Billionaire Investor Philippe Laffont Has Nearly 23% of Coatue Management's Portfolio Invested in 3 Artificial Intelligence (AI) Stocks

The article highlights Coatue Management’s ~23% AI concentration in Taiwan Semiconductor (8.8% of portfolio), Micron (7.5%), and a valued SpaceX stake (~$3.2B). TSMC is cited as handling 70%+ of advanced AI chip manufacturing (73% of third-party foundry market in 1Q26) and is said to have committed an additional $100B to its Arizona campus, bringing total planned investment to $265B, while Micron’s AI-driven memory demand is expected to stay supply-constrained through 2027/2028 and it signed 16 multiyear customer agreements targeting at least $100B revenue through 2030. Overall, the news is a bullish read-through on AI semiconductor and compute supply-chain demand rather than a specific earnings/guidance catalyst.

Analysis

This is less a “buy the names” note than a confirmation that AI upside still accrues to the bottleneck owners. The market keeps rewarding assets with long lead times and scarce substitutability: advanced foundry capacity and high-bandwidth memory. That favors TSM and MU over designers, but the second-order effect is valuation risk — these are now crowded consensus longs, so the next leg likely requires another revision cycle, not just continued enthusiasm.

For MU, the key setup is duration of scarcity. If DRAM/HBM tightness persists into 2027-2028, earnings power can stay elevated long enough to re-rate the stock beyond a normal cyclical peak multiple. The risk is that AI capex eventually shifts from “can’t get enough supply” to “inventory digestion,” and memory usually rolls over fastest when utilization loosens. TSM is higher quality but also faces a different tax: geopolitical premium plus massive capex can cap near-term FCF expansion even as strategic value rises.

The contrarian view is that the consensus may be underestimating how much of the AI value chain is already priced. TSM may be the safest business but not the best risk/reward after a strong run; MU likely has more upside if spot pricing and contract pricing stay firm. The public-market read-through on SpaceX is weak because the position is private and illiquid — more a sentiment signal than a tradeable catalyst. What would break the thesis: any sign of hyperscaler spending normalization, DRAM/NAND price inflection, or TSM commentary that foundry utilization / advanced packaging bottlenecks are easing faster than expected.

Key Decisions for Investors

• Prefer MU over TSM on a 3-6 month horizon: long MU / short TSM as a relative-value pair if you want AI exposure without paying the full scarcity premium; thesis breaks if TSM keeps outgrowing estimates and MU spot pricing rolls over.
• For outright exposure, add MU on pullbacks and use a 12-18 month horizon; risk/reward is better than chasing TSM after the run, with upside tied to sustained DRAM/HBM tightness and downside capped only if memory supply normalizes earlier than 2027.
• Avoid initiating new aggressive longs in TSM at current levels; if already long, consider call overwriting or trimming into strength and re-entering only on a valuation reset or after clearer evidence that capex is translating into higher FCF.
• Set a watch item on memory pricing and hyperscaler capex commentary over the next 1-2 earnings cycles; a downshift in those data points would be the cleanest falsifier for MU and would likely compress the whole AI supply-chain basket.
• No direct public-market trade on SPCX from this note; treat it as sentiment, not signal. Use it only as a reminder that private AI infrastructure remains hot, not as justification for paying up in public names.