Anthropic Eyed $7B MatX Deal to Build Custom AI Chips

A Bold Bet on Custom Silicon

Anthropic reportedly explored acquiring MatX, a small but closely watched AI chip startup, in a deal valued at roughly $7 billion. The talks, if finalized, would mark one of the largest chip-related acquisitions in the AI industry and underscore how aggressively leading model developers are moving to control their own hardware destiny rather than relying solely on outside suppliers 1. For Anthropic, the rationale is straightforward: as its AI operations scale, having tailored silicon could reduce dependence on general-purpose chips and give the company more leverage over cost, performance and supply in an increasingly hardware-constrained market 1.

Why Chips Have Become the New Battleground

The reported interest in MatX fits into a broader pattern of AI labs racing to secure computing power however they can, whether through in-house design, strategic investment, or acquisition. Nvidia, still the dominant supplier of AI accelerators, has reportedly paused new revenue-sharing arrangements with cloud partners amid antitrust scrutiny, a sign that the financing and dealmaking structures underpinning AI infrastructure are drawing regulatory attention even as demand for compute keeps climbing 4. Against that backdrop, a frontier lab like Anthropic seeking to buy a chip startup outright — rather than simply licensing or leasing capacity — reflects an industry-wide push to hedge against bottlenecks and pricing pressure from a concentrated chip supply chain.

Capital Is Flowing Differently Across the AI Ecosystem

The MatX talks arrive at a moment when venture capital's relationship with AI is being reshaped in multiple directions at once. OpenAI has launched its own venture fund, committing $400 million as the sole investor to back early-stage AI companies, a move that blurs the line between AI labs and traditional venture firms and gives incumbents more influence over which startups get funded 3. Meanwhile, traditional VCs are reportedly conducting urgent portfolio reviews, recognizing that deals struck in 2024 may already look mistaken given how fast AI capabilities and valuations are moving 2. Broader trends point to 2026 venture activity being dominated by AI megarounds and defense-tech investment, with access to computing power emerging as a decisive factor in which startups thrive 6.

Wealth and Risk Concentrate Around AI

The appetite for AI exposure extends beyond venture funds. Wealthy investors have reportedly been increasing their stock allocations toward AI giants like Microsoft, even as private-market vehicles — venture capital, private credit and real estate funds — have recorded markdowns 5. Taken together, the MatX talks, Nvidia's shifting deal structures, OpenAI's fund, and cautious VC repositioning all point to an AI sector where capital, chips and startups are becoming ever more tightly, and riskily, intertwined.