The Best AI Stocks Wall Street Is Sleeping On in 2026

The article argues that Figma, SentinelOne, and Meta Platforms may be undervalued AI-linked stocks despite competitive concerns and recent share weakness. Figma reported $304 million in Q4 revenue, up 40%, with $38 million in positive free cash flow and 136% net dollar retention, while SentinelOne grew ARR 22% and revenue 20% to $271 million in Q4. Meta’s Q1 revenue rose 33% to $56.3 billion, free cash flow was $12.4 billion, liquidity stood at $81 billion, and the stock trades at 22x earnings even after lifting 2026 capex guidance to $125 billion-$145 billion.

Analysis

The market is still pricing AI as a single trade, but this piece highlights three very different business models: monetization of workflow, monetization of security, and monetization of attention. The key second-order effect is that lower-cost AI tools do not automatically commoditize the premium products; they often expand usage while shifting demand toward the highest-trust, highest-precision layers of the stack. That favors the platforms with embedded collaboration and switching costs, while pressuring vendors whose differentiation is mostly feature-level and easier to replicate.

FIG looks more interesting than the headline drawdown suggests because the economics imply product-market fit is still improving even as expectations reset. Positive free cash flow with strong retention means the market is effectively valuing the company as if growth must decelerate sharply within 12 months; if monetization per seat remains resilient, multiple expansion can happen before absolute earnings inflect. The main risk is not competition today, but a gradual buyer-power squeeze from enterprise procurement over the next 2-4 quarters if AI-assisted design compresses seat counts faster than Figma can upsell advanced workflows.

S is a different setup: the stock can work if investors eventually accept that AI is not just an attack vector but also a distribution advantage for defense vendors with telemetry breadth. The contrarian point is that agentic security tools may increase testing volume, but they also raise the value of platforms that can observe, correlate, and automate response across endpoints; that can support premium retention even in a crowded market. META is the cleanest balance sheet-backed winner because capex fear has created a valuation discount just as operating leverage from AI-driven ad tools is likely to show up over the next 2-3 quarters; the market may be overpaying for capex uncertainty and underpricing the optionality of its data moat.

Key Decisions for Investors

• Long FIG on a 3-6 month horizon into post-earnings volatility; target a re-rating toward software peers if retention stays >130% and FCF remains positive. Risk: if net retention rolls over below 120%, the de-rate can persist despite revenue growth.
• Pair trade: long META / short PLTR for a 2-4 quarter horizon. META offers self-funded AI investment with earnings support; PLTR remains more vulnerable to multiple compression if AI enthusiasm broadens beyond narrative-driven names.
• Speculative long S with a 6-9 month horizon, sized smaller than META/FIG. Thesis is that security spending proves stickier than feared if customers keep scaling ARR; risk/reward improves if the market stops treating AI attack tools as a reason to sell the whole category.
• Buy META on dips below the post-capex-announcement range, looking for a 12-month total return driven by earnings growth plus multiple normalization. Use a tight risk budget: if ad growth or FCF weakens for two consecutive quarters, thesis breaks.
• Avoid owning ADBE as the relative loser unless you have a clear catalyst for product re-acceleration; the market is likely to keep rewarding native AI-first workflows over legacy suites over the next several quarters.