Wall Street Sees an AI Bubble Forming and Is Gaming What Pops It

Three years after ChatGPT ignited an AI-driven rally, investor sentiment is cooling as recent selloffs—from Nvidia to Oracle’s stock tumble after disclosure of soaring AI-related capex and bond-funded data-center builds—highlight the risk that outsized spending may not translate quickly into profits. The ecosystem’s funding needs are immense: OpenAI is reported to plan roughly $1.4 trillion of spending and to burn about $115 billion through 2029, Nvidia has pledged up to $100 billion of investments, and Big Tech is projected to spend over $400 billion on data-center capex in the next year, driving rising depreciation and pressure on free cash flow, buybacks and dividends. Valuations are mixed—some AI-exposed names (Palantir, Snowflake) trade at very high multiples while core incumbents like Nvidia, Alphabet and Microsoft remain below 30x—which leaves investors weighing a potential rotation or multiple compression if revenue growth and monetization lag the massive capital deployment.

Analysis

Three years after ChatGPT sparked the AI investment surge, market signals show rising skepticism: Nvidia experienced a selloff, Oracle’s shares plunged after reporting significantly higher AI-related capital expenditures and a cloud-sales miss, and reports that OpenAI plans roughly $1.4 trillion of spending and may burn about $115 billion through 2029 have traders questioning sustainability. Nvidia has pledged up to $100 billion in investments that are recycling capital into the ecosystem, and Oracle’s credit-risk gauge reached its highest level since 2009 after issuing large bond financings for data centers, illustrating how funding and leverage are central vulnerabilities.
Big Tech remains the engine of AI spending — Alphabet, Microsoft, Amazon and Meta are projected to spend more than $400 billion on capex in the next 12 months — but rising depreciation (about $10 billion in late‑2023 to nearly $22 billion most recently and estimated near $30 billion next year) is pressuring free cash flow and could compress buybacks and dividends. Bloomberg Intelligence projects 18% earnings growth for the Magnificent Seven in 2026, the slowest in four years, leaving room for multiple weakness if revenue monetization of AI lags.
Valuations are bifurcated: names like Palantir (~180x) and Snowflake (~140x) trade at stretched multiples while Nvidia, Alphabet and Microsoft sit below 30x and the Nasdaq 100 is about 26x projected profits versus dot‑com highs above 80x. That dispersion suggests concentrated speculative risk and a plausible rotation or multiple compression scenario if AI investments fail to produce commensurate revenue and cash‑flow gains; credit-market signals and capex/depreciation trajectories are likely to be the earliest hard indicators.

Key Decisions for Investors

• Trim or avoid exposure to highly speculative, ultra‑high‑multiple AI names (for example PLTR and SNOW) until there is consistent evidence of revenue monetization and positive cash‑flow trajectories
• Maintain selective, sized positions in large-cap incumbents (NVDA, GOOGL, MSFT) given their scale and relative valuation below 30x, but monitor quarterly capex, depreciation and FCF guidance and be prepared to reduce exposure if margins or cash generation deteriorate
• Use credit-market signals (widening spreads or rising issuer credit risk like Oracle’s recent move) as an early warning to cut levered infrastructure and data-center bets that rely on continued external funding
• Implement risk-management measures — such as modest hedges, position sizing reductions, or pairs trades versus broad tech benchmarks — to protect against a rotation or multiple compression while awaiting clearer monetization proof