The $725 Billion Bet: Why the AI Investment Window Is Still Open

Everyone who missed Nvidia at $200 asked the same question at $400. Then again at $600. Then at $800.

“Is it too late?”

The $725 Billion Bet: Why the AI Investment Window Is Still Open

Here’s what those investors didn’t understand — and what most people asking the same question in May 2026 still don’t: the right question isn’t about price history. It’s about what the underlying capital commitment tells you about where this cycle actually is.

Microsoft, Alphabet, Amazon, Meta, and Oracle have collectively committed $725 billion in AI infrastructure spending for 2026 alone. That’s up 77% from last year. Wall Street analysts now project the figure will cross $1 trillion in 2027. These are not venture bets. These are multi-decade capital allocations backed by signed cloud contracts.

Google Cloud’s backlog just hit $460 billion — double what it was three months ago. Microsoft has an $80 billion Azure order backlog it physically cannot fulfill because it doesn’t have enough power to run the servers. The company has signed a 20-year nuclear energy deal to restart Three Mile Island specifically for AI data center load.

When the world’s most sophisticated capital allocators sign 20-year nuclear power contracts, you are not at the end of a cycle. You are in the middle of an industrial buildout.

The Part Most Investors Get Wrong

The “AI bubble” narrative focuses on valuation multiples. It misses the demand confirmation.

Nvidia’s forward earnings are projected to grow 75% this year — to $8.34 per share. At approximately 45 times forward earnings, its PEG ratio sits below 1.0. That technically makes it undervalued relative to its growth rate.

Morningstar’s May 2026 analysis found Microsoft trading 31% below fair value — the most undervalued mega-cap AI stock by conventional metrics. The stock that everyone thinks is expensive is, by one of the most rigorous valuation frameworks in the industry, genuinely cheap.

The disconnect between narrative and data is the opportunity.

Three Categories Worth Owning Right Now

The chipmaker. Nvidia generated $193.7 billion in data center revenue in its last fiscal year — up 68% year-over-year. It has $1 trillion in forward commitments from its Blackwell and Vera Rubin processors. This isn’t projection. It’s backlog.

The infrastructure enabler. The binding constraint on AI expansion is no longer chips. It’s electricity. 40% of planned AI data center projects are delayed by power grid bottlenecks, not hardware availability. Companies solving the power problem — Constellation Energy, Vertiv Holdings, Broadcom — are capturing structural, compounding demand that most AI narratives completely ignore.

The undervalued hyperscaler. Microsoft is building at a pace that keeps it capacity-constrained through at least year-end. Azure’s cloud backlog confirms demand. And Morningstar says you’re buying it at a 31% discount to fair value. For long-duration investors, that is the definition of a margin of safety.

[embed]Middle East Tensions Explained: What It Means for Oil Prices Middle East tensions and oil prices in 2026: Brent hit $120, Hormuz shut. How this war reshapes your fuel costs…factandview.com

The Real Risk Nobody Is Talking About

It isn’t that AI fails. The risk is investor impatience.

Amazon’s free cash flow is going negative this year — by $17 to $28 billion — because of AI capex. Meta’s drops 90%. These companies are spending aggressively because their backlogs justify it. But quarterly earnings reports will show ugly FCF numbers. Analysts will downgrade. Headlines will call it a bubble.

That is the most likely source of a 20–30% correction in AI stocks in 2026. Not fundamental failure — multiple compression driven by impatience with a capital cycle that is working exactly as designed.

That correction, if it comes, is the entry point — not the exit.

One Number That Settles the Debate

$460 billion.

That’s Google Cloud’s signed contract backlog as of Q1 2026. It doubled in three months. You don’t double your signed contract backlog on hype. You double it on enterprise customers who need AI infrastructure now and are willing to commit capital years in advance to secure it.

The AI revolution is not a narrative. It is a $725 billion annual capital commitment that is accelerating, backed by $2 trillion in signed cloud contracts, with nuclear power agreements to match.

The window is not closed. It is wide open — for investors who understand what they’re actually buying.

Want the Full Breakdown?

This is the short version. The complete analysis covers:

• All five categories of AI stocks with specific tickers, valuation metrics, and risk profiles
• Three scenarios — base case, upside, and downside — with exact portfolio positioning for each
• Seven FAQ answers on Nvidia, ETF selection, capex ROI, crash probability, and where Morningstar sees value
• Internal links to related Fact and View analysis on inflation hedging, market valuations, and institutional positioning

👉 Read the full article: Is It Too Late to Invest in AI Stocks in 2026?

Fact and View covers the intersection of geopolitics, economics, and markets. No sponsored content. No advertiser influence.

For informational purposes only. Not investment advice. Consult a qualified financial advisor before making investment decisions.