OpenAI Weighs Google TPU Deal to Diversify AI Chips

By Grace Kim

OpenAI is reviewing a purchase of Google-developed Tensor Processing Units, or TPUs (AI accelerator chips), to expand computing capacity and reduce dependence on Nvidia, Bloomberg reported on October 27. Contract terms remain unfinalized, and neither Google nor OpenAI has confirmed the discussions.

What does an OpenAI-Google chip deal change?

For AI infrastructure buyers, the direct answer is supply. If OpenAI completes the purchase, it would gain another source of accelerators beyond Nvidia’s graphics processing units, or GPUs, and accelerate supply-chain diversification in a market centered on Nvidia. The deal would also broaden the computing options available for developing and running AI models.

OpenAI is pursuing that diversification alongside its own chip program. The company has said it plans to begin using Jalapeno, an in-house AI chip developed with Broadcom, later this year. Richard Ho, who leads OpenAI’s in-house chip development team, previously worked on Google’s TPU program.

Google began developing TPUs more than a decade ago to improve the speed and efficiency of its search services. More recently, the company expanded the business into AI computing chips that can replace Nvidia GPUs. TPU distribution is widening as well: Oracle said on October 27 that it plans to offer Google TPUs through its cloud infrastructure.

Amin Vahdat, a Google executive, said: “Next-generation AI development requires a variety of computing resources.” That supports OpenAI’s stated strategy of combining Google-designed accelerators with its Broadcom-developed chip while reducing reliance on Nvidia. It does not settle the commercial questions, because the companies have disclosed neither finalized terms nor a confirmed agreement.

A completed purchase would ease the market’s Nvidia-centered structure by adding another accelerator platform and broadening access to Google TPUs through cloud distribution. The report gives no purchase volume, contract value or deployment schedule, so buyers cannot yet quantify the shift.

Crypto-market read-through

The accompanying forecast treats the semiconductor review as news about AI supply-chain diversification, with any crypto-market inflow expected to remain indirect. Its short-term rating is Neutral, at 44% confidence.

The report cites a strong U.S. dollar, a 24-hour decline in total crypto market capitalization, and Bitcoin’s 59.42% market share as confirming a risk-off trend, or reduced appetite for riskier assets.

If the stock-market rally continues, the forecast calls for a limited crypto rebound. If the strong dollar and weak altcoins (cryptocurrencies other than Bitcoin) persist, it expects sideways trading with a downward bias.

The mid-term rating is also Neutral, at 46% confidence. The report links expanded AI computing resources and improved cost efficiency with favorable investor sentiment toward technology stocks, but says actual demand for cryptocurrencies remains unconfirmed. Declining gold prices and stable Treasury yields buffer risk aversion, while Bitcoin’s rising dominance signals a delay in capital diversification across altcoins.

If broader AI investment lifts risk appetite in the stock market, the model expects a moderate crypto improvement. Renewed dollar-led liquidity absorption would instead produce selective weakness.

The long-term rating turns Positive, at 49% confidence. The forecast expects OpenAI and Google’s semiconductor diversification to expand AI infrastructure over time and strengthen the basis for technology adoption across the digital-asset industry.

If the shift spreads computing costs and expands cloud options, it could indirectly support blockchain development. Wider productivity gains and recovering risk appetite could drive a structural crypto uptrend, but entrenched macroeconomic tightening and a strong dollar would limit the upside.

For infrastructure buyers, the established fact is a reported review, not a purchase. If talks become a contract, the change concerns OpenAI’s supplier mix and capacity options; price, volume and launch date remain unpublished.