Summary

Neural networks have become the organising principle of the global computing industry. This report provides a comprehensive analysis of the hardware, models, software and applications behind modern AI—from transformers, mixture-of-experts and post-transformer architectures through data-centre GPUs and custom accelerators to edge NPUs, neuromorphic, photonic and analog computing—with market forecasts to 2040 by segment, architecture family and region.

It maps the full value chain: architectures and algorithms, neural network hardware, software/models/platforms, and the markets and applications driving demand across cloud, enterprise, consumer devices, automotive, healthcare, robotics, finance, defence, telecoms and science. Regional analysis covers North America, China, South Korea, Japan, Taiwan, Europe and the UK, alongside the competitive landscape, investment and M&A, regulation and standards, and the energy, memory, geopolitical and capital-intensity risks shaping the sector.

The report profiles 195 companies, each rated on commercial readiness, and includes scenario-based forecasts, a technology roadmap and an outlook to 2040.

Report contents include:

• Executive summary: scope and market definition, key findings, market size and forecast 2025–2040, growth drivers and restraints, technology inflection points, competitive landscape, deal activity and strategic recommendations.
• Introduction: definitions, historical development, architecture taxonomy, training vs. inference, scaling laws and compute demand, the value chain and methodology.
• Neural network architectures and algorithms: MLPs, CNNs, RNNs/LSTMs, transformers and attention, mixture-of-experts, post-transformer and liquid networks, graph/topological, generative (GANs, VAEs, diffusion, world models), spiking, physics-informed/neural operators/KANs, quantum, biological/organoid, memory-augmented, and model compression/on-device AI.
• Neural network hardware: GPUs, hyperscaler custom ASICs, challenger accelerators, China's and Korea/Japan's ecosystems, edge NPUs, neuromorphic, photonic and optical interconnect, analog/in-memory/thermodynamic computing, memory/packaging/interconnect, and hardware market forecasts.
• Software, models and platforms: frameworks and compilers, foundation model developers, open-weight vs. proprietary models, inference economics and token demand, MLOps and inference serving, training data and labelling, and memory/agent infrastructure.
• Markets and applications: hyperscale cloud and AI labs, enterprise and agentic AI, consumer devices, automotive, healthcare, industrial/robotics/physical AI, financial services, defence, telecoms and science, with end-use forecasts.
• Global market forecasts 2025–2040 by segment, hardware vs. software, training vs. inference, cloud vs. edge, architecture family and region, with scenario analysis.
• Regional analysis: North America, China, South Korea, Japan, Taiwan, Europe, the UK, and rest of world.
• Competitive landscape, investment and M&A: market structure, venture funding, M&A and licensing, IPOs, strategic compute partnerships and the start-up landscape.
• Regulation, standards and ethics (EU AI Act, US export controls, UK and China policy, standards, sector regulation, liability/copyright/litigation); market challenges and risks (energy, memory supply, geopolitics, capital intensity, talent, safety); and future outlook to 2040 with roadmap and predictions.
• 195 company profiles rated on commercial readiness, plus appendices (year-by-year forecasts, glossary) and references.

Companies profiled include: 1X Technologies, Agibot (Zhiyuan Robot), Aidoc, Akhetonics, Alchip Technologies, Aleph Alpha, Alibaba Group (Qwen, T-Head), Alif Semiconductor, Ambarella, AMD, AMI Labs, Ampere Computing, Analog Devices, Anduril Industries, Anthropic, Arago, Arlequin AI, Arm Holdings, Aspinity, Astera Labs, Axelera AI, Axera Semiconductor, Ayar Labs, Baidu (incl. Kunlunxin), Baseten, Biren Technology, Black Forest Labs, Black Sesame Technologies, Blaize, Blumind, BrainChip, Broadcom, ByteDance, Cambricon Technologies, Cartesia, Celestial AI, Cerebras Systems, ChipAgents, Chipmind, CogniFiber, Cohere, CoreWeave, Cortical Labs, Crusoe, d-Matrix, Darktrace, Databricks, DeepSeek, DeepX, EdgeCortix, ElevenLabs, EnCharge AI, Enflame Technology, Engramme, Etched, Extropic, Ferroelectric Memory Company (FMC), Field AI, Figure AI, FinalSpark, Fireworks AI, Fractile, and more.

ページTOPに戻る

Table of Contents

1 EXECUTIVE SUMMARY
1.1 Report scope and market definition
1.2 Key findings
1.3 Market size and forecast 2025-2040
1.4 Growth drivers and restraints
1.5 Technology inflection points 2027-2040
1.6 Competitive landscape summary
1.7 Investment and deal activity 2025-2026
1.8 Strategic outlook and recommendations
2 INTRODUCTION
2.1 Defining neural networks
2.2 Historical development
2.3 Taxonomy of neural network architectures
2.4 Training and inference
2.5 Scaling laws and compute demand
2.6 The neural network value chain
2.7 Research methodology
3 NEURAL NETWORK ARCHITECTURES AND ALGORITHMS
3.1 Multilayer perceptrons and feedforward networks
3.2 Convolutional neural networks
3.3 Recurrent networks and LSTMs
3.4 Transformers and attention
3.5 Mixture-of-experts models
3.6 Post-transformer architectures
3.7 Liquid neural networks and continuous-time models
3.8 Graph and topological neural networks
3.9 Generative architectures: GANs, VAEs, diffusion and world models
3.10 Spiking neural networks
3.11 Physics-informed networks, neural operators and KANs
3.12 Quantum neural networks
3.13 Biological neural networks and organoid intelligence
3.14 Memory-augmented neural networks
3.15 Model compression, small language models and on-device AI
3.16 Architecture adoption outlook
4 NEURAL NETWORK HARDWARE
4.1 Hardware landscape
4.2 GPUs
4.3 Hyperscaler custom ASICs
4.4 Challenger accelerator companies
4.5 China's accelerator ecosystem
4.6 South Korean and Japanese accelerator developers
4.7 Edge neural processing units
4.8 Open-source neural network hardware
4.9 Neuromorphic processors
4.10 Photonic neural networks and optical interconnect
4.11 Analog, in-memory and thermodynamic computing
4.12 Memory, packaging and interconnect
4.13 Hardware market forecasts
5 SOFTWARE, MODELS AND PLATFORMS
5.1 Frameworks and compilers
5.2 Foundation model developers
5.3 Open-weight vs. proprietary models
5.4 Inference economics and token demand
5.5 Platforms, MLOps and inference serving
5.6 Training data and labelling
5.7 Memory and agent infrastructure
5.8 Software and model forecasts
6 MARKETS AND APPLICATIONS
6.1 Hyperscale cloud and AI labs
6.2 Enterprise and agentic AI
6.3 Consumer devices
6.4 Automotive and autonomous mobility
6.5 Healthcare and life sciences
6.6 Industrial, robotics and physical AI
6.7 Financial services
6.8 Defence, security and government
6.9 Telecommunications and quantum-secure communications
6.10 Science and materials discovery
6.11 End-use market forecasts
7 GLOBAL MARKET FORECASTS 2025-2040
7.1 Forecast approach
7.2 Total market
7.3 By segment
7.4 Hardware vs. software
7.5 Training vs. inference
7.6 Cloud vs. edge
7.7 By architecture family
7.8 By region
7.9 Scenario analysis
8 REGIONAL ANALYSIS
8.1 North America
8.2 China
8.3 South Korea
8.4 Japan
8.5 Taiwan
8.6 Europe
8.7 United Kingdom
8.8 India, Middle East, Russia, Australia and rest of world
9 COMPETITIVE LANDSCAPE, INVESTMENT AND M&A
9.1 Market structure
9.2 Venture funding
9.3 M&A and licensing
9.4 IPOs and public markets
9.5 Strategic compute partnerships
9.6 Start-up landscape
10 REGULATION, STANDARDS AND ETHICS
10.1 EU AI Act
10.2 US export controls
10.3 United Kingdom
10.4 China
10.5 Standards and certification
10.6 Sector regulation: healthcare and drug development
10.7 Liability, copyright and litigation
11 MARKET CHALLENGES AND RISKS
11.1 Energy and power
11.2 Memory and packaging supply constraints
11.3 Geopolitics and supply-chain concentration
11.4 Capital intensity and valuation risk
11.5 Talent, data and intellectual property
11.6 Safety, security and reliability
11.7 Risk assessment
12 FUTURE OUTLOOK 2027-2040
12.1 Technology roadmap
12.2 Ten predictions for 2040
12.3 Opportunity assessment
12.4 Wildcards and discontinuities
13 COMPANY PROFILES (195 company profiles)
14 APPENDICES
14.1 Detailed year-by-year forecasts
14.2 Glossary
15 REFERENCES

ページTOPに戻る

List of Tables/Graphs

List of Tables TABLE 1 Neural network market segments and scope definitions TABLE 2 Key findings summary TABLE 3 Global neural network market by segment, 2025-2040 (US$ billion) TABLE 4 Key market drivers TABLE 5 Key market restraints and challenges TABLE 6 Leading companies by market segment TABLE 7 Largest neural network funding rounds, M&A and licensing deals, 2025-2026 TABLE 8 Strategic implications by stakeholder type TABLE 9 Key milestones in neural network development, 1943-2026 TABLE 10 Taxonomy of neural network architectures and primary applications TABLE 11 Training vs. inference: workload, hardware and economics TABLE 12 Selected frontier models: architecture and parameter counts, 2020-2026 TABLE 13 Modelled hardware cost per million output tokens by model size, 2024-2040 (US$) TABLE 14 Value-chain layers, value capture and representative companies TABLE 15 Forecast methodology and key assumptions TABLE 16 Reconciliation with related Future Markets reports TABLE 17 CNN architectures and commercial applications TABLE 18 Transformer variants and efficiency techniques TABLE 19 Selected mixture-of-experts models: total and active parameters TABLE 20 Comparison of post-transformer architectures TABLE 21 Post-transformer model developers TABLE 22 Liquid AI: model families, platforms and funding TABLE 23 Graph and topological neural network developers and applications TABLE 24 World-model developers and funding TABLE 25 Spiking vs. conventional artificial neural networks TABLE 26 Spiking neural network hardware: digital, hybrid and analog implementations compared TABLE 27 Neuromorphic photonic architectures for spiking and recurrent neural networks TABLE 28 Scientific machine learning architectures TABLE 29 Quantum neural network and quantum machine learning developments, 2025-2026 TABLE 30 Biological computing platforms TABLE 31 Model compression and efficiency techniques TABLE 32 Typical efficiency gains from reduced numerical precision in edge NPUs (relative to FP32) TABLE 33 Sparsity techniques and their impact on neural network inference efficiency TABLE 34 Share of neural network compute by architecture family, 2024-2040 (%) TABLE 35 Neural network hardware categories compared TABLE 36 Neural network hardware energy efficiency by technology category TABLE 37 Data-center GPU roadmaps, NVIDIA and AMD, 2024-2028 TABLE 38 NVIDIA and AMD data-center revenue, 2024-2026 TABLE 39 Data-center GPU unit shipments and blended ASP, 2025-2040 TABLE 40 Hyperscaler custom neural network accelerators TABLE 41 AI ASIC unit shipments by programme, 2025-2040 (thousand units) TABLE 42 Challenger neural network accelerator companies: technology, funding and valuation TABLE 43 Total cost of ownership per accelerator-hour: merchant GPU vs. custom and start-up ASICs, 2025 TABLE 44 Wafer-scale and multi-die accelerator yield economics TABLE 45 Chinese neural network accelerator suppliers TABLE 46 Chinese AI chip IPOs, 2025-2026 TABLE 47 China AI accelerator capability gap to the NVIDIA flagship by workload type TABLE 48 Korean and Japanese neural network chip developers TABLE 49 Edge NPU suppliers by application TABLE 50 Edge AI consolidation: M&A, 2025-2026 TABLE 51 Edge NPU performance and power requirements by application TABLE 52 Edge neural network accelerators: published performance and efficiency figures TABLE 53 Estimated wafer and die costs for edge AI SoCs by process node TABLE 54 Smartphone NPU performance and efficiency evolution, 2017-2024 TABLE 55 Robot compute platforms: performance, power and target applications TABLE 56 Open-source and RISC-V neural network accelerator initiatives TABLE 57 Neuromorphic processors: commercial and research TABLE 58 Photonic computing and optical interconnect companies TABLE 59 Optical neural network processor architectures compared TABLE 60 Photonic neural network processors: vendor performance and efficiency claims TABLE 61 Optical switching and light-source suppliers for neural network clusters TABLE 62 Analog, in-memory and probabilistic computing companies TABLE 63 Energy cost of computation versus data movement TABLE 64 Memory technologies for in-memory neural network computing TABLE 65 Processing-in-memory and compute-in-memory products from memory suppliers and specialists TABLE 66 Emerging non-volatile memory devices for neural network synapses and in-memory compute TABLE 67 HBM supplier revenue share, 2024-2026 TABLE 68 HBM generations and per-accelerator memory capacity TABLE 69 AI rack bill of materials by component, 2023-2028 TABLE 70 Interconnect energy efficiency by technology TABLE 71 Memory bandwidth versus accelerator compute, 2026-2037 (representative flagship accelerator) TABLE 72 HBM stack shipments, pricing and revenue forecast, 2026-2037 (memory report forecast) TABLE 73 Memory tiers for neural network inference TABLE 74 Data-center NN accelerator revenue by type, 2025-2040 (US$ billion) TABLE 75 Edge NPU revenue by device, 2025-2040 (US$ billion) TABLE 76 Neuromorphic and novel NN hardware revenue by technology, 2025-2040 (US$ billion) TABLE 77 Low-power/high-efficiency AI semiconductor market by application, 2025-2036 (US$ billion; scope of the dedicated Future Markets report) TABLE 78 Optical computing report forecast: PIC content in AI accelerator interconnects and processors, 2026-2036 (US$ million) TABLE 79 Optical computing report forecast: optical processors by application, 2026-2036 (US$ billion) TABLE 80 Neural network frameworks and compiler stacks TABLE 81 Leading foundation model developers: models, funding and valuation TABLE 82 Leading open-weight model families TABLE 83 Token-processing indicators, 2024-2026 TABLE 84 Inference cost per million output tokens by model class and hardware, 2024-2040 (US$) TABLE 85 Memory consumption by inference workload class TABLE 86 Neural network platform, MLOps and inference-serving vendors TABLE 87 Data labelling and synthetic data providers TABLE 88 AI memory-layer companies TABLE 89 Software, models and platforms revenue, 2025-2040 (US$ billion) TABLE 90 Hyperscaler capital expenditure: 2025 actual and 2026 guidance TABLE 91 Data-center processor demand by facility type, 2025-2040 (US$ billion) TABLE 92 Sovereign AI build-outs by country, 2025-2030 TABLE 93 Enterprise neural network applications TABLE 94 On-device neural networks in consumer electronics TABLE 95 Flagship smartphone neural processing units, 2025-2026 generation TABLE 96 On-device LLM inference capability of AI PCs, 2026 TABLE 97 Automotive neural network compute platforms and driving stacks TABLE 98 SAE automation levels and in-vehicle neural network compute requirements TABLE 99 Waymo operating metrics, 2022-2026 TABLE 100 Neural network applications in healthcare and selected products TABLE 101 Robot foundation models and physical AI platforms TABLE 102 Robot foundation models: architecture, training scale and status TABLE 103 Physical AI technical constraints: current performance versus deployment thresholds TABLE 104 Global humanoid robot shipments, revenue and prices, 2025-2040 TABLE 105 Neural network applications in financial services TABLE 106 Neural network companies in defence and security TABLE 107 Neural network applications in telecoms and secure communications TABLE 108 Neural networks in scientific discovery TABLE 109 Neural network market by end-use sector, 2025-2040 (US$ billion) TABLE 110 Future Markets physical AI market forecast by sub-market, 2026-2040 (US$ billion; full-system revenue, not neural network value) TABLE 111 Physical AI value pool by stack layer, 2026-2036 (share of revenue excluding wearables) TABLE 112 Forecast assumptions by segment TABLE 113 Global neural network market, 2024-2040 TABLE 114 Segment growth summary TABLE 115 Hardware vs. software, models and platforms, 2025-2040 (US$ billion) TABLE 116 Data-center accelerator spend: training vs. inference, 2025-2040 (US$ billion) TABLE 117 Cloud vs. edge deployment, 2025-2040 (US$ billion) TABLE 118 Future Markets edge AI chip forecast by application, 2026-2036 (US$ billion; dedicated edge AI chips report basis) TABLE 119 Compute-linked market value by architecture family, 2025-2040 (US$ billion) TABLE 120 Neural network market by region, 2025-2040 (US$ billion) TABLE 121 Bear, base and bull scenarios, 2025-2040 (US$ billion) TABLE 122 US neural network policy and infrastructure programmes TABLE 123 China's neural network ecosystem: chips, models and policy TABLE 124 South Korean neural network programmes and investments TABLE 125 Japanese neural network programmes and companies TABLE 126 European neural network companies and funding TABLE 127 UK compute, research and policy initiatives TABLE 128 Sovereign AI programmes outside the US, China and Europe TABLE 129 Data-center NN accelerator vendor share, 2025 and 2026 (estimates) TABLE 130 Neural network funding rounds, 2025-2026 TABLE 131 M&A and licensing deals, 2024-2026 TABLE 132 Neural network-related IPOs, 2025-2026 TABLE 133 Major compute and model partnerships TABLE 134 EU AI Act implementation timeline, including Digital Omnibus changes TABLE 135 US export controls affecting neural network hardware, 2022-2026 TABLE 136 UK legislative and policy developments TABLE 137 Chinese AI regulation and procurement measures TABLE 138 AI management and safety standards TABLE 139 FDA-EMA guiding principles for AI in drug development TABLE 140 Notable AI litigation and rulings, 2025-2026 TABLE 141 Data-center power demand indicators TABLE 142 AI server rack configuration and power roadmap, 2025-2040 TABLE 143 Data-center cooling configurations: density, PUE and overhead TABLE 144 Embodied carbon across the AI hardware hierarchy TABLE 145 AI accelerator memory requirements and cost, 2024-2030 TABLE 146 Memory and storage content and cost per AI server node, 2026-2037 TABLE 147 Top 15 strategic choke points in the AI hardware supply chain TABLE 148 Capital expenditure vs. revenue indicators TABLE 149 AI-driven EDA approaches by design-flow stage TABLE 150 Risk matrix: likelihood and impact TABLE 151 Technology readiness and commercialisation timeline, 2027-2040 TABLE 152 Photonic neural network technology roadmap, 2026-2036 TABLE 153 Technology readiness of neural-network-driven physical AI applications, 2026 TABLE 154 HBM technology roadmap, per-stack specifications TABLE 155 Ten predictions for the neural network market to 2040 TABLE 156 Opportunity matrix by segment TABLE 157 Companies profiled (index) TABLE 158 Commercial readiness stages: definitions TABLE 159 Commercial readiness of profiled companies by company group (number of companies) TABLE 160 Profiled companies by commercial readiness stage List of Figures FIGURE 1 Global neural network market by segment, 2024-2040 (US$ billion) FIGURE 2 Neural network technology roadmap, 2025-2040 FIGURE 3 Timeline of neural network milestones, 1943-2026 FIGURE 4 Neural network architecture taxonomy FIGURE 5 Training compute of notable neural network models, 2012-2025 (FLOP, log scale) FIGURE 6 Neural network market value chain FIGURE 7 Compute cost of attention vs. sequence length: quadratic vs. linear architectures FIGURE 8 Share of neural network compute by architecture family, 2025, 2030 and 2040 FIGURE 9 Data-center NN accelerator revenue by type, 2024-2040 (US$ billion) FIGURE 10 Edge NPU revenue by device, 2024-2040 (US$ billion) FIGURE 11 Neuromorphic and novel NN hardware revenue by technology, 2024-2040 (US$ billion) FIGURE 12 Valuations of leading model developers, 2026 FIGURE 13 Google monthly tokens processed, 2024-2026 (log scale) FIGURE 14 Software, models and platforms revenue, 2024-2040 (US$ billion) FIGURE 15 Neural network market share by end-use sector, 2025 vs. 2040 FIGURE 16 Global neural network market, 2024-2040 (US$ billion, base case) FIGURE 17 Segment mix, 2025, 2030 and 2040 FIGURE 18 Hardware vs. software, models and platforms, 2024-2040 (US$ billion) FIGURE 19 Data-center accelerator spend: training vs. inference, 2024-2040 (US$ billion) FIGURE 20 Regional share, 2025, 2030 and 2040 FIGURE 21 Scenario comparison, 2024-2040 (US$ billion) FIGURE 22 Data-center NN accelerator vendor share, 2025 FIGURE 23 Largest neural network funding rounds, 2025-2026 (US$ billion, log scale) FIGURE 24 Neural network start-up landscape by layer FIGURE 25 Neural network regulatory timeline, 2024-2028 FIGURE 26 Neural network market risk heat map FIGURE 27 Commercialisation timeline for emerging neural network technologies, 2025-2040 FIGURE 28 Commercial readiness of profiled companies by company group FIGURE 29 Agibot A2 FIGURE 30 Cerebras WSE-3 FIGURE 31 MemryX MX3 FIGURE 32 SiMa.ai Modalix