2 Stocks to Buy Hand Over Fist Before the Nasdaq Soars Higher in 2026

CoreWeave reported blistering growth—revenue tripled to $3.56 billion in the first nine months of 2025, with an estimated $5.1 billion for full-year 2025 and a Q3 backlog of $55.6 billion (up from $15 billion year‑over‑year)—and trades at roughly 12x sales. Palantir's AIP adoption drove Q3 2025 earnings to more than double year‑over‑year, remaining deal value jumped 91% to $8.6 billion, and customer count rose 45% YoY; consensus expects ~40% earnings growth in 2026 while the stock trades near 172x forward EPS. The piece argues both firms' sizable backlogs and accelerating AI-driven demand justify premium multiples and position them to potentially outperform in 2026, supporting a constructive view for growth-focused investors.

Analysis

Market structure: Winners are GPU/hyper-scale infra providers (CoreWeave CRWV, NVDA and data‑center REITs) and frontier AI software (PLTR) that convert data into productivity; losers include legacy CPU/cloud capacity and low‑margin managed services as pricing power shifts toward GPU supply holders. CoreWeave’s $55.6B backlog vs. 2025 revenue implies multi‑year demand > supply, supporting sustained pricing power and higher utilization-driven margins, while NVDA’s tight GPU cadence remains the choke point. Cross‑asset: a tech re‑rating will pressure long-duration bond yields and lift risk assets; expect rising implied vols on AI names and a small upward impulse to power/energy commodities from increased data center demand.

Risk assessment: Tail risks include abrupt halting of AI spend (macro or client budget cuts), GPU supply shocks (NVIDIA production/allocations), regulatory limits on model deployment, and customer concentration (large contracts from Meta/OpenAI). Immediate moves (days) will be sentiment‑driven; short term (weeks–months) hinge on quarterly RDV/backlog converts; long term (2–4 years) depends on data‑center build cycles, energy costs, and financing availability. Hidden dependencies: power/real estate bottlenecks, circular financing and warranty/lease obligations that can amplify leverage.

Trade implications: Direct play CRWV for secular infra exposure; use options on PLTR to capture asymmetric upside while limiting capital at risk given 172x forward EPS. Pair trades favor hardware beneficiaries (NVDA, suppliers) vs. expensive AI SaaS names with weak RDV conversion. Entry: stagger buys over 4–8 weeks to avoid front‑running GPU guidance; exit/scale on 40–60% moves or on two consecutive negative conversion quarters.

Contrarian angles: Consensus underestimates execution and margin pressure as providers rapidly expand — CRWV’s 12x sales prices in flawless buildout; PLTR’s valuation embeds near‑perfect RDV monetization. Historical parallel: 2017 capex spikes later moderated by supply and energy constraints; unintended consequences include energy price spikes or GPU rationing that could cut revenue growth by >30% in 12–18 months.

Key Decisions for Investors

• Consider establishing a 2–3% long position in CoreWeave (CRWV) using a 4‑week dollar‑cost‑average entry; target a 30–60% upside over 12 months tied to backlog conversion, set an initial stop‑loss at 25% and trim half of the position if quarterly revenue misses consensus by >10%.
• Allocate 1–1.5% of portfolio to 9–12 month Palantir (PLTR) call debit spreads (buy nearer‑ATM and sell 25–40% OTM) to capture upside while capping premium; take profits if two consecutive quarters show RDV growth >50% or exit if RDV growth falls below 20% across two quarters.
• Implement a relative‑value pair: go long NVDA or NVDA suppliers via 3–6 month calls sized 1.5–2% of portfolio while shorting high‑PE generic AI/SaaS names (equal notional 1.5–2%) to capture hardware capex leverage; rebalance if NVDA shipment guidance misses by >10%.
• Apply macro risk triggers: reduce broad growth/QQQ exposure by 3–5% if 10‑yr UST rises more than 50 bps within 90 days; monitor monthly GPU spot prices and CRWV customer‑concentration disclosures and cut CRWV exposure by half if top‑3 clients exceed 50% of backlog or backlog conversion <30% over 12 months.