What's happening
According to a June 8, 2026 report from The Information, Google has placed an order with Intel Foundry to manufacture more than 3 million Tensor Processing Units (TPUs) in 2028. The order represents approximately half of Google's projected TPU output for the 2027–2028 period, during which the company is expected to produce a total of 6 million TPUs. The development marks a concrete step toward supply chain diversification for one of the world's largest consumers of custom AI silicon, with Intel Foundry positioned as an alternative manufacturing source alongside the dominant incumbent, TSMC.
In a separate but related development, Nvidia is reported to be testing Intel's manufacturing technology for the development of processors that integrate four GPUs into a single package. Nvidia, which carries a market capitalization of approximately $5.04 trillion and reported revenue of $253.49 billion, has historically relied on TSMC for the fabrication of its high-end data center GPUs, including the A100 and H100 series. The simultaneous exploration by both Google and Nvidia signals a broader industry evaluation of Intel Foundry as a viable production alternative for advanced AI-oriented silicon.
Why it matters for markets
Intel's foundry business has been a central strategic priority for the company as it works to compete with TSMC and Samsung in contract chip manufacturing. The Google TPU order — covering more than 3 million units in 2028 alone — represents a tangible, large-scale commercial commitment to Intel Foundry at a time when AI infrastructure investment is driving unprecedented demand for advanced semiconductors. Intel reported annual revenue of $53.76 billion and carries a market capitalization of $542.41 billion; the foundry segment has been a key area of investment as the company seeks to diversify beyond its traditional CPU business. Intel shares rose more than 13% on the day the report was published, reflecting the scale of market attention on the development.
For Alphabet, which reported revenue of $422.50 billion and operates Google Cloud as a core growth business, securing a second manufacturing source for its custom TPU hardware reduces concentration risk in its AI infrastructure supply chain. TPUs are purpose-built accelerators central to Google's internal AI workloads and its cloud offerings. Allocating half of a projected 6 million-unit production run to Intel Foundry rather than a single supplier represents a structural shift in how the company manages chip procurement at scale.
The broader implication for the semiconductor supply chain is the potential erosion of TSMC's near-monopoly position in advanced AI chip fabrication. If Intel Foundry can demonstrate the process reliability and yield necessary to fulfill the Google order and satisfy Nvidia's technical evaluation, it could alter the competitive dynamics of the foundry market — a segment that has grown in strategic importance as governments and corporations alike seek to reduce dependence on geographically concentrated manufacturing.
Sectors and assets to watch
Intel (INTC) is the primary ticker to monitor, as the Google order and Nvidia evaluation directly concern its Intel Foundry Services business. The company's 52-week range of $18.97 to $132.75 reflects the significant volatility Intel has experienced as investors have assessed its foundry ambitions against execution challenges. The current price of $107.92 and market cap of $542.41 billion provide context for the scale of the market response to the June 8 report. Any further announcements regarding foundry customer wins, process node milestones, or yield data will be closely watched by market participants.
Alphabet (GOOGL) and Nvidia (NVDA) are also directly named in the reported developments. TSMC, though not publicly traded on U.S. exchanges under that ticker, is the incumbent supplier whose market position is most directly affected by this reported diversification. Broader semiconductor equipment and materials companies that support advanced foundry processes — as well as other hyperscale cloud providers that rely on custom silicon — may also face indirect implications as the industry evaluates whether Intel Foundry can scale to meet demand from multiple large customers simultaneously.
What to watch next
Key developments to monitor include any official confirmation from Google, Nvidia, or Intel regarding the reported orders and technology evaluations, as the sourcing for this story remains a single publication. Investors and analysts will be watching for Intel's process node readiness and whether its manufacturing technology can meet the specifications required for advanced AI accelerators at the volumes described. Progress disclosures in Intel's quarterly earnings reports — particularly any updates on foundry revenue, customer pipeline, and capital expenditure commitments — will provide measurable indicators of whether the reported momentum is translating into financial results. Any response from TSMC regarding its own capacity commitments or pricing arrangements with Google and Nvidia will also be relevant to assessing the pace and depth of any supply chain shift.