Michael Burry adds to beaten-down stocks while warning of echoes of dot-com bubble

Michael Burry said he added to beaten-down names including MercadoLibre, Adobe, PayPal, Zoetis and built a full-sized position in Lululemon, while warning that the AI rally resembles the late-1990s dot-com bubble. He cited Apollo data showing 87% of venture capital funding is going to AI-related companies, with AI-linked borrowers accounting for nearly half of investment-grade bond issuance and about 38% of high-yield issuance. The piece is primarily a sentiment and positioning update, but Burry’s warning could reinforce risk-off positioning in AI-linked and momentum trades.

Analysis

This is less a call on the individual names than a regime signal: when a prominent contrarian starts leaning into internationally exposed, cash-generative franchises while explicitly attacking the crowded growth complex, it usually means dispersion is about to widen further. The second-order effect is that capital may rotate within equities rather than leave risk altogether — away from expensive AI-adjacent duration and toward “boring” compounding businesses that have been punished for geography, not fundamentals. That favors quality-at-a-discount factors and hurts anything relying on multiple expansion and narrative momentum.

The more interesting transmission channel is credit. If AI borrowers continue to dominate IG/HY supply, marginal balance-sheet capacity gets crowded out for everyone else, increasing funding costs and reducing the market’s willingness to underwrite weaker business models. That creates a setup where equity leaders can keep levitating even as credit conditions for the broader market quietly deteriorate — a classic late-cycle split that can persist for months before it breaks.

Among the mentioned names, the highest convexity is probably in the payments and software complex, where sentiment has already been damaged and incremental bad news can still drive de-rating. Conversely, the Latin America e-commerce and specialty consumer franchises may have the best asymmetry if the market begins rewarding local execution over U.S.-centric AI exposure. The key contrarian miss is that this may not be a wholesale value rotation; it may simply be the first stage of a factor unwind where only the cheapest, most hated, and most self-funded names outperform while the rest of the market still trades like momentum.

Near term, the catalyst is not earnings growth so much as factor positioning: if AI leadership stalls for even a few sessions, crowded longs can unwind quickly, but if the AI tape keeps levitating, these shorts can be painful on timing. Over a 3-6 month horizon, the risk is that the market keeps ignoring the warning until debt markets or guidance resets force a broader repricing. The cleanest setup is to own the names with durable cash generation and avoid paying up for the remaining duration-sensitive compounders.

Key Decisions for Investors

• Long MELI vs short a basket of AI-high-beta winners on a 3-6 month horizon; seek a 1:2 risk/reward if the market starts rewarding international cash compounding over narrative growth.
• Reduce or hedge exposure to ADBE/PYPL via call overwrites or outright shorts for 1-3 months; these names have limited near-term upside if factor rotation continues, with downside amplified by multiple compression.
• Initiate a relative-value long ZTS / short high-multiple software pair for 2-4 quarters; ZTS has lower execution risk and less dependence on sentiment recovery.
• Buy downside protection on a broad tech proxy with a 2-4 month tenor; if AI leadership rolls over, the unwind in crowded positioning can be fast and disorderly.
• For opportunistic investors, add MELI on weakness with a staggered entry over 2-6 weeks; use a tight stop if international-exposure discount fails to narrow after the next earnings cycle.