A Blow to Nvidia? What Anthropic's IPO Prospectus Revealed About the Chipmaker's Dominance

In its IPO filing, the AI company Anthropic disclosed that it uses semiconductors from four different manufacturers—Amazon, Google, Nvidia, and AMD—for its computations. This multi-chip strategy could pose a threat to Nvidia’s monopoly and the company’s pricing power in the market, according to some analysts surveyed by Business Insider.

The AI lab Anthropic manages its computing resources “as a single, interchangeable pool from several strategic partners,” according to the company’s filing with the regulator in preparation for its initial public offering. This diversification is achieved through software that distributes tasks across different chips depending on where they are most efficient.

According to Anthropic, building this flexibility “took years and significant investment.” Over time, it could make major AI players less dependent on Nvidia and give them more leverage in negotiations, analysts warn.

“Nvidia’s pricing power could be under real threat,” Patrick Moorhead, CEO of Moor Insights & Strategy, told Business Insider. “A customer capable of distributing workloads across four platforms negotiates in a completely different way than one who cannot.”. However, he adds that switching between chips remains difficult, and only the largest companies in the market have the resources to do so. Furthermore, Anthropic continues to spend billions on computing power from Nvidia itself, the expert emphasized.

Other giants, such as OpenAI, are likely also using a multi-chip approach, but for smaller players, such investments and efforts may simply not pay off, agrees Gil Luria, head of technology research at D.A. Davidson. However, he did not include Nvidia among the main beneficiaries of Antrophic’s upcoming IPO. According to Luria’s assessment, the key beneficiaries of the offering will be semiconductor manufacturer Broadcom, which produces chips for Google, as well as Google itself and Amazon.

The multi-chip approach makes Anthropic less vulnerable to fluctuations in Nvidia's hardware supply, noted IDC analyst Ashish Nadkarni.

Brad Gasworth, head of global research and market analysis at Circular Technology, believes that Anthropic’s strategy does not pose a serious short-term threat to Nvidia. He is confident that having a wider selection of processors does not negate the chipmaker’s advantages in performance, networking, and software.

A more significant risk for Nvidia lies in the possibility that AI computing could become a standardized, mass-market commodity. If supply exceeds demand, this will lead to a decline in prices for renting computing power, warned Citadel CEO Ken Griffin. “Graphics processing units, like airlines, have very low marginal operating costs,” he told Bloomberg last week. “As soon as you have excess capacity, prices start to plummet.”

Nvidia did not respond to Business Insider's request for comment.

In its IPO filing, Anthropic also acknowledged its dependence on the three largest tech giants—Google, Amazon, and Microsoft. Sales through these companies’ cloud platforms accounted for 48% of the AI lab’s revenue in the first half of 2026. At the same time, these providers have become, for the startup, investors, computing power suppliers, a sales channel, and direct competitors all at once. In its prospectus, the company explicitly identifies this combination of roles as the main source of risk.

This article was AI-translated and verified by a human editor