2 AI Stocks That Could Go Parabolic

CoreWeave reported a $67B AI-cloud revenue backlog in Q4 2025 (up from $55B the prior quarter) but generated $4.75B negative free cash flow in 2025 with $3.1B liquidity and debt rising to >$21B (from $14B); it plans $30–35B in capex this year while analysts forecast $12.4B revenue in 2026 (+142% YoY) and $23B in 2027. Nokia’s partnership with Nvidia on AI-RAN has helped the stock rise >60% over the last year; 2025 sales were €20B (+3% YoY) with net profit €660M (-49%), and management forecasts €2.0–2.5B operating profit for 2026 with a forward P/E around 22.

Analysis

CoreWeave sits at a structural inflection point: demand elasticity for AI-optimized hosting is high but the business model is capital-intensive and GPU-dependent. That combination creates a bimodal outcome over 12–24 months — either successful conversion of backlog into high-margin recurring revenue or meaningful equity dilution / margin compression if GPU pricing, allocation, or interest-cost dynamics deteriorate.

Nokia’s Nvidia tie-up is a classic ecosystem play where the winner is not just the vendor but the vendor who becomes the preferred integrator for a dominant silicon provider. If telecom operators lean into Nvidia-centric AI-RAN stacks over the next 18–36 months, Nokia can re-rate rapidly; conversely, execution delays or multi-vendor acceptance (including custom silicon from hyperscalers) would leave much of the re-rating as headline-driven upside with limited fundamental support.

Second-order winners include software integrators and testing/validation firms that reduce operator deployment risk, while legacy silicon vendors and vertically integrated hyperscalers (if they push x86/FPGA alternatives) face margin pressure. Key near-term catalysts are NVDA product cycles and supply policy (60–90 day impact window for enterprise procurement), telecom standardization milestones (6–24 months), and any financing event from CoreWeave (which would reset valuation assumptions immediately).

The consensus underprices event risk on CoreWeave’s capex path and overprices certainty of Nokia’s re-rating; both create tradable dispersion between infrastructure growth and hardware ecosystem adoption. Tactical positioning should be skewed to defined-loss option structures that capture convex upside while protecting against financing or supply shocks.

Key Decisions for Investors

• CRWV: Buy a 9–12 month call spread (long ATM call, short 1.5x strike) sized 1–2% of equity book to capture backlog conversion upside while capping downside. Hedge tail risk with a smaller NVDA 3–6 month put position (0.3–0.6% notional). Target 2–3x payoff if execution succeeds; expect potential >30–40% equity drawdown on a dilutive financing event.
• NOK: Initiate a 6–18 month asymmetric long using long-dated calls (or equity) sized 2–3% of book paired with a 6–12 month short NVDA position sized 0.5–1% to neutralize pure GPU-platform beta. Risk/reward: if AI-RAN adoption accelerates, aim for +50–100% upside; stop-loss on 30% drawdown if guidance/partner milestones slip.
• INTC: Buy a 6–12 month put spread to express downside risk from platform displacement and slower data-center cycles. Keep position size small (0.5–1%) as a macro/structural hedge against semiconductor share shifts; expected payoff 2:1 vs premium if Intel loses share to Nvidia-aligned ecosystems.
• Portfolio sizing & risk control: total directional exposure to these themes should not exceed 6–8% of AUM. Use defined-loss option structures and monitor three triggers to trim positions: (1) CRWV equity financing announcement, (2) NVDA supply policy shift, (3) missed Nokia integration milestones — each should prompt re-evaluation within 48–72 hours.