Buy TI and Amtech to Play the AI Boom in Semiconductors

The article is constructive on semiconductors overall, citing WSTS/IDC forecasts for global semiconductor sales growth of 89.9% in 2026 to $1.5 trillion, with memory chips up 249.5% and total IC growth at 101.5%. It highlights AI infrastructure, auto, and industrial demand as key tailwinds, but notes near-term headwinds from tariffs, geopolitics, supply-chain disruptions, and inflation. The piece favors Texas Instruments and Amtech Systems, with TXN 2026 revenue/earnings growth expected at 17.4%/40.6% and ASYS estimates raised 28% for FY2026.

Analysis

The market is pricing a bifurcated semiconductor cycle: AI/data-center demand is still pulling the entire group higher, but the second-order beneficiary set is narrowing to suppliers with either pricing power or direct exposure to capacity buildout. That favors TXN as a “quality laggard” rather than a pure AI torque trade: its analog/embedded mix should lag the headline growth rate, but the hidden lever is operating leverage from years of internal capex once utilization inflects. ASYS is the higher-beta expression of the same thesis, but it is much more cyclical and should be treated as a capital spending proxy on SiC adoption, not as a straight AI winner.

The bigger underappreciated effect is that memory inflation can be both a boom and a tax on the rest of the ecosystem. If DRAM/NAND remain this tight, non-AI OEMs will likely ration inventory, delay design wins, and push out fab tool orders, which means the “good” semiconductor news is actually a headwind for broad-based demand in autos, industrials, and consumer electronics over the next 2-4 quarters. That argues for relative-value exposure to names with secular content growth and away from companies dependent on discretionary end demand or China-heavy sales channels.

Geopolitics and tariffs add a separate margin squeeze: they don’t just raise end-product prices, they also increase working capital and inventory risk, which can hit cash conversion before it shows up in revenue. The most important reversal signal is not a growth slowdown but a capex pause in AI infrastructure or a policy shock that cools spending expectations; that would hit ASYS first, then the broader equipment complex. Conversely, if AI capex keeps compounding, the market may continue to underestimate how long this cycle can sustain elevated multiples even with mediocre macro conditions.

Key Decisions for Investors

• Long TXN on any 3-5% pullback over the next 2-6 weeks; prefer it as a lower-volatility way to own the onshoring/industrial content theme with better downside protection than the high-beta equipment names.
• Long ASYS only as a tactical position, sized small, with a 3-6 month horizon; best entry is on post-earnings weakness because the stock likely over-responds to any AI capex commentary. Use a 2:1 upside/downside framework.
• Pair trade: long TXN / short a China- or consumer-exposed semiconductor name if available in-book; the thesis is that tariff and inventory pressure will hit less strategic end markets before it dents industrial/auto design content.
• Buy near-dated calls on ASYS only if AI infrastructure capex prints confirm acceleration; otherwise avoid chasing the move after a 400%+ run, as multiple compression risk is high if bookings normalize.
• Reduce exposure to the most memory-sensitive hardware names if you hold them elsewhere in the book; memory inflation is likely to delay non-AI demand and could create a 2-3 quarter air pocket in broad semiconductor end demand.