ClaimsSatya Nadella

“we’ve been at this in a variety of different forms for a long, long time in terms of building our own silicon.”

Satya Nadella, Microsoft chairman and CEO, on earnings call 28 January 2026 Microsoft FY2026 Q2 Earnings Conference Call (Microsoft Investor Relations) archived copy

Factual AI compute, hardware and data

Microsoft has been building its own silicon in various forms for a long time.

Mostly true True in substance, wrong in a detail.

Factual: Statements about the present or the past, checked against the record: did they hold up?

The second judge said resolved true; the verdict shown is the first judge's.

Said for an organisation they lead or work for; it counts on both records.

What would count as true or false

Judged as of 2026-01-28. True if the record shows that Microsoft has designed or built its own silicon in various forms over a long span (many years, e.g., at least about a decade), such as Xbox custom chips, HoloLens HPU, Azure FPGA (Project Catapult/Brainwave), Pluton security processor, Azure Boost/Cobalt/Maia chips. Mostly true if Microsoft has a substantial custom-silicon history but its in-house design work is somewhat shorter or narrower than 'a long, long time' implies (e.g., mainly co-designed chips for several years). Mixed if the history is limited to a few scattered projects of disputed in-house status. Mostly false if Microsoft's own silicon efforts only began recently (e.g., within the last 2–3 years) and the earlier work was mostly buying third-party chips. False if Microsoft had no in-house or custom silicon efforts before recent years.

Set before the research began.

Evidence (4)

  1. E1 Project Catapult - Microsoft Research Microsoft Research · 7 December 2021 Own record

    “Project Catapult began in 2010 when a small team, led by Doug Burger and Derek Chiou, anticipated the paradigm shift to post-CPU technologies.”
  2. E2 Microsoft Infrastructure - AI & CPU Custom Silicon Maia 100, Athena, Cobalt 100 SemiAnalysis · 15 November 2023

    “It should be noted that while Microsoft is currently behind on deploying custom silicon in their datacenters relative to Google and Amazon, they have a long history of silicon projects.”
  3. E3 Microsoft Infrastructure - AI & CPU Custom Silicon Maia 100, Athena, Cobalt 100 SemiAnalysis · 15 November 2023

    “Microsoft has also internally developed multiple generations of root of trusts that are found on every single server they install in their datacenters.”
  4. E4 Microsoft Reveals Its First Custom Processors ExtremeTech · 15 November 2023

    “But while other companies invested in custom chip designs, Microsoft has not—until now.”

Reasoning

Microsoft Research's own project page dates Project Catapult, its datacenter FPGA work, to 2010 (E1). An independent analysis (E2, E3) says Microsoft has "a long history of silicon projects," including several generations of in-house root-of-trust chips on every datacenter server. However, ExtremeTech described the 2023 Maia and Cobalt chips as Microsoft's first custom processors (E4). That suggests its in-house processor design work is narrower than "a long, long time" implies, even though its wider silicon work goes back more than a decade.

Cite this claim

Satya Nadella, “Microsoft has been building its own silicon in various forms for a long time.” (said 28 January 2026). Verdict: mostly true. Reliability API, claim c_0445f0cb5d43, method 0.1.0+28da144b, dataset release v1.1.2. https://epistacklabs.com/claims/c_0445f0cb5d43

This claim is in dataset release v1.1.2, with the same verdict.

Reply to this claim

For Satya Nadella, or someone speaking for them: a correction or counter-evidence. Replies are read before anything is shown; an approved reply shows your name, role, date, statement and link. Your email address is never shown and is deleted 30 days after the decision (privacy).

More factual claims by Satya Nadella

  1. Mixed Factual 28 January 2026
    “same thing in inferencing as well.”

    AI inferencing, like AI training, requires a lot of compute and a lot of storage located close to that compute.

  2. Resolved true Factual July 2025
    “Through software optimization alone, we are delivering 90% more tokens for the same GPU compared to a year ago.”

    Through software optimization alone, Microsoft delivers 90% more GPT-4o-family tokens per GPU than a year ago.

  3. Mostly true Factual 15 November 2023
    “right as we speak, we have some of the Maia stuff powering GitHub copilot for example.”

    As of November 2023, some Maia chips are powering GitHub Copilot.

  4. Resolved true Factual 15 November 2023
    “we built an entire rack which is liquid cooled for Maia.”

    Microsoft built an entire liquid-cooled rack for its Maia AI chip.

  5. Mixed Factual 4 June 2026
    “I was able to use a model that is using 500B, I was able to use a 5B, and have it really deliver the same outcome, why would I not use that?”

    In a Land O'Lakes agent example shown at Build, a 5B-parameter model delivered the same outcome as a 500B-parameter model.

All of them

Verdicts are made automatically; how verdicts are made.