 2025-2035
Compute Servers	22.60	46.3%	45.39	7.22%
Storage	5.15	10.6%	11.12	8.00%
Networking and Interconnect	2.35	4.8%	5.36	8.59%
Software	3.37	6.9%	6.68	7.08%
Professional and Managed Services	4.39	9.0%	8.72	7.10%
Cloud HPC Services	10.94	22.4%	38.51	13.41%

Cloud HPC services grow fastest among offerings at 13.41% a year. Storage and networking grow faster than compute servers because AI-enhanced workloads move more data per flop and need faster parallel file systems.

By System Class: Departmental and Workgroup Systems Lead with 36.1%

System class (compute servers)	2025 (USD billion)	Share 2025	2035 (USD billion)	CAGR 2025-2035
Leadership-Class Systems	2.59	11.5%	7.63	11.40%
Large Supercomputers	5.70	25.2%	14.25	9.59%
Divisional Systems	6.15	27.2%	11.75	6.68%
Departmental and Workgroup Systems	8.15	36.1%	11.76	3.74%

Departmental and workgroup clusters remain the largest class by value because there are about 51,500 of them sold each year, but they grow slowest as small users shift to cloud capacity. By 2035, large supercomputers overtake both smaller classes.

By Workload: Simulation and Modeling Leads with 63.4%

Workload	2025 (USD billion)	Share 2025	2035 (USD billion)	CAGR 2025-2035
Simulation and Modeling	30.94	63.4%	48.39	4.57%
AI-Enhanced HPC	11.76	24.1%	54.30	16.53%
High-Performance Data Analytics	6.10	12.5%	13.08	7.93%

AI-enhanced HPC overtakes simulation and modeling by 2035 and reaches 46.9% of the market. Simulation keeps growing in absolute terms because climate, materials, engineering and national security codes remain simulation-based.

By End-User: Government and Defense Lead with 22.6%

End-user	2025 (USD billion)	Share 2025	2035 (USD billion)	CAGR 2025-2035
Government and Defense	11.03	22.6%	23.62	7.91%
Academic and Research	8.74	17.9%	17.48	7.18%
Manufacturing and Engineering	7.52	15.4%	18.41	9.37%
Life Sciences and Healthcare	5.91	12.1%	18.06	11.83%
Financial Services	4.78	9.8%	12.39	9.99%
Energy and Utilities	4.05	8.3%	7.99	7.03%
Others	6.78	13.9%	17.83	10.15%

Life sciences and healthcare grow fastest among end-users at 11.83% a year. Others, which include media, telecom, retail and transport, grow at 10.15% as AI-enhanced analytics spread. Energy and utilities grow slowest, at 7.03%, because DC Market Insights expects seismic and reservoir simulation budgets to grow more slowly than AI-led workloads.

DC Exclusive: The HPC Workload × End-User Matrix

At USD 7.33 billion in 2025, simulation and modeling for government and defense is the single largest cell of the matrix and 15.0% of all HPC spending. The matrix below cross-tabulates what the High Performance Computing Market is used for against who buys it. It is a DC Market Insights model, built from 31 inputs: 7 end-user totals, 3 workload totals and 21 workload-mix weights, one for each cell, balanced so that every row and column sums to its published total.

2025, USD billion	Simulation and Modeling	AI-Enhanced HPC	Data Analytics	Total
Government and Defense	7.33	2.64	1.06	11.03
Academic and Research	5.63	2.35	0.75	8.74
Manufacturing and Engineering	5.60	1.20	0.72	7.52
Life Sciences and Healthcare	2.98	2.12	0.80	5.91
Financial Services	2.05	1.16	1.58	4.78
Energy and Utilities	3.10	0.60	0.35	4.05
Others	4.24	1.69	0.85	6.78
Total	30.94	11.76	6.10	48.80

Three patterns stand out. Life sciences and healthcare spend 36.0% of their HPC budget on AI-enhanced work, the highest share of any buyer. Financial services spend 33.0% on high-performance data analytics, as risk and pricing models run on large data sets. Manufacturing and engineering spend 74.5% on simulation, so computer-aided engineering software and CPU-dense clusters remain the core of that segment.

The system-price model behind these figures has two inputs per system class that any reader can check: the number of systems sold each year and their average price. Multiplying them gives compute server spend; attach ratios for storage, networking, software and services give on-premises spending, and cloud HPC services complete the market. The model and its assumptions are explained on our research methodology page.

Regional Insights: Which Region Leads the High Performance Computing Market?

North America leads the High Performance Computing Market with 41.8% of 2025 value, and the Middle East and Africa is the fastest-growing region at 10.49% a year, because DC Market Insights expects sovereign compute programs to add national systems from a small base.

North America Leads with 41.8% Share

North American buyers spent USD 20.40 billion on HPC in 2025, and DC Market Insights forecasts USD 45.85 billion by 2035, a CAGR of 8.44%. The United States hosts El Capitan, Frontier and Aurora, and DOE’s October 2025 program adds nine systems, including Lux and Discovery at Oak Ridge, which are part of a USD 1 billion public-private investment with AMD, HPE and Oracle.

Asia Pacific Holds 26.1% Share

Asia Pacific is worth USD 12.74 billion in 2025 and USD 33.81 billion by 2035, a CAGR of 10.25%, the second-fastest region. China holds the No. 1 TOP500 system, Japan is building FugakuNEXT, and India and South Korea add national and industrial systems from smaller bases.

Europe Holds 25.4% Share

Europe is worth USD 12.40 billion in 2025 and USD 27.67 billion by 2035, a CAGR of 8.36%. Joint EuroHPC procurement, JUPITER in Germany and the planned national supercomputer in Edinburgh anchor spending, while the AI factories program extends HPC access to companies.

Middle East and Africa Holds 3.5% Share

The Middle East and Africa is worth USD 1.71 billion in 2025 and USD 4.63 billion by 2035, growing fastest at 10.49% a year. Saudi Arabia accounts for about a third of the region in 2025 in the DC Market Insights allocation, and its share rises as national AI and research programs mature.

Latin America Holds 3.2% Share

Latin America is worth USD 1.56 billion in 2025 and USD 3.82 billion by 2035, a CAGR of 9.36%. Brazil accounts for close to half of regional spending in the DC Market Insights allocation.

Country Analysis: Which Countries Spend the Most on High Performance Computing?

The United States is the largest country in the High Performance Computing Market at USD 18.85 billion in 2025, 38.6% of global value, followed by China at USD 5.60 billion (11.5%) and Germany at USD 3.04 billion (6.2%). The 12 largest countries account for 83.2% of 2025 spending.

Rank	Country	Region	2025 (USD billion)	Share 2025	2035 (USD billion)	CAGR 2025-2035
1	United States	North America	18.85	38.6%	42.18	8.39%
2	China	Asia Pacific	5.60	11.5%	14.03	9.61%
3	Germany	Europe	3.04	6.2%	6.64	8.14%
4	Japan	Asia Pacific	2.87	5.9%	6.83	9.07%
5	United Kingdom	Europe	2.21	4.5%	5.15	8.84%
6	France	Europe	2.01	4.1%	4.32	7.95%
7	Canada	North America	1.55	3.2%	3.67	8.99%
8	Italy	Europe	1.19	2.4%	2.60	8.13%
9	South Korea	Asia Pacific	1.02	2.1%	2.81	10.66%
10	India	Asia Pacific	0.90	1.9%	3.52	14.54%
11	Brazil	Latin America	0.72	1.5%	1.72	9.12%
12	Australia	Asia Pacific	0.65	1.3%	1.69	10.04%

India is the fastest-growing large country market at 14.54% a year, ahead of South Korea at 10.66%, as national programs and industrial users add GPU systems. The country figures are a DC Market Insights allocation of the regional totals, weighted by each country’s estimated share of installed and newly procured HPC capacity.

Forecast Scenarios: How Could the 2035 Forecast Change?

The High Performance Computing Market reaches USD 115.78 billion by 2035 in the base case, with a range of USD 96.45 billion to USD 137.70 billion across the low and high scenarios.

Scenario	Assumption from 2026	2035 (USD billion)	CAGR 2025-2035
Low	Public HPC budgets flat in real terms; AI-enhanced share of spending capped near 35%	96.45	7.05%
Base	Leadership systems rise to 11 a year; AI-enhanced HPC reaches 46.9% of spending	115.78	9.02%
High	Sovereign AI and science programs add systems faster; cloud HPC grows 3 points a year faster	137.70	10.93%

The low case is what a pause in public funding looks like: leadership systems slip, mid-size centers extend the life of existing clusters, and cloud absorbs only part of the gap. The high case assumes that national AI programs keep funding GPU systems at research centers and that power and cooling are available on time. The split between cloud and on-premises purchases is the main swing factor for vendors, since it changes who buys the hardware more than how much is bought.

Competitive Insights: Who Leads the High Performance Computing Market?

• HPE
• Dell Technologies
• Lenovo
• NVIDIA
• AMD
• Intel
• Bull
• Fujitsu
• IBM
• NEC
• Penguin Solutions
• Inspur
• Microsoft Azure
• Amazon Web Services
• Google Cloud
• Oracle Cloud Infrastructure
• DDN

HPE is the market leader among system builders, with an estimated 15% of 2025 HPC compute server spending, about USD 3.39 billion (DC Market Insights estimate, set against HPE’s fiscal 2025 Server segment revenue of USD 17.75 billion reported to the US Securities and Exchange Commission, which also includes general-purpose servers). HPE’s Server segment rose 10% from USD 16.10 billion in fiscal 2024 and includes integrated systems for HPC and supercomputing, including exascale, sold as HPE Cray EX and HPE Cray XD. HPE built El Capitan and partners on Lux and Discovery at Oak Ridge. Dell Technologies builds Doudna at NERSC with NVIDIA, and Bull, which Atos sold to the French State in March 2026, generated about EUR 0.7 billion of revenue in fiscal 2025 across HPC, quantum, business computing and AI. Among processor suppliers, Intel provides the processors in 53.20% of TOP500 systems and AMD in 38.40%, while NVIDIA supplies the GPUs for Solstice, Equinox and Doudna. Fujitsu partners with RIKEN on FugakuNEXT, and Microsoft Azure operates Eagle, the No. 7 system on the June 2026 TOP500 list.

Recent Developments

• In May 2025, DOE announced Doudna, a Dell and NVIDIA system for NERSC at Berkeley Lab, due in 2026 and designed to deliver more than ten times the performance of Perlmutter (Source: DOE announcement).
• In June 2025, the UK government confirmed up to GBP 750 million for a new national supercomputer at the University of Edinburgh (Source: UK government press release).
• In August 2025, RIKEN launched an international initiative with Fujitsu and NVIDIA to develop FugakuNEXT, targeting up to a hundredfold increase in application performance over Fugaku and operation around 2030 (Source: RIKEN news release).
• In October 2025, DOE, AMD, HPE and Oracle announced the Lux and Discovery systems at Oak Ridge National Laboratory as part of a USD 1 billion public-private investment (Source: ITBrief news report).
• In March 2026, Atos completed the sale of Bull, its advanced computing activities, to the French State for an enterprise value of up to EUR 404 million including earn-outs (Source: Atos press release).
• In June 2026, LineShine at the National Supercomputing Centre in Shenzhen became the No. 1 system on the TOP500 list at 2.198 exaflop/s (Source: TOP500 list release).

We follow these High Performance Computing Market deals, launches and capacity announcements as they happen in our data center industry updates.

Methodology: How We Built the High Performance Computing Market Model

DC Market Insights built this market bottom-up from 16 annual cohorts (2020 to 2035) of system purchases in 4 system classes, each with a unit count and an average price, which gives compute server spend of USD 22.60 billion in 2025. Attach ratios then add storage at 22.8% of compute server spend, networking and interconnect at 10.4%, software at 11.2% of hardware and services at 14.6% of hardware, and a separate cloud HPC series adds rented capacity. In all, the sizing rests on 49 inputs: 17 unit anchors, 16 price anchors, 12 attach-ratio anchors and 4 cloud anchors. The total was split by 6 offerings, 4 system classes, 3 workloads, 7 end-users and 5 regions, and every split sums to the market total. The result was checked top-down against public records: the USD 600 million cost of El Capitan and the EUR 500 million budget of JUPITER, which also covers six years of operation, for the leadership class, HPE’s USD 17.75 billion Server segment and Bull’s EUR 0.7 billion of revenue for vendor scale, and the TOP500 list for the count and mix of large systems. Country values allocate each regional total by country weights for installed and new HPC capacity. The analyst named on this page built the model, and the reviewer checked the arithmetic, the sources and the dates before publication.