3 Chip Giants Still Worth Buying Despite Massive Gains

Micron is up 297% year to date and 831% over the past 12 months, Western Digital is up 333% YTD and 1,160% over 12 months, and TSMC has returned 54% YTD and 116% over the past year. The article argues all three remain attractively valued relative to AI-driven demand, citing Micron's 10x forward P/E, Western Digital's 39x forward P/E and 0.66 PEG, and TSMC's 29x forward P/E. Overall, it is a bullish valuation-driven case for further upside in AI chip and storage stocks.

Analysis

The market is discounting a longer-duration AI capex cycle than most semis have historically enjoyed, but the cleaner read is that the profit pool is migrating from accelerators to the memory/storage bottlenecks that sit one layer deeper in the stack. That matters because these businesses have far less substitution risk and far more pricing leverage once capacity is spoken for; the market is still underappreciating how long constrained supply can support margins even after the headline AI spend rate moderates. In that setup, MU and WDC benefit not just from end-demand growth but from a structural repricing of inventory risk, with forward earnings likely to stay upwardly revised for multiple quarters.

The main second-order effect is on the rest of the ecosystem: when memory and storage pricing tightens, hyperscalers have to absorb higher system costs, which can temporarily compress ROI on AI deployments and delay marginal workloads. That is bullish for the highest-quality platform owners that can pass through costs or bundle software, but it is a hidden tax on smaller AI infrastructure players and on enterprises still piloting deployments. TSM sits in the most resilient position because it is effectively tolling the entire industry; even if chip winners rotate, the foundry keeps collecting volume.

The contrarian risk is that consensus is extrapolating sold-out supply too far into 2026-2029 and assuming zero normalization in lead times. Memory and HDD markets can reprice violently when incremental capacity arrives or when customers pull forward demand, so the key timing risk is 6-18 months out, not next week. The best short setup is not the names in the article, but the suppliers and equipment vendors that get paid to expand this boom—those are the first beneficiaries of capex, and often the first to fade when the pricing cycle crests.

Key Decisions for Investors

• Long MU vs. SOXX for 3-6 months: MU’s earnings revisions should outpace the basket if memory pricing stays tight; risk is a sudden inventory normalization, which would hit MU first.
• Long WDC on pullbacks with a 6-12 month horizon: the stock still has room if 2026-2027 revenue estimates continue to rise, but size smaller than MU because HDD is more exposed to customer concentration and lead-time compression.
• Hold/add TSM as the lowest-beta AI infrastructure winner for 12+ months: best risk/reward if the market rotates from capex beneficiaries to cash-flow compounding; downside is multiple compression, not demand failure.
• Pair long TSM / short NVDA for relative-value exposure to AI capex: if AI spending remains robust but product competition intensifies, the foundry’s toll-collector model should be more durable than the accelerator OEM multiple.
• Use call spreads instead of outright longs in MU and WDC over the next earnings cycle: upside remains meaningful, but implied expectations are elevated and a delay in pricing strength would likely compress multiples before fundamentals roll over.