The Chips Powering AI Could Nearly Triple to $443 Billion by 2030
A new market outlook sees AI accelerators growing fast as data centers and edge devices demand more specialised silicon.
The market for AI accelerator chips — the specialised processors that train and run artificial intelligence models — is projected to nearly triple by the end of the decade, growing from $151 billion in 2025 to $443 billion by 2030, according to a recent report cited by Yahoo Finance.
The forecast reflects how central these chips have become to the AI boom. As generative AI spreads and large language models are deployed more widely, demand for graphics processing units and neural processing units is rising across both cloud data centers and edge devices, the report finds.
Data centers are driving the spend
North America currently leads the market, powered by hyperscaler spending on data-center infrastructure. The largest cloud providers have been pouring capital into the facilities and hardware needed to train and serve AI models, and accelerators sit at the heart of that buildout.
Asia Pacific is the fastest-growing region, according to the report, driven by China's AI-chip design capabilities and India's computing initiatives. The regional split underscores that the race to build AI infrastructure is now global, with governments as well as companies investing in domestic chip capacity.
Substantial investment in data-center and edge infrastructure is one of the main forces behind the projected growth, alongside the wider adoption of generative AI. As more AI workloads move closer to where data is generated — in devices, factories and vehicles — demand for accelerators designed for edge environments is expected to rise alongside cloud demand.
A market with risks attached
The outlook is not without caveats. The report notes potential headwinds from economic constraints and from technological obsolescence — a real risk in a field where each new generation of chips can quickly make the last one look slow. Simply Wall St, which published the analysis, notes that its coverage is general in nature, based on historical data and analyst forecasts, and is not financial advice.
The projection lands in a week of choppy trading for the big technology names. According to the same report, Microsoft finished at $535.93, down 3.6%, and Alphabet settled at $282.57, down 2.2%, while Apple closed at $344.17, down 1.5%, after announcing fixed-income offerings including notes due in 2028. EPAM Systems rose 20.6% to $165.70 after announcing a share-repurchase program of up to $1.8 billion running to the end of 2029, while Sandisk fell 22.3% to $806.43.
The market swings are a reminder that the infrastructure buildout behind AI is being financed in public markets, where sentiment can turn quickly even as the long-term demand picture points upward. If the projection holds, the companies that design and manufacture accelerators will be supplying one of the fastest-growing hardware markets of the decade.
Written by the BidAsk Newsroom. Based on reporting from the sources below.
