S&P 5007,785.76▼0.2% Nasdaq26,729.16▼0.3% Dow53,732.41▼0.2% Russell 2K3,068.42▲0.5% 10-Yr4.70%+6bp VIX14.25−0.38 WTI$82.40▲1.4% Gold$4,432.00▲1.6% EUR/USD1.157▲0.4% BTC$63,324▲0.8% Nikkei68,309▲1.2%
At close · Fri, Aug 14, 2026
Daily Market Updates.

US Markets

HomeUS MarketsSectorsMicrosoft has 2.2m AI chips installed, despite 1.8m ta…

Microsoft has 2.2m AI chips installed, despite 1.8m target

A Guardian investigation cites internal documents showing Microsoft installed more than 2 million AI chips, while a reported target of 1.8 million by end-2024 may not reflect current ramp pace.

A Guardian investigation says Microsoft has an apparent mismatch between its stated AI capacity and the number of advanced AI chips the company has in operation in its datacentres worldwide. The outlet reports that Microsoft targeted having 1.8 million AI chips installed by the end of 2024, but internal documents it viewed indicate Microsoft has 2.2 million chips installed nearly two years later, during a major datacentre expansion.

The story frames the issue as a constraint in the global AI build-out, noting that AI chips are central to training and running models and that leading tech firms need large quantities to keep pace. It says the discrepancy could indicate Microsoft’s newest datacentres may not be fully operational, or that they lack chips they need.

The Guardian also highlights that Nvidia makes the microchips used in AI infrastructure, and that its supply chain is tightly held. The outlet says Nvidia generally does not disclose how many chips it sells or to whom, and that clients such as large technology companies also do not reveal their chip holdings, making it difficult for outsiders to track whether AI infrastructure is growing as fast as industry claims.

The article adds that Microsoft says it has built AI infrastructure quickly, citing CEO Satya Nadella’s statement that Microsoft would double its global datacentre footprint by mid-2027. It also notes Microsoft has invested roughly $280 billion since 2022 in land, buildings, and computational infrastructure for AI, including more than $41 billion in the past quarter.

More like this

Sources

Get the close, explained.

One email every trading day: what moved, why it moved, and what's on deck tomorrow. Read in 3 minutes.

Free. Unsubscribe anytime.