5 Artificial Intelligence (AI) Stocks That Have More Than Doubled This Year and Can Still Go Higher

The article highlights five AI infrastructure stocks—Bloom Energy, SanDisk, Lumentum, Micron Technology, and Intel—that have already more than doubled in some cases and still have room to run. Bloom Energy reported Q1 revenue up 130% year over year, raised full-year guidance to 80% growth, and posted $70 million in net income; SanDisk revenue rose 251% with management guiding for sevenfold fourth-quarter revenue growth. Micron, Lumentum, and Intel also showed strong AI-driven demand, with revenue up 70%, 90%, and improving turnaround momentum, respectively.

Analysis

This is not an AI-demand story so much as a capital-expenditure translation story: the first-order winners are the picks-and-shovels names that sit one layer down from model training and inference, where spend is less discretionary and more utility-like. The market is still underpricing how quickly hyperscaler and enterprise AI budgets are propagating into power, storage, optics, and memory bottlenecks; those bottlenecks tend to create short, violent upcycles because supply additions lag demand by quarters, not weeks.

The most interesting second-order effect is that the beneficiaries are increasingly non-linear relative to the core AI capex cycle. Memory and optical interconnect suppliers should see margin expansion persist even if headline AI capex growth decelerates, because mix shift and scarcity pricing can keep earnings growing faster than unit demand. The energy-infrastructure angle is also underappreciated: if on-site power becomes a gating item for data-center deployment, companies that can monetize uptime and speed-to-power can capture demand that would otherwise be delayed, not lost.

The contrarian risk is that this basket is crowded on the same narrative, so revisions may be the real catalyst rather than absolute beats. The near-term threat is that any moderation in hyperscaler spending or a normalization in component lead times could compress these names quickly, especially the higher-beta storage and optics plays. Longer term, the market may be overestimating the durability of today’s pricing power for commodity-like components; when capacity catches up, the winners are the names with genuine structural moats, not just cyclical scarcity.

Intel is the weakest of the set on quality: the market is paying for a turnaround with a much lower probability of sustaining premium margins versus the pure infrastructure beneficiaries. That makes it the best relative short against the stronger names if the sector digests the recent move. The cleanest setup is to own the secular bottlenecks and fade the legacy turnaround story, with the main risk being a broad AI multiple rerating that lifts all boats before fundamentals reassert.

Key Decisions for Investors

• Long BE / short XLU or utility proxies for 3-6 months: if data-center power constraints persist, BE should outperform traditional power names by a wide margin; downside is high if customer timing slips or financing conditions tighten.
• Long SNDK and MU on pullbacks for 1-2 quarters: memory scarcity plus AI mix shift can keep earnings revisions positive longer than consensus expects; take profits if supply commentary turns constructive or pricing inflects down.
• Long LITE versus short INTC in a pair trade over 3-6 months: optics has cleaner direct AI exposure and better margin expansion, while INTC remains a lower-quality turnaround with execution risk.
• Use call spreads on NVDA only as a hedge, not the main expression: if the infrastructure trade is right, component suppliers may still outperform on a relative basis even if NVDA remains strong; keep position size modest because the stock is less mispriced.
• Avoid chasing the strongest movers after earnings; instead wait for 10-15% drawdowns to initiate. These names are trading on expectations, so entry discipline matters more than headline growth rates.