These Artificial Intelligence (AI) Stocks Have Crushed Nvidia in 2026 With Gains of 67% and 121%. They Can Still Soar Higher

Lumentum reported 72% year-over-year revenue growth in the first nine months of fiscal 2026 to just over $2 billion, with EPS up 4.5x to $5.27, and guided current-quarter revenue to $985 million versus $480.7 million a year ago. Applied Materials posted Q2 revenue of $7.91 billion and EPS of $2.86, and guided to $8.95 billion in revenue and $3.36 in non-GAAP EPS, implying 23% revenue growth and 36% EPS growth in the current quarter. The article argues both stocks remain positioned for further upside from AI-driven spending.

Analysis

The market is repricing the AI stack from compute-only to compute-plus-connectivity-plus-capacity expansion. That matters because the next leg of AI capex is less about raw accelerator demand and more about removing throughput constraints across data movement, packaging, and process equipment; names tied to those bottlenecks can outperform even if headline GPU spending moderates. The strongest second-order read-through is that every dollar of incremental AI infrastructure spend is increasingly accruing to the picks-and-shovels layer, not just the chip designers.

Lumentum’s move is the cleaner operating leverage story: when network bandwidth becomes the gating item, optical content per server rack rises faster than unit shipment growth. That creates a multi-quarter earnings surprise profile, but also a valuation trap if order growth normalizes even modestly; the stock is now priced for near-flawless execution. Applied Materials is the broader-cycle beneficiary, but its upside depends on a sustained WFE upcycle and on AI-related foundry/DRAM capex staying concentrated rather than dispersing into less favorable nodes.

The consensus seems to underappreciate how uneven the supply chain winners are. If hyperscalers keep shifting spend toward higher-density clusters, component vendors with the best margin mix can outgrow the original equipment names for several quarters, while pure-play semi equipment exposure should lag until foundry customers broaden capex beyond a handful of leading-edge programs. Conversely, if AI capex pauses, these are the first names where multiple compression can outrun fundamental deterioration because expectations are already elevated.

The key risk is not demand collapse but sequencing: a digestion period of 1-2 quarters where bookings remain strong but convert slower than bulls expect. That would likely hit LITE first, then AMAT, and could happen even with intact 12-month fundamentals. The best setup is to own the winners on pullbacks, not strength, and use the group’s momentum to express relative views rather than outright beta.

Key Decisions for Investors

• Long LITE on a 2-4 week pullback or post-earnings volatility reset; target a 12-month hold with upside tied to continued data-center mix expansion, but size modestly given elevated multiple and binary sentiment risk.
• Pair trade: long LITE / short NVDA for 3-6 months as a connectivity-vs-compute expression; if AI capex remains strong but shifts toward bottleneck removal, the spread can work even if NVDA consolidates.
• Long AMAT as a 6-12 month cyclical AI capex beneficiary, but favor entry only on market-wide semiconductor weakness; the setup offers better upside if WFE bookings stay above guidance and margin leverage follows.
• Use AMAT vs TSM/large-cap foundry exposure as a relative-value trade: long AMAT / short TSM basket if you expect equipment suppliers to capture more of the next capex wave than end-demand fabs.
• If chasing momentum, buy call spreads rather than stock in LITE and AMAT to cap valuation risk; the risk/reward is better when expressing upside over the next 1-2 quarters with defined downside.