1 Under-the-Radar Stock Riding the AI Boom (Hint: It's Not Nvidia)

Amkor reported Q1 revenue of $1.68 billion, up 27% year over year and a record for the quarter, with AI data center revenue also at a record level. The company signed a 10-year agreement with TSMC for advanced packaging and testing in Arizona, supporting its AI infrastructure growth story, though shares have already more than doubled this year and trade around 50x earnings. Management guided Q2 revenue to $1.75 billion-$1.85 billion and EPS of $0.42-$0.52, but 2026 capex of $2.5 billion-$3.0 billion and 2027 Arizona startup costs add execution risk.

Analysis

The market is starting to price AMKR not as a low-growth back-end supplier but as a capacity-constrained toll collector on the AI packaging bottleneck. The second-order implication is that the real economic value shifts from leading-edge wafer scaling toward who can add qualified capacity fastest in the U.S., which should support a multi-year re-rating for OSATs with credible domestic footprints and penalize smaller peers that lack balance-sheet depth. That said, the setup also creates a classic “good story, bad stock” risk: when capex surges ahead of revenue, earnings quality deteriorates before the operating leverage appears.

TSMC’s long-duration commitment matters more as a signaling event than as near-term revenue. It effectively de-risks the Arizona buildout for other tier-1 customers by validating the site as a strategic, not speculative, node in the supply chain; that should increase the odds of follow-on commitments from hyperscaler-adjacent chip designers and memory vendors over the next 12-24 months. The flip side is that this could compress future returns for AMKR if the market is extrapolating a monopoly-like scarcity premium that disappears once additional packaging capacity comes on line across the industry.

The main contrarian concern is timing mismatch: investors are paying for 2028 optionality while 2027 is likely to be the margin trough as start-up costs and under-absorbed fixed costs hit. If AI packaging demand merely grows steadily instead of inflecting sharply, the stock can de-rate even while fundamentals improve. For the broader group, the more attractive expression may be owning the demand owner and shorting the bottleneck provider if valuation outruns utilization.

Near term, the best catalyst is not the Arizona completion date but evidence that advanced packaging revenue is becoming less customer-specific and more multi-program. Any delay in customer qualification, capex slip, or evidence that alternative OSAT capacity is being booked would hit the multiple fast. Conversely, if management guides to sustained backlog visibility into 2027, the stock can remain momentum-supported despite the rich valuation.

Key Decisions for Investors

• Maintain a tactical long AMKR, but only on pullbacks toward prior breakout levels; target a 6-12 month hold with upside tied to continued AI packaging bookings, and use a tight stop if the market starts discounting Arizona delays or margin compression.
• Pair trade: long TSM / short AMKR for 6-9 months if you believe the market is overpaying for execution risk; TSM captures the strategic control point while AMKR absorbs capex and ramp risk.
• Avoid chasing AMKR after the rerating; instead, sell cash-secured puts or use a bullish put spread to get long only if implied volatility remains elevated and the market offers a better entry.
• Monitor INTC and other secondary beneficiaries for lagged upside over the next 12-18 months; if domestic packaging bottlenecks persist, suppliers with adjacent U.S. manufacturing exposure can benefit without AMKR's valuation burden.
• If AMKR trades above 55-60x forward earnings on no incremental guidance upside, consider a partial hedge via short semiconductor infrastructure baskets, since the stock is vulnerable to multiple compression before the Arizona revenue arrives.