3 Under-the-Radar Artificial Intelligence (AI) Stocks With Explosive Potential

Weave Communications reported 2025 revenue of $239M (+17%) with $12.9M free cash flow and is rolling out an omnichannel AI receptionist after acquiring TrueLark, expanding its addressable market by $7B to $22B. Evolv Technologies' Q3 2025 revenue surged 57% to $42.9M and the company raised full-year 2025 revenue guidance to $142M–$145M. Pagaya posted 2025 revenue of $1.3B (+26%), adjusted EBITDA of $371M (+76%), GAAP net income of $80M, and guided 2026 GAAP net income of $100M–$150M (roughly +50% at the midpoint).

Analysis

The tactical edge in these stories is not the AI models themselves but the data and integration hooks they create. Companies that stitch AI into existing high-friction workflows (scheduling, screening, underwriting) can convert recurring human costs into measurable unit economics improvements — the real moat is a combination of longitudinal training data, downstream capital relationships, and integration depth with incumbents' plumbing.

Second-order competitive dynamics favor platform partners and middleware: cloud GPUs, voice/NLP licensors, and EHR/loan-origination vendors that offer “one-click” embeds can compress margins for standalone point vendors. Conversely, hardware-dependent players face a slower cadence of revenue tied to replacement cycles and retrofit friction, which creates a multi-year runway but also lumpy earnings and capital intensity.

Key risks are execution and regulatory/accuracy shocks. A high-profile mis-classification (security false negative or healthcare scheduling error) or a privacy enforcement action would force conservative retraining and reduce conversion rates for 6–18 months; macro wage inflation or rising LLM inference costs could accelerate adoption on one hand, or make SaaS price sensitivity bite on the other. The contrarian read: the market is over-indexing on feature launches; winners will be those that lock partners via contractual SLAs, revenue shares, or capital partnerships rather than those that simply ship a better bot.

Key Decisions for Investors

• Long PGYWW (6–18 months): position sized 3–5% portfolio. Rationale: asymmetric upside from a data-network flywheel and embedded distribution; downside limited to warrant premium and credit performance cycles. Use a 25–30% trailing stop or hedge with short-dated puts if underwriting metrics weaken.
• Long EVLVW (3–12 months): 2–4% portfolio exposure or buy-call spread. Rationale: durable data advantage in low-competition physical-security niches; reward from faster enterprise rollouts vs execution risk from municipal procurement timelines. Use a stop if backlog-to-revenue conversion lags two consecutive quarters.
• Pair trade — long PGYWW / short WEAV (dollar-neutral, 6–24 months): allocate balanced notional. Rationale: overweight data-infrastructure (pricing power, capital partnerships) vs point-solution exposure to bundling by EHR incumbents. Take profits if spread narrows >40% or if WEAV signs exclusive enterprise distribution.
• Portfolio hedge: buy NVDA 3-month 10–15% OTM puts sized to cover 3–5% portfolio downside. Rationale: protects against a short-term AI-hype drawdown that would hit multiples across specialty AI names even if fundamentals remain intact.