AI spending is surging, but a hidden risk is getting overlooked

The real question is Wall Street Nowadays, it is wondered whether the bull run will continue in 2026.

For 2023, 2024 and 2025 S&P 500calendar year returns stand at 24%, 23% and 16%. Analysts unanimously predict solid years ahead.

But there is a problem.

Bloomberg’s comprehensive look at the most important predictions for 2026, artificial intelligence Companies are at the center of the market while saying the biggest danger is not being part of the AI ​​scene. However, the focus seems to be on software, while the real problem is infrastructure.

This helps explain why the data center ecosystem has become a deal-making machine.

Bloomberg AI will add about $70 billion to data center merger and acquisition talks in 2025, he said. This is not a random number; This is a sign that buyers think size matters right now.

Example of a headline: SoftBank agrees to acquisition Digital Bridge $16 per share in cash. The deal is expected to be completed in the second half of 2026. The company also stated that the bid was 15% higher than the previous close and 49% higher than the 30-day VWAP.

DigitalBridge CEO Marc Ganzi called this a moment of AI-driven demand and, more importantly, chance to play offensively with the help of balance.

If you want “market proof” that this isn’t just a hype rally, read the business transcripts now.

The alphabet emerged in itself 2025 Q2 earnings release It will spend about $85 billion on capital expenditures in 2025, more than the $75 billion it had planned.

Most of this money will go to technical infrastructure. It was also stated that capital expenditures will increase again in 2026 due to customer demand and the opportunity for expansion.

More AI Stocks:

Microsoft in the first quarter of fiscal year 2026, spent $34.9 billion on capital expenditures. Half of this was for “short-lived assets” like GPUs and CPUs, the other half was for long-lived assets and large data center sites (including financial leases).

In simple terms, this means that Big Tech isn’t just “interested” in AI; It means spending money to create a second economy around land, electricity, cooling, switches, racks and fibre.

The International Energy Agency is very concerned. The organization predicts that data centers around the world will use approximately 415 TWh of electricity in 2024, which will correspond to 1.5% of all electricity use.

The base case scenario says this number could rise to almost 945 TWh by 2030, which could be four times more. This is not an easy solution and poses a problem for infrastructure development.

Related: Nvidia’s $20 billion Groq game is its 2026 plan

The IEA estimates that by 2030, the United States may need about 240 TWh more electricity for data centers than now (and China may need about 175 TWh more).

That’s why the AI ​​market in 2026 no longer looks like “pick the best model.” It looks more like this:

• Who can gain power the fastest?

• Who can approve it and build it on time?

• Who can afford long-term assets without affecting profits?

Alphabet acknowledged that increased capital spending hurt its bottom line, largely due to higher depreciation.

It also revealed that depreciation increased to $5 billion in the second quarter, which is $1.3 billion more than in the same period last year.

Related: This $40 PC correction could be the quietest MSFT signal of 2026

Microsoft got even more specific: Most of the money goes to “short-lived assets,” including CPUs and GPUs; That’s just a polite way of saying that this hardware is selling out fast.

This is important in 2026 because the market will ask tougher questions.

• Are these investments making money fast enough?

• Are companies building long-term competitive advantages or are they just wasting money on the most expensive upgrade cycle in the world?

• Power becomes a limiting factor: IEA’s base case arithmetic It shows that demand for data centers is growing fast enough that grid concerns may become critical on a larger scale, not just in the news.

• Capex continues to rise and Wall Street is starting to give it a rating. Alphabet raises 2025 capital spending forecast to 85 billion dollars He predicted it would climb further in 2026. Microsoft spent $34.9 billion On capital goods every three months. This is the kind of spending that makes investors care about more than just the “AI narrative.”

• Mergers and acquisitions are accelerating as buyers seek larger businesses. Bloomberg’s 2025 data center figure is $70 billion. When everyone wants to be exposed at the same time, things that used to be boring become important.

Related: Novo and Lilly shift GLP-1 strategy abroad: US could be next

This story was first published by . Street First appeared on January 4, 2026 Economy section. Add TheStreet at: Preferred Source by clicking here.