Missed Out on the Initial AI Rally? These 3 AI Stocks Are Just Getting Started.

The article argues that Meta Platforms, Micron, and Nebius still look inexpensive despite recent AI-driven rallies, citing Meta's 33% Q1 revenue growth and sub-19x forward P/E, Micron's expected 193% revenue growth this year and 57% next year, and Nebius's projected 550% growth this year. Nebius also reported 684% year-over-year revenue growth in Q1 and has partnerships with Meta and Nvidia that could support continued momentum. Overall, it is a bullish stock-picking piece on AI beneficiaries rather than a company-specific catalyst.

Analysis

The setup is less about “AI winners” broadly and more about which layer of the stack still has pricing power. META is the cleanest quality/growth arb: if ad demand remains resilient, it can keep compounding while still trading like a mature mega-cap, but the bigger second-order driver is its capex intensity pulling through demand to semis, networking, and power infrastructure. That makes META a bellwether for the entire AI spend cycle rather than just a standalone multiple re-rate.

MU is the higher-beta beneficiary of a supply-constrained memory market, but the key risk is that the market is extrapolating shortage economics too far into 2026. Memory is cyclical even in AI-led upturns; if hyperscaler procurement normalizes or new capacity comes on faster than expected, margin expansion can compress abruptly. Still, the demand shock from AI infrastructure is real enough that the first-order earnings inflection likely persists for several quarters, making pullbacks more interesting than chasing strength.

NBIS is the most reflexive name here: the growth rate is extraordinary, but the stock likely trades on perceived scarcity of AI compute rather than durable operating leverage. The hidden variable is customer concentration and partner dependence; if a single large relationship slows, the market will probably de-rate the entire narrative faster than fundamentals change. NVDA and INTC are indirect signposts: NVDA remains the toll collector on the buildout, while INTC is the relative loser if scarce leading-edge supply continues flowing to preferred ecosystem partners instead of open-market buyers.

The consensus miss is that this is not one trade but three different durations: META is a quality multiple expansion story over 12-24 months, MU is a cyclical earnings momentum trade over 3-9 months, and NBIS is a momentum/speculative growth name with binary execution risk. The current optimism looks justified for the group, but the dispersion is likely to widen, not narrow, as investors start separating durable cash generation from borrowed growth.

Key Decisions for Investors

• Long META vs. QQQ on a 6-12 month horizon: buy on any 5-8% pullback, target a move to market-multiple parity; downside is capped if ad growth decelerates but capex keeps sentiment supported.
• Trade MU as a tactical momentum long, not a core hold: prefer call spreads 3-6 months out to capture continued shortage pricing while limiting downside if memory supply normalizes faster than expected.
• Pair long NBIS / short INTC for a 3-9 month relative-value trade: thesis is that scarce AI compute allocation accrues to preferred neocloud partners, while legacy foundry narratives stay structurally challenged.
• Own NVDA as the highest-quality hedge against an AI capex continuation regime, but use it only as an overlay; if META/MU/NBIS weaken together, that signals a capex air pocket and warrants trimming.
• If NBIS gaps higher on partnership headlines, monetize into strength: the risk/reward becomes poor quickly because valuation will outrun realized revenue unless quarterly execution keeps surprising.