FBI: Americans lost a record $21 billion to cybercrime last year

U.S. victims lost nearly $21.0B to cyber-enabled crimes in 2025, up 26% y/y from $16.6B, with IC3 complaints topping 1.0M (from 859k). Crypto-related crime caused the largest loss (> $11B across 181,565 cases) and investment fraud produced $8.6B (49% of scam incidents); Americans 60+ lost $7.7B (+37% y/y). The FBI initiated 3,900 FFKC interventions, freezing $679M of $1.16B targeted, and for the first time included AI-related scams (22,300 complaints, $893M losses); most-targeted critical infrastructure sectors were healthcare, manufacturing, financial services, IT, and government.

Analysis

The secular rise in fraud and sophisticated social-engineering attacks creates asymmetric demand across three buckets: enterprise detection/response, identity/authentication, and regulated custody/incident remediation. Vendors that can convert incident volume into recurring professional services and telemetry sales will see higher gross retention even if headline product multiples face compression; expect multi-year upside to margins from managed detection and IR contracts that have 2–4x higher gross margins than pure software deals.

A wave of regulatory and insurance repricing is a second-order accelerator. Larger, regulated platforms with scale-in underwriting and compliance capabilities will gain share as counterparties and insurers prefer counterparties that demonstrably lower loss frequency; conversely, thinly capitalized consumer-facing intermediaries will see CAC and capital costs rise materially, compressing free cash flow even if top-line transaction volumes remain intact.

AI-driven deception changes the threat model and shortens detection windows, favoring solutions that combine real-time behavioral signals with provenance/cryptographic attestation. This flips value toward firms that own the edges (identity providers, CDN/bot-management, and cloud infra) and can instrument traffic for signal; boutique point products without strong data network effects risk being outcompeted.

Time horizons: immediate (0–3 months) — newsflow and government procurement notices move equities; medium (3–12 months) — contract awards, insurance cycle repricing, and compliance upgrades; long (12–36 months) — standards/crypto custody consolidation and embedding of attestation-based identity. Reversals occur if mass attribution + fund freezes materially reduce fraud profitability or if market begins to price in rapid commoditization of AI-detection tech.

Key Decisions for Investors

• Long Palo Alto Networks (PANW) — buy a 9–12 month call spread (buy ATM, sell 20–25% OTM) sized 1–2% NAV. Rationale: captures enterprise spend on integrated detection/IR and OT/critical-infrastructure modules with limited gamma; target 30–50% upside vs max loss ~premium, stop if 25% downside in 60 days.
• Long Leidos (LDOS) or Booz Allen Hamilton (BAH) — buy outright shares (1–2% NAV) with a 6–18 month horizon. Rationale: exposure to government/critical-infra cybersecurity budgets and higher-margin professional services; asymmetric payoff if large contract awards accelerate—trim into 30–40% gains.
• Pair trade: Long Coinbase (COIN) 6–12 month calls (or stock) / Short Robinhood (HOOD) (equal dollar) — 1% NAV each leg. Rationale: favors regulated custody and analytics incumbents over consumer platforms with greater fraud/KYC friction. Risk: crypto price shock; hedge delta with small inverse crypto ETF if necessary.
• Tactical option: Buy Cloudflare (NET) 6–9 month calls (0.5–1% NAV) as a hedge against bot/traffic-based fraud growth. Rationale: benefits from demand for edge instrumentation and bot management; cap downside to premium while retaining upside from multiple expansion if ARR acceleration occurs.