3 AI Stocks I'd Be Happy to Hold Through Any Pullback

Broadcom reported Q1 2026 AI revenue of $8.4 billion, up 106% year over year, with total revenue rising 29% to $19.3 billion and management citing a path to more than $100 billion in AI chip revenue by 2027. Alphabet highlighted $60.4 billion in Google Search revenue in Q1 2026, up 19%, alongside $64.4 billion in trailing 12-month free cash flow and $126.8 billion in cash. ASML raised full-year net sales guidance to 36 billion-40 billion euros after posting Q1 sales of 8.8 billion euros, reinforcing the article's bullish case for these AI infrastructure leaders.

Analysis

The market is still underappreciating how the AI stack is bifurcating into two durable monetization layers: proprietary acceleration at the chip-design level and control points in compute manufacturing. That favors AVGO and ASML simultaneously, because custom silicon adoption increases the value of Broadcom’s design win while also locking in upstream equipment intensity for leading-edge nodes. The second-order effect is that Nvidia’s dominance becomes less linear than consensus assumes; as hyperscalers diversify into custom accelerators, the spend does not disappear, but it reallocates toward whichever supplier can embed itself deepest into the customer’s roadmap.

GOOGL is the cleanest quality compounder here because its AI optionality is effectively financed by Search and distribution, which lowers the probability of a valuation reset if AI monetization is slower than expected. The hidden risk is not model quality but product cannibalization: any acceleration in AI answers that reduces search query volume could pressure ad pricing before new formats fully offset it. That said, Alphabet’s balance sheet and owned distribution give it a much longer runway than pure-play AI beneficiaries, so the downside is more about multiple compression than existential share loss.

ASML’s setup is more nuanced: the near-monopoly status creates pricing power, but it also makes quarterly lumpy guidance vulnerable to timing, not demand destruction. The key question for the next 6-12 months is whether customer capex is pulled forward enough to offset digestion at foundries and memory makers; if not, the stock can remain fundamentally strong while underperforming on order timing noise. The contrarian view is that the consensus is too willing to extrapolate AI capex in a straight line, when in reality the supply chain will likely see alternating waves of over-ordering and digestion, creating better entry points than momentum chasing.

Key Decisions for Investors

• Maintain a core long AVGO position on 6-12 month horizon; add on 5-8% pullbacks, with the thesis that custom AI content per hyperscaler dollar is still early and switching costs make revenue sticky.
• Pair long GOOGL vs short a high-multiple AI infrastructure proxy over 3-6 months; use it as a quality hedge against AI capex deceleration because Alphabet has self-funding optionality while weaker names depend on perpetual re-acceleration.
• Use ASML weakness tied to quarterly order noise to buy call spreads or outright equity on 1-3 month horizons; risk/reward improves when the market confuses timing volatility with demand inflection.
• Avoid chasing NVDA here; relative upside appears capped if more hyperscaler spend migrates into custom silicon, making NVDA better suited as a trading hedge than a long-duration overweight.
• If AVGO rallies sharply into a broad AI tape, consider trimming into strength and reallocating to GOOGL/ASML, where the moat is less dependent on a single product cycle.