I've Been Buying Artificial Intelligence (AI) Stocks for 10 Years. Here's the 1 Lesson This Correction Taught Me.

The article argues that recent selloffs in AI-related stocks are overdone, citing Microsoft down 12% YTD, Alphabet down nearly 10% in March, and Palantir down 20% in 2026 despite Palantir revenue rising 70% to $1.4B. It highlights Microsoft’s $37.5B quarterly capex, Alphabet’s planned $185B cloud spend, and Palantir’s $4.26B in quarterly contract value as reasons the businesses remain fundamentally strong. The piece is opinion-driven rather than new company-specific news, so the likely market impact is limited.

Analysis

The important signal here is not that AI capex is large, but that the market is beginning to discriminate between capex-funded growth and self-funding growth. That creates a second-order advantage for the platforms with durable cash generation and broad distribution, because they can outspend smaller software names without needing immediate multiple expansion. In practice, this means the current pullback is more likely to compress the weakest “AI narrative” names first, while the true infrastructure winners can regain leadership once investors stop treating capex as a one-factor negative.

Microsoft and Alphabet are likely being mispriced on the wrong horizon. Their core businesses can absorb several quarters of elevated spending, and the key question is whether incremental AI monetization shows up inside 6-12 months; if it does, today’s drawdown becomes a buying window, not a warning. The bigger risk is not balance-sheet strain, but a duration shock: if AI payback gets pushed out another 2-3 quarters, growth investors may rotate into lower-duration cash-return stories and keep multiples suppressed.

Palantir is the most interesting divergence because it is being treated like a hypergrowth hyperscaler despite having a different economic model. If enterprise and government demand stays strong, the stock can keep working even as broader AI sentiment cools; if not, the valuation has very little cushion and the downside is magnified. The market is missing that this is as much a positioning and sentiment trade as a fundamentals trade, which means price can overshoot in both directions over a 1-3 month window.

The contrarian setup is that the selloff may be creating a relative-value opportunity inside AI rather than an outright bear signal. The best expression is to own businesses where AI is a monetization layer on top of entrenched cash flows, and fade names where AI requires continuous external financing to sustain the story. That should outperform if the market remains skeptical about capex efficiency, which is likely until the next earnings season proves adoption is translating into operating leverage.

Key Decisions for Investors

• Add MSFT on weakness over the next 2-6 weeks; prefer staged entries because the risk/reward improves if the market continues to punish capex headlines, while downside is buffered by the core software franchise.
• Buy GOOGL vs. short a basket of high-multiple AI/software names over 1-3 months; the pair benefits if ad cash flows continue to subsidize AI spend while the market penalizes less self-funded peers.
• Hold PLTR only as a tactical momentum long, not a core compounder; use tight risk controls because a 15-20% retracement is plausible if positioning unwinds, but upside can extend sharply if enterprise bookings keep surprising.
• Express the view with a basket long of MSFT/GOOGL and short of weaker capex-dependent AI beneficiaries; this is the cleanest way to isolate the market’s growing distinction between cash-generative platforms and story stocks.
• Use post-earnings volatility to buy 3-6 month calls on NVDA only if supply chain commentary confirms demand is still outrunning capacity; otherwise, the stock is vulnerable to multiple compression even if fundamentals remain strong.