- Microsoft (MSFT) has dropped data center projects totaling 2 gigawatts in the U.S. and Europe over the last six months due to an oversupply relative to demand, opting not to support extra OpenAI workloads.
- Alphabet’s Google and Meta Platforms are absorbing the excess capacity internationally and in the U.S., respectively, as Microsoft maintains its $80 billion AI infrastructure spend this fiscal year.
- Investor doubts grow over U.S. tech’s hefty AI investments – Alphabet at $75 billion and Meta up to $65 billion – amid DeepSeek’s low-cost AI reveal in January, though Microsoft insists on strong regional growth despite earlier cuts of a couple hundred megawatts.
Microsoft’s (MSFT) decision to abandon data center projects totaling 2 gigawatts of electricity capacity across the U.S. and Europe over the past six months reflects a strategic recalibration in its AI infrastructure ambitions, driven by an oversupply relative to current demand forecasts, according to a Reuters report citing TD Cowen analysts. The $2.94T market cap software giant, which saw its shares dip 1.5% to $389.17 on Wednesday, has scaled back new capacity leasing, a move largely attributed to its choice not to support additional training workloads for OpenAI, the ChatGPT developer it has long backed. This shift comes amid growing investor skepticism about the massive AI investments by U.S. tech firms, with the emergence of Chinese startup DeepSeek in January showcasing cost-effective AI technology that has raised questions about the necessity of such extensive spending.
The pullback has created opportunities for competitors, as TD Cowen’s supply chain checks reveal Alphabet’s Google stepping in to absorb excess capacity in international markets, while Meta Platforms fills the gap in the U.S. Microsoft, however, remains steadfast in its broader vision, asserting that its $80 billion AI infrastructure budget for this fiscal year remains intact, with a spokesperson noting, “while we may strategically pace or adjust our infrastructure in some areas, we will continue to grow strongly in all regions.” This follows an earlier report from TD Cowen in February, which highlighted Microsoft’s cancellation of leases for a couple of hundred megawatts with at least two private data center operators, signaling a pattern of reassessment that predates the latest 2-gigawatt reduction.
The broader AI landscape underscores the stakes, with Alphabet committing $75 billion to its AI buildout this year – 29% above Wall Street’s expectations – and Meta pledging up to $65 billion, both defending their investments as essential to maintaining a competitive edge despite DeepSeek’s disruptive low-cost model. Microsoft and Meta executives have echoed this sentiment, arguing post-January that such spending is non-negotiable in the race to dominate AI innovation. Meanwhile, Nvidia-backed AI cloud startup CoreWeave, a key Microsoft customer, reported no contract cancellations this month, countering a Financial Times claim and suggesting that not all partners view Microsoft’s adjustments as a retreat. As the industry navigates this tension between ambition and efficiency, Microsoft’s measured step back from 2 gigawatts of capacity highlights a pragmatic response to an evolving market, balancing its long-term AI goals with immediate demand realities.
WallStreetPit does not provide investment advice. All rights reserved.
Leave a Reply