AI Chip Market to Reach $677.59 Billion by 2035 as Inference and Custom Silicon Reshape Global Compute Demand

DC Market Insights forecasts 13.85% CAGR through 2035; GPUs hold 76.9% of 2025 market value while AI ASICs and SoCs expand at 19.61% annually

LONDON, Oct. 8, 2026 /PRNewswire/ -- The global AI Chip Market was valued at $185.26 billion in 2025 and is projected to reach $677.59 billion by 2035, expanding at a compound annual growth rate of 13.85%, according to a new study published by DC Market Insights. The market is moving from a training-led buildout toward a broader compute cycle in which inference, custom cloud silicon, on-device neural processing units and sovereign AI programs increasingly determine chip volumes, architecture choices and capital allocation.

Full report: https://www.dcmarketinsights.com/report/ai-chip-market

AI Chip Spending Scales Beyond Training

DC Market Insights estimates that the AI Chip Market expanded from $13.57 billion in 2020 to $185.26 billion in 2025 as data center accelerators moved into volume production. The firm forecasts a further rise to $260.01 billion in 2026, followed by slower but still substantial growth through 2035. Data center accelerators accounted for 91.6% of market value in 2025, illustrating how strongly the current market remains tied to hyperscale and AI data center investment.

The next leg of growth is increasingly linked to inference. AI inference represented $85.60 billion, or 46.2% of 2025 value, compared with $86.92 billion for training. DC Market Insights projects inference to grow at 17.76% annually and reach $439.05 billion by 2035, or 64.8% of total market value. The shift matters because inference is persistent: deployed models serve users and agents for years, creating recurring demand for cost-efficient accelerators across cloud, enterprise, automotive and edge environments.

Custom AI Silicon Gains Share While GPUs Remain the Largest Segment

GPUs generated $142.39 billion in 2025 and represented 76.9% of the market, making them the dominant AI chip category. DC Market Insights forecasts GPU revenue to reach $440.85 billion by 2035 at an 11.96% CAGR. AI ASICs and SoCs, however, are the fastest-growing major chip type. Their value is projected to increase from $32.64 billion in 2025 to $195.65 billion in 2035, a 19.61% CAGR, lifting their share from 17.6% to 28.9%.

The increase in custom silicon reflects efforts by hyperscale cloud providers and AI labs to lower inference cost per token and improve performance per watt. Google, Amazon Web Services and Microsoft are deploying their own TPU, Trainium and Maia architectures, while Broadcom is a major design and supply-chain partner for custom accelerators. Merchant suppliers are responding by expanding inference-focused products, strengthening rack-scale platforms and integrating higher-bandwidth memory and networking more tightly with compute.

Gigawatt-Scale Agreements Extend Visibility Into 2029

The report identifies large supply agreements as one of the clearest indicators of future demand. In October 2025, AMD and OpenAI announced a 6 GW GPU agreement beginning with 1 GW of Instinct MI450 capacity in the second half of 2026. OpenAI and Broadcom also agreed to deploy 10 GW of custom AI accelerators, with deployments targeted to begin in the second half of 2026 and complete by the end of 2029.

DC Market Insights estimates that one gigawatt of AI capacity can carry roughly $17 billion to $22 billion of accelerator content at 2025 prices. Commitments at this scale push AI semiconductor demand beyond conventional annual procurement cycles and increase the strategic value of foundry capacity, advanced packaging, high-bandwidth memory, substrate availability, interconnects and power infrastructure.

Edge and Device AI Expands the Addressable Market

Although data centers dominate market value, edge and device AI chips are projected to grow faster, at 16.64% annually. Their share rises from 8.4% of the market in 2025 to 10.7% in 2035. The category includes automotive AI SoCs, client NPUs in PCs and smartphones and industrial or robotics accelerators used in smart cameras, machines and edge gateways.

DC Market Insights values automotive AI chips at $6.42 billion in 2025 and industrial and edge AI chips at $3.94 billion. Client NPU value in AI-capable PCs and smartphones totaled an estimated $5.13 billion. The expansion of Copilot-class PCs, higher TOPS requirements, automated driving, robotics and machine vision creates a second demand pool that is less concentrated than hyperscale data center spending.

North America Leads, but Asia Pacific and the Middle East Gain Momentum

North America accounted for $104.42 billion, or 56.4% of global AI chip value in 2025, and is forecast to reach $349.15 billion by 2035. Asia Pacific represented 26.8% of the market at $49.69 billion and is projected to reach $196.82 billion by 2035, expanding at a 14.76% CAGR. Europe held a 10.0% share, while the Middle East and Africa represented 4.8% and Latin America 2.1%.

The Middle East and Africa is the fastest-growing region in the report at 16.60% annually, supported by sovereign AI investment in Saudi Arabia and the United Arab Emirates. At the country level, the United States was the largest market at $96.30 billion, equal to 52.0% of global value. China ranked second at $22.94 billion. India is projected to be the fastest-growing major country market at 20.46% annually, reaching $23.15 billion by 2035.

Competitive Landscape Centers on Merchant GPUs and Custom Accelerators

NVIDIA leads the AI Chip Market with an estimated 72% of 2025 value, according to DC Market Insights. AMD remains the principal merchant GPU challenger, while Broadcom is the largest custom AI ASIC partner in the firm's estimate. Google, Amazon Web Services and Microsoft increasingly shape demand through in-house accelerators, while Intel, Qualcomm, Huawei, Mobileye, Horizon Robotics, Marvell Technology, Cambricon Technologies and Hailo compete across data center, automotive and edge applications.

Recent supplier disclosures underscore the scale of demand. NVIDIA reported $89.00 billion in data center revenue for the quarter ended July 26, 2026, up 117% from a year earlier. Microsoft introduced its Maia 200 inference accelerator in January 2026, and Qualcomm is entering data center inference with AI200 and AI250 products. The competitive frontier is therefore broadening from peak training performance toward inference efficiency, cost per token, memory capacity, software compatibility and supply certainty.

2035 Forecast Range Spans $528.13 Billion to $823.51 Billion

DC Market Insights' base case places the AI Chip Market at $677.59 billion in 2035. A low case, which assumes slower delivery of AI power capacity, flat real capital spending after 2028 and weaker edge and device demand, produces a $528.13 billion market. A high case in which grid capacity and HBM supply keep pace with announced programs and edge demand exceeds the base model reaches $823.51 billion.

The scenario range highlights the market's dependence on infrastructure outside the chip itself. Power delivery, data center construction, high-bandwidth memory and advanced packaging can constrain accelerator deployments even when end-user demand remains strong. For investors and suppliers, this makes AI chip growth inseparable from the wider data center buildout and from the supply chains supporting each new generation of accelerated systems.

About the AI Chip Market Study

The study covers chips whose primary role is AI computation, including data center GPUs, custom AI ASICs and SoCs, client NPUs, FPGAs and other AI accelerators sold for data centers, vehicles, consumer devices and industrial edge systems. The report uses 2025 as the base year, 2020-2024 as the historical period and 2026-2035 as the forecast period.

DC Market Insights built the model bottom-up from unit and pricing assumptions across data center accelerators, automotive AI SoCs, client NPUs and industrial or edge AI chips, then reconciled the totals against supplier disclosures and regional deployment indicators. The report was written by Amit Jain and reviewed by Deepti Agrawal, Senior Editor, Research.

Get Free Sample Report  - https://www.dcmarketinsights.com/report/ai-chip-market

Related Reports from DC Market Insights

• Data Center Accelerator Market — $169.77 billion in 2025 to $605.42 billion by 2035; 13.56% CAGR.
• Data Center GPUs Market — $153.59 billion in 2025 to $679.88 billion by 2035; 16.04% CAGR.
• AI Server Market — $246.33 billion in 2025 to $1,098.47 billion by 2035; 16.13% CAGR.
• Data Center Chip Market — $314.87 billion in 2025 to $1,021.48 billion by 2035; 12.49% CAGR.
• AI Data Center Market — $21.74 billion in 2025 to $123.60 billion by 2035; 18.91% CAGR.

About DC Market Insights

DC Market Insights is the data center research and consulting practice of Credence Research, founded in 2015. The firm sizes markets, forecasts them to 2035 and advises investors, operators, vendors and governments on commercial due diligence, site selection, market entry and power strategy, market sizing, competitive intelligence and policy and investment decisions. DC Market Insights is supported by more than 200 analysts and consultants and the wider organization completes more than 450 consulting projects a year.

Its research spans the full data center value chain, including facilities, power, cooling, construction, AI compute, cloud, interconnection, software, services, storage and networking. Each report carries the name of the analyst who built it and is reviewed by a second analyst or editor before publication.

Media and Research Contact

DC Market Insights

Email: [email protected]

United States: +1 628 262 7656

United Kingdom: +44 7453 598 606

Office: Tower C-1105, S 25, Akash Tower, Vishal Nagar, Pimple Nilakh, Pune, MH 411027, India

Contact: www.dcmarketinsights.com/contact-us

DC Market Insights Data & Tools

Explore DC Market Insights proprietary data and decision-support tools: Data Center Map | Data Center Operators | Deal Tracker | Power & Policy Tracker | Data Center PUE Calculator

SOURCE DC Market Insights