Google is in preliminary talks with Samsung Electronics to manufacture a critical piece of its next-generation AI chip, a move that could reshape the pecking order in the semiconductor foundry business.

The negotiations center on a memory input-output die for Google’s upcoming Tensor Processing Unit, codenamed Icefish. Samsung would produce the component using its advanced 2-nanometer process technology, one of the most cutting-edge fabrication methods in existence.

Why Google is looking beyond TSMC

For years, TSMC has been the go-to manufacturer for virtually every major tech company designing custom silicon. Google’s TPUs have been no exception. But capacity constraints at TSMC are forcing Google to think creatively about its supply chain.

Google’s TPUs have been powering AI workloads across its data centers since 2016. Nearly a decade in, the company’s appetite for custom silicon has only grown as it pours resources into generative AI, search improvements, and cloud computing services.

Samsung’s aggressive play for AI chip market share

Samsung is planning to invest over 110 trillion won, roughly $73 billion, in 2026 alone. That money is earmarked for AI semiconductor facilities and research and development.

Samsung is already landing major clients. In July 2025, the company announced a $16.5 billion multi-year deal with Tesla to supply AI6 chips.

Samsung has historically been known more for its memory chips, DRAM and NAND, than for its logic foundry business. TSMC has dominated contract manufacturing for advanced processors, capturing the lion’s share of orders from companies that design but don’t fabricate their own chips. Samsung’s transition into advanced contract manufacturing is an attempt to change that narrative, and the Google discussions fit squarely into that strategy.

What this means for investors

For investors watching the semiconductor space, the Google-Samsung discussions point to a structural shift worth paying attention to. TSMC has enjoyed something close to a monopoly in advanced chip manufacturing for the better part of a decade. If major customers like Google begin splitting orders between TSMC and Samsung, that duopoly dynamic could compress TSMC’s pricing power over time.

Samsung’s stock and strategic positioning stand to benefit if it can reliably deliver on 2nm manufacturing at scale. The Tesla deal already demonstrates that major clients are willing to bet on Samsung for mission-critical AI silicon.

Investors should watch two things closely. First, whether Samsung can actually deliver competitive yields on its 2nm process. Yield rates, the percentage of functional chips produced per wafer, are the make-or-break metric in semiconductor manufacturing. Second, keep an eye on whether other major chip designers follow Google’s lead. If Amazon’s Annapurna Labs or Meta’s custom silicon teams start diversifying toward Samsung, that would confirm a genuine market shift rather than a one-off deal driven by specific capacity needs.

The talks remain in early stages, and there’s a meaningful gap between preliminary negotiations and a signed production agreement.