3 Growth Stocks That Could Skyrocket in the Second Half of 2026

Arista Networks reported its first $3B revenue quarter (Q2 2026) while maintaining a 45% GAAP operating margin and delivering GAAP diluted EPS of $0.095, with analyst targets implying up to 57% upside. Micron posted $41.5B revenue (+346% YoY) and guided to $51B revenue and $32 diluted EPS (high end), with targets suggesting the stock could more than double. Eli Lilly grew revenue 48% to $23B driven by Mounjaro and Zepbound and raised full-year 2026 guidance to $85B–$87B, supported by an indicated ~28% upside—overall a strong momentum/read-through for large-cap growth tied to data center/AI memory and GLP-1 demand.

Analysis

The cleaner expression here is not “buy the winners,” but own the businesses where incremental demand turns into free cash flow fastest. ANET is the highest-quality pick because the AI capex cycle is still in the phase where every new rack architecture needs networking spend; that tends to be a lagged but durable beneficiary versus compute names that get repriced first. The second-order loser is the legacy switching stack and any vendor whose share depends on broad enterprise refreshes rather than hyperscale density.

MU is more interesting as a torque trade than a core compounder. The market is likely extrapolating current HBM scarcity too far forward; if supply additions from SK Hynix, Samsung, and MU itself land on schedule, the next 2-4 quarters can still look strong while the 6-18 month setup becomes much more cyclical. That makes MU vulnerable to multiple compression even if earnings keep rising, especially once investors stop paying for scarcity.

LLY is the most structurally resilient name, but it is also the one where expectations can absorb bad news the worst because the market already treats GLP-1 demand as a long-duration annuity. The real watch item is not demand, but manufacturing, payer access, and whether compounding can stay ahead of competitive pricing pressure from NVO and follow-on obesity entrants. If supply expands cleanly, the upside is slower but more durable than MU’s; if not, revenue momentum can persist longer than consensus models assume.

Key Decisions for Investors

• Long ANET vs short CSCO or JNPR for 1-3 months: own the AI networking beneficiary with better operating leverage; exit if hyperscaler capex commentary softens or ANET guides margin compression.
• Tactical long MU via 3-6 month call spread on pullbacks, not strength: play near-term earnings power while capping downside; take profits into any further guidance beat because the forward multiple is the vulnerable part.
• Overweight LLY for 6-18 months as a quality-duration compounder, but keep sizing moderate: the thesis breaks if prescription growth slows or if management signals supply normalization is lagging demand.
• Set a watchlist alert on HBM pricing and DRAM spot prices: a sustained 10%-plus decline would be the first clean signal that MU’s scarcity premium is fading.