As U.S. weighs profit-sharing in AI firms, Nvidia-linked supply worries push Google and Nvidia to explore Intel

Market coverage on Monday highlighted two forces that could shape the AI industry in the near term: a U.S. push to consider government financial stakes in major AI companies, and signs that some of the biggest AI chip buyers are actively mapping out alternatives as Taiwan Semiconductor Manufacturing Co. struggles to keep up with demand.

According to Reuters reporting carried by, senior U.S. officials have held preliminary discussions with major AI companies about the government potentially buying equity in them. The concept described in that reporting centers on the firms voluntarily ceding shares to the government, with returns directed toward public purposes, including the possibility of a dividend-like payout to American households.

On June 5, 2026, U.S. President Donald Trump told reporters his team is looking into the idea and said he intends to meet with AI executives at the White House “probably next week.” Another Reuters report, also republished by, framed the proposal as a form of partnership with the American public and noted that OpenAI and Anthropic are preparing for major IPO milestones, while the administration continues to emphasize cybersecurity testing of advanced AI models before public release.

On the supply side, Mobile World Live, citing The Information, reported that Alphabet’s Google has placed an order with Intel to manufacture more than 3 million specialized AI chips in 2028. The same report said Google plans to use Intel for at least some of its tensor processing units, after testing Intel’s manufacturing capabilities, and that the shift reflects mounting strain on TSMC’s ability to meet surging demand for advanced foundry capacity.

The implications for Nvidia are direct, even if Nvidia itself has not publicly confirmed any change in fabrication strategy. Mobile World Live also reported that The Information said Nvidia is evaluating Intel’s manufacturing technology for a forthcoming processor design that could combine multiple graphics chips into a single unit. TipRanks, summarizing The Information’s reporting, similarly said that Google and Nvidia are “quietly turning to Intel” as a backup manufacturer for their most advanced processors, attributing the move to TSMC capacity pressure.

Nvidia’s business model depends on access to advanced semiconductor manufacturing, even though Nvidia designs the chips. In its FY2025 annual report, Nvidia described reliance on third-party foundries for semiconductor fabrication, naming Taiwan Semiconductor Manufacturing Company as an example supplier. The filing also warned that disruptions at upstream suppliers could affect the availability of Nvidia’s solutions, lead times, and financial results. Nvidia has also continued to emphasize onshoring, including a company blog post describing the celebration of the first NVIDIA Blackwell wafer produced on U.S. soil through TSMC Arizona, underscoring the strategic importance of capacity expansion even as supply constraints persist.

Still, major details remain unresolved. The U.S. equity-stake discussions are described as preliminary and voluntary in the Reuters-based reporting, and Reuters said it could not immediately confirm the report about the specific equity-stake mechanism. On the chip front, the Intel backup framing comes from reporting summaries and third-party accounts, and Nvidia did not provide additional public confirmation about the status, timing, or scale of any Intel-linked manufacturing plan.

What to watch next is whether the White House or affected AI companies move from concept to specifics, such as the structure of any government equity purchases, the governance terms, and the timeline tied to upcoming IPOs. For semiconductors, investors and customers are likely to look for clearer indicates on whether Nvidia and other AI chip buyers formalize Intel contingency manufacturing, and whether TSMC’s capacity ramps are sufficient to reduce the perceived need for backup foundry plans.