Bubble Fears Not Spoiling Billion-Dollar AI Buildout Plans in ’26 … Yet

Up, up and away? The companies expected to pour more than $500 billion into hyperscaling artificial intelligence this year (including Magnificent 7 members such as Microsoft, Amazon and Google) are still betting big on the technology following that trajectory.

The less appealing alternative is up, up and pop, manifesting bubble fears that were inflamed when the Financial Times reported in December that Oracle had lost its “primary backer” for a $10 billion Michigan data center. Afterward, the cloud computing giant’s shares sank 5.4% and are down 32% in the past three months, weighing on other AI-related stocks, including Nvidia, Broadcom and Advanced Micro Devices. 

AI Heft v. Hype

Are AI naysayers just being pessimistic? When Reuters asked tech leaders and investing analysts to look into their Magic 8 Balls in October, the resounding response was “Ask Again Later.”

What’s clearer is that the data centers powering AI globally will need $6.7 trillion in cumulative capital investment by 2030 to keep pace with computing demand, with another $1.3 trillion for power generation and transmission, according to McKinsey & Co. With an estimated half a million skilled workers needed this year to fill the construction industry’s intensifying labor shortages, AI visionaries could see their capital plans stalled while the labor resources required to build data centers catch up:

• “I don’t think there is a bubble, but we do see some constraints in terms of construction capacity not keeping up with all the new investments,” ABB CEO Morten Wierod told Reuters. “There is not enough people and resources to build all this.”
• Cue Wall Street’s overvaluation fears, already heightened when Nvidia became the first company to hit a $4 trillion market cap in July, then $5 trillion in October.

Yet even with billions backing AI capital spending, the Magnificent 7 stocks, which focus heavily on AI and account for more than a third of the S&P 500’s market value, will most likely be insulated if the market goes bust on profitability, said Andy Wu, an associate professor of business administration at Harvard Business School. 

1 +1 = 1 trillion: “While generative AI can do amazing things, it is also perhaps the most wasteful use of a computer ever devised,” Wu told the Harvard Gazette. “If you do 1+1 on a calculator, that’s one calculation. If you do 1+1 in generative AI, that is potentially a trillion calculations to get an answer. That consumes a huge amount of chip capacity and electricity.”