Palantir stock plunges after 'Big Short' investor Michael Burry says Anthropic is eating its lunch

Anthropic's ARR was cited as rising from $9B to $30B in months, and Palantir (PLTR) shares dropped roughly 7% after Michael Burry's deleted post claiming Anthropic is 'eating Palantir's lunch.' Burry has disclosed a sizable short via long-dated PLTR puts (around Sep 2025) and argues PLTR is a low-margin, consulting-heavy business that took ~20 years to reach $5B. A Pentagon ban on Anthropic forced contractors to remove Claude from Palantir's Maven systems, highlighting both competitive pressure from plug-and-play AI and regulatory/operational risk to PLTR.

Analysis

The market reaction overstates the speed at which enterprise operational platforms can be displaced by plug-and-play model APIs. Enterprises with classified or regulated data face certification, audit and integration friction that typically imposes a 12–36 month migration lag; that lag preserves recurring revenue and gives incumbents time to rebundle capabilities. Palantir’s deployment-heavy revenue base implies gross margin sensitivity to utilization — a 5–10% permanent drop in Forward Deployed Engineer (FDE) billable hours would plausibly compress reported gross margin by ~200–500bps over 1–2 years as headcount and fixed-cost absorption reprice.

Second-order winners include cloud providers and model hosts who monetize API demand (incremental $/token revenue, higher data egress), and mid-tier systems integrators that can repackage model APIs into regulated workflows; these firms capture implementation dollars as enterprises avoid risky re-platforming. Conversely, pure product OEMs that lack certification footprints or billing relationships with government/health customers will see elongating sales cycles and higher churn. If a vendor ban or procurement clampdown widens, defense primes and on-prem software vendors should see an incremental share gain in multi-year RFPs.

Key catalysts: near-term is sentiment/flow (days-weeks) driven by headlines and option flows; medium-term (3–12 months) is contract renewals and any mandated technology substitutions; long-term (12–36 months) is product roadmap (incumbent LLM partnerships or native model development). Reversals occur if incumbents announce credible managed-model partnerships, materially shift pricing away from labor-led services to SaaS, or if regulatory constraints limit model API penetration into sensitive workloads.

Key Decisions for Investors

• Buy a cost‑defined PLTR bearish option structure: buy 12–24 month put spread (long-dated put / short further-OTM put) sized as 1–2% of book. Rationale: captures multi-year platform risk while capping premium; payoff ~3–6x if equity halves by expiry; max loss = premium paid.
• Pair trade: short PLTR equity vs long MSFT (0.6x notional hedge) over 6–12 months. Rationale: isolates platform/defense exposure from broad AI/cloud upside; target >3:1 upside if PLTR underperforms by 25–40% while MSFT is flat-to-up due to model/provider moat.
• Long selective defense/secure-ops exposure (examples: LHX, GD) 6–18 months as a hedge against accelerated gov procurement toward certified vendors. Rationale: downside protection if sensitive workloads revert to vetted incumbents; expected 10–25% relative upside in a constrained vendor environment.
• Tactical volatility trade: sell 2–6 week PLTR straddles after IV spikes, size small and hedge with long wings or buy cheap long-dated puts as tail protection. Rationale: monetize headline-driven IV; risk-managed if stock gaps materially on contract news.