Incode

Incode's founding bet — build the entire identity stack on proprietary AI rather than reselling third-party document and biometric engines — has become its defining edge in a market suddenly defined by deepfakes and synthetic identity. Owning the models end to end lets Incode iterate anti-spoofing defenses faster than orchestration vendors dependent on outside suppliers, and the market is rewarding it: in November 2025 the company was reported in talks to raise $150–300M at a valuation of up to $3B, more than double its 2021 unicorn mark, on the back of roughly $170M in ARR, over 4.1 billion identity checks processed in 2024, and a customer roster spanning eight of America's top ten banks and four of Latin America's top five. Its August 2025 acquisition of AuthenticID fused Incode's AI with an enterprise-grade fraud operation, and it contributed biometrics to a $37M Login.gov contract, planting a flag in high-assurance US government identity. The backer list — General Atlantic, SoftBank, J.P. Morgan, Capital One, Coinbase — reads as a who's-who of financial institutions that are also potential customers. In a world where every remote interaction must now defend against AI-generated faces and injection attacks, the vertically integrated player that controls its own models and has already scaled to billions of checks holds a compounding data-and-accuracy advantage.

A $3B target against roughly $170M of ARR prices Incode at a demanding multiple in a category where the leaders — Socure, Persona — are equally well-capitalized and racing at the same deepfake-defense problem, and where the underlying technology is commoditizing as every vendor bolts liveness and injection-attack detection onto its stack. Owning the full model pipeline is a differentiator but also a cost and focus burden: Incode must fund its own R&D across document, biometric, and liveness domains simultaneously while orchestration rivals cherry-pick best-of-breed components and undercut on price. The AuthenticID acquisition brings integration risk — merging two fraud stacks, data networks, and go-to-market motions rarely proceeds cleanly — and the reported raise is still preliminary talks, not closed capital, in a funding environment that has punished identity companies that missed growth targets. Government contracts like Login.gov are prestigious but procurement-slow, politically exposed, and low-margin relative to the enterprise multiple implied. Deep customer concentration among the largest banks means a handful of renewals carry outsized weight, and those same institutions increasingly build or multi-source verification to avoid single-vendor dependence. If deepfake defense becomes table stakes rather than a moat, the vertical-integration premium compresses toward the orchestration crowd's economics.