A PYMNTS Company

Google’s $12.2 Billion Marvell Deal Deepens Ties Across AI Supply Chain

 |  August 19, 2026
Google search

A sweeping chip agreement between Google and Marvell Technology is putting a fresh spotlight on the increasingly close financial relationships between powerful artificial-intelligence companies and the suppliers they rely on, potentially adding another dimension to competition concerns surrounding the AI industry.

    Get the Full Story

    Complete the form to unlock this article and enjoy unlimited free access to all PYMNTS content — no additional logins required.

    yesSubscribe to our daily newsletter, PYMNTS Today.

    By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions.

    Under the agreement, Marvell will help develop custom chips for Google and has granted the technology company a warrant allowing it to acquire as many as 58.97 million Marvell shares for $206.58 apiece, according to Reuters. The shares would be worth about $12.18 billion if the warrant were fully exercised, Reuters reported.

    The arrangement could make Google Marvell’s fifth-largest investor while simultaneously deepening the commercial relationship between the two companies, according to Reuters. Marvell could receive roughly $120 billion in revenue through fiscal 2033 if Google meets targets tied to the warrant, the news organization reported.

    Such arrangements are attracting attention because the AI boom is creating a web of relationships in which the biggest buyers of computing infrastructure can also become significant financial stakeholders in companies supplying that infrastructure. While the Reuters report did not say regulators had opened an antitrust investigation into the Google-Marvell agreement, it said the transaction could intensify concerns about increasingly interconnected relationships across the AI sector.

    Those links are emerging as technology companies spend heavily to secure the chips, networking equipment and data-center capacity required to develop and operate AI systems.

    Read more: French Publishers Seek Antitrust Action Against Google Over AI Search Summaries

    Google has increasingly relied on its own tensor processing units, or TPUs, as an alternative to graphics processors supplied by Nvidia. Demand for such internally designed processors has grown as companies look for lower-cost computing options and chips optimized for inference — the process of running already-trained AI models — according to Reuters.

    The Marvell agreement covers multiple technologies associated with Google’s TPU infrastructure, including processors involved in operating AI models as well as components used for data storage and networking, Reuters reported. That broad scope could give Marvell a larger role in one of the most important pieces of Google’s AI infrastructure.

    The deal may also reshape competition among Google’s suppliers. Broadcom has been Google’s principal partner for custom chips, according to Reuters, and its shares dropped more than 5% after the Marvell agreement was disclosed. Marvell climbed nearly 8%, while shares of Google parent Alphabet were little changed.

    Still, the agreement does not necessarily indicate that Google is replacing Broadcom. Morningstar analyst William Kerwin described the transaction to Reuters as a significant victory for Marvell but said he viewed it as evidence that Google is expanding the number of suppliers participating in a growing market rather than simply shifting business away from Broadcom.

    Related:

    From a competition perspective, that distinction could matter. Adding another chip supplier may strengthen competition for Google’s business. At the same time, a customer taking a potentially multibillion-dollar equity position in a major supplier illustrates how traditional boundaries between buyers, vendors and investors are becoming less distinct as AI infrastructure spending accelerates.

    The structure isn’t unprecedented. AMD agreed last year to supply OpenAI with AI processors in an arrangement potentially representing tens of billions of dollars in annual revenue while also giving OpenAI an option to acquire roughly 10% of AMD, according to Reuters. The news organization also reported that Nvidia recently agreed to provide a financial backstop of as much as $105 billion for an Ohio data-center project being leased by OpenAI.

    Those transactions underscore a developing challenge for competition authorities: determining whether investments and supply agreements primarily create new capacity and competition or whether financial interdependence among leading AI companies and infrastructure providers could ultimately reduce their incentives to compete independently.

    For Google, the Marvell partnership comes as custom processors become more important to its AI strategy. Reuters reported that a recent reorganization of Google’s AI operations increased the influence of executives closely connected with Google Cloud, bringing additional attention to the chips and infrastructure underpinning the company’s AI services.

    For Marvell, the economics are potentially transformative. The prospective revenue through 2033 and Google’s ability to become one of its largest shareholders give the chipmaker both a substantial commercial opportunity and a closer relationship with one of the world’s largest technology companies.

    Whether that relationship ultimately draws regulatory scrutiny remains uncertain. What is already clear is that the race for AI computing power is producing corporate alliances that extend beyond ordinary supplier contracts — and that those ties are likely to become an increasingly important part of the competition-policy debate.

    Source: Reuters