My Top 2 AI Stocks Flying Under the Radar for May 2026

Aehr Test Systems is up 379% YTD and still trades at 14x earnings, supported by a backlog that reached about $92 million after a $41 million hyperscaler contract. ServiceNow has fallen 42% YTD, but Q1 revenue still rose 22% year over year, subscription guidance for 2026 was raised to 22%-22.5%, and RPO grew 25% to $27.7 billion despite delayed Middle East contracts tied to the war in Iran. The article is bullish on both names as overlooked AI plays, though the tone is more selective than broad-based.

Analysis

AEHR is a classic “picks-and-shovels plus backlog” setup, but the more important signal is that hyperscaler validation can re-rate the entire wafer-level test niche. If the $41M order is truly tied to broader AI accelerator qualification, this is not just revenue visibility for one company; it can force a reassessment of testing bottlenecks across the semiconductor capex chain, benefiting adjacent equipment names and potentially compressing margins for smaller test vendors that lack differentiated throughput. The main second-order effect is that investors may start underwriting a multi-year testing super-cycle rather than a one-off order spike.

The risk on AEHR is that backlog quality, not backlog size, determines whether this is a durable inflection. A single large customer can pull forward bookings without translating into clean conversion if qualification cycles slip or if platform mix changes; that means the stock can remain very momentum-sensitive over the next 1-2 quarters. The current multiple is only cheap if FY27 estimates hold; if revenue recognition slips by even one quarter, the valuation support disappears quickly because the market is already pricing in operating leverage.

NOW looks less like a broken story and more like a timing mismatch between contract timing and fundamental demand. The market is likely underestimating how much of the re-rating comes from the combination of a lower multiple and a still-growing pipeline: once delayed enterprise deals close, the stock can gap higher without needing heroic top-line acceleration. The contrarian read is that AI is not structurally displacing NOW in the near term; instead, AI is increasing enterprise workflow complexity, which should expand budget share for automation platforms that sit above infrastructure spend.

The key failure mode for NOW is that the market may keep treating every delay as permanent demand erosion, especially with geopolitical noise muddying the quarter-to-quarter read. That creates a good setup for a “buy the air pocket” trade over 3-6 months, but not for an aggressive short-dated call unless the contract closure cadence visibly improves. The asymmetric opportunity is a valuation reset trade, not a rapid fundamental acceleration trade.

Key Decisions for Investors

• Long AEHR on pullbacks over the next 2-4 weeks; use a staggered entry because the setup is momentum-dependent. Risk/reward is attractive only if you believe FY27 conversion is real; stop if backlog fails to translate into sequential revenue growth in the next print.
• Pair trade: long NOW / short a basket of expensive enterprise software names with weaker pipeline visibility for a 3-6 month mean-reversion trade. NOW’s setup is lower valuation plus intact demand, while the short leg captures the market’s continued punishment of premium multiples.
• Buy 6-12 month AEHR call spreads instead of outright equity if you want exposure to the backlog re-rate without taking full event risk. The spread structure fits a thesis where upside is driven by multiple expansion plus estimate revisions, but timing is uncertain.
• For NOW, sell cash-secured puts on weakness 1-2 months out if the stock remains near lows and the market is still pricing delayed contracts as permanent damage. This expresses a contrarian long bias with defined entry and improves basis if the name stabilizes.
• Monitor AEHR for customer concentration and conversion updates; if the next quarter confirms bookings-to-revenue conversion, add to the position. If not, treat the move as a hype-driven squeeze and exit quickly.