TLDR: A new forecast from Omdia reveals that the AI data center chip market is set to expand significantly, reaching $286 billion by 2030. However, the rapid growth observed in recent years is expected to slow beyond 2026, primarily due to the increasing adoption of custom ASICs like Google’s TPUs and advancements in AI model efficiency. While GPUs remain dominant, alternative chip solutions are gaining market share.
According to Omdia’s latest ‘AI Processors for Cloud and the Data Center Forecast,’ the artificial intelligence data center chip market is on a trajectory to reach an impressive $286 billion by the year 2030. This projection highlights the continued robust expansion of AI infrastructure. However, the report also signals a deceleration in the market’s hypergrowth, with signs indicating a slowdown in expansion beyond 2026.
The boom in AI infrastructure has led to a substantial upward revision of previous forecasts. In 2024, an estimated $123 billion worth of GPUs and AI accelerators were shipped. This figure is projected to climb to $207 billion in 2025, before reaching the $286 billion mark by 2030. While the market experienced an extraordinary growth of over 250% year-on-year between 2022 and 2024, the growth rate from 2024 to 2025 is anticipated to be around 67%.
Omdia predicts that AI infrastructure spending, as a proportion of total data center expenditure, will peak in 2026. During this period, virtually all incremental spending within data centers will be driven by AI, before gradually tapering off towards 2030.
Alexander Harrowell, Principal Analyst for Advanced Computing at Omdia, identified several key drivers fueling this growth. These include ‘the adoption of AI applications, the increasing democratization of AI fine-tuning, and the use of reasoning models, which generate large numbers of tokens that are not used in the final output.’ Conversely, Harrowell also pointed to factors that are contributing to the slowing demand for AI computing. These include ‘the shift towards smaller, specialized models’ and ‘generation-on-generation improvements in AI model efficiency.’ He elaborated that these improvements encompass ‘better curation of training datasets, advances in fundamental model design, and operational optimizations in inference.’
While NVIDIA currently maintains a dominant position in the market, Omdia’s observations for 2025 indicate a growing traction for alternatives to traditional GPUs. This includes the rise of custom ASIC chips, such as Google’s Tensor Processing Units (TPUs), as well as merchant Application-Specific Standard Products (ASSPs) from companies like Huawei (Ascend series), Groq, and Cerebras. AMD has also made notable strides with its Instinct line of GPUs, bolstered by significant software investments made in 2024.
Also Read:
- Nvidia’s China AI Chip Market: Navigating Export Controls, Production Halts, and Revenue Sharing
- Global AI Investments Projected to Soar 31.9% Annually, Reaching $1.3 Trillion by 2029
The comprehensive Omdia forecast covers a wide array of AI processing hardware, including GPUs, AI ASICs, ASSPs, FPGAs tailored for AI applications, and CPUs equipped with integral AI acceleration hardware. The report provides detailed estimates segmented by market, global region, vertical industries, specific applications, memory content, compute performance, and power consumption bands.
