Microsoft's new AI chip delayed to 2026: report

Microsoft's next-generation Maia AI chip (Braga) production is delayed to 2026 from an initial 2025 target, reportedly due to design changes and staffing issues. This setback undermines Microsoft's strategy to reduce reliance on external suppliers like Nvidia, especially as the Braga chip is also expected to underperform Nvidia's Blackwell. The delay further intensifies competitive pressure as rivals Google and Amazon continue to advance their own custom AI silicon, highlighting Microsoft's challenge in securing its AI infrastructure and cost efficiencies.

Analysis

Microsoft's proprietary AI chip program has encountered a material setback, with its next-generation Maia chip, codenamed Braga, now delayed until 2026. This delay, reportedly caused by design changes and high staff turnover, directly undermines the company's strategic goal of reducing its significant reliance on costly Nvidia GPUs. The challenge is compounded by reports that the Braga chip, when released, will offer performance considerably below Nvidia's current Blackwell architecture. This dual issue of a delayed timeline and a performance deficit suggests Microsoft's capital expenditures on third-party hardware will remain elevated for longer than planned, potentially pressuring cloud margins. The setback is amplified by the competitive landscape, where rivals Google and Amazon are advancing their own custom silicon; Google unveiled its seventh-generation TPU in April, and Amazon's Trainium3 is expected this year. This divergence highlights a growing competitive gap, positioning Microsoft at a disadvantage in the race for cost-efficient, vertically-integrated AI infrastructure.

Key Decisions for Investors

• Investors in Microsoft should scrutinize upcoming capital expenditure guidance and Azure's gross margins for signs of pressure from continued, high-cost dependency on Nvidia hardware.
• This development reinforces the investment thesis for Nvidia, as the failure of a major customer to develop a viable near-term alternative solidifies its market dominance and pricing power in the AI accelerator space.
• Consider the strengthening competitive positions of Alphabet and Amazon, as their progress in custom silicon could translate into a cost and performance advantage over Microsoft in the cloud and AI services market.
• Monitor for any further updates on Microsoft's chip development timeline or performance benchmarks, as any acceleration or improvement could alter the current negative outlook.