Wall Street pushed AI to the limit; now the floor is giving out. After months of heavy spending on artificial intelligence infrastructure, the foundation is starting to show strain. Returns are lagging far behind the scale of investment, raising doubts about how quickly this buildout can pay for itself. If the AI bubble pops, the fallout won’t be contained to Silicon Valley.

Bubble math: At the heart of the mounting concerns is a fundamental mismatch between how much is being spent and how little is being earned. JPMorgan Chase estimates that about $5T will go into AI infrastructure by 2030, with AmazonAMZN

, AlphabetGOOGL

, MetaMETA

, and MicrosoftMSFT

alone planning roughly $670B in spending this year. Relative to US GDP, that level of investment exceeds nearly every major capital cycle in American history.

• OpenAI and Anthropic generate about $25B and $19B annually but still trail spending, with WSJ noting the former missed growth and could potentially face funding risks.
• Still, Equity Armor Investments CEO Luke Rahbari called OpenAI misses “a distraction,” saying forecasts can swing 25–50% and the real issue is modeling AI economics.

The Calm Before the Strain

The buildout is being funded through corporate bonds, private credit, and junk debt that ultimately tie back to retirement accounts, pension funds, and bank deposits. That exposure looked easier to stomach after a strong late-April earnings cycle briefly calmed nerves, but the relief is starting to look like a cover-up. Alphabet reported 63% YoY growth in Google Cloud, while Microsoft and Amazon posted double-digit gains in cloud. All beat Wall Street expectations on revenue and EPS, reinforcing the view that AI infrastructure spending will eventually translate into sustained cloud demand.

• AppleAAPL
  stood apart, posting $111.2B in revenue and strong profit growth while continuing to print from hardware as others burn cash on AI.
• However, the “Magnificent Seven” now make up over 30% of the S&P 500, highlighting extreme concentration as their performance increasingly drives the broader market.

Echoes of 2008: Policy experts warn policymakers are repeating pre-crisis mistakes, allowing bubble risks to build without clear response plans. The rise of circular financing, opaque debt, and interconnected bets mirrors the financial engineering seen before the Great Recession, yet proposals to address a potential AI unwind remain limited in Washington. The momentum feels unstoppable, but the risks underneath are building faster than the system can absorb.