VeryAI raises $10M to stop deepfakes and AI fraud with smartphone palm-scan identity verification

The internet is starting to lose a basic signal of trust: proof that the person on the other side of the screen is actually human. Generative AI can now create realistic faces, voices, and entire identities within minutes. Fraudsters are already using those tools to bypass identity checks, impersonate users, and open fake accounts. Traditional defenses such as CAPTCHAs, facial recognition, and two-factor codes are struggling to keep pace.

Now, a new startup called VeryAI is betting that the key to proving someone is human might be sitting right in front of a smartphone camera.

VeryAI announced today that it has raised $10 million in seed funding, led by Polychain Capital, to build what it calls a “Proof of Reality” platform. The round includes participation from Berggruen Institute and Anagram. The financing marks the company’s first external funding and arrives alongside the debut of its first product: a palm-scan identity system that works through a standard smartphone camera.

New startup VeryAI raises $10M to tackle AI-generated identities with palm biometrics

The idea behind the company is simple. The internet is filling up with synthetic identities. AI tools can now produce realistic faces, voices, and videos in minutes. That shift is starting to weaken common authentication methods such as facial recognition, CAPTCHAs, and one-time codes.

Security researchers say attackers are moving faster than many systems can respond. Industry data shows the time required to compromise a system has risen by 22 percent since 2023, with breaches now occurring in an average of 48 minutes.

VeryAI thinks palm biometrics offer a cleaner solution. A palm contains a dense pattern of lines and features that are highly distinctive. It is rarely photographed or shared publicly, which lowers the chance that someone could scrape it from the internet, as they can with faces.

Using the company’s system, a user simply holds up their hand to a smartphone camera. The platform analyzes the palm and converts it into a mathematical representation used for identity checks. No specialized hardware is required.

According to the company, the system produces a false acceptance rate of about one in ten million when verifying a single hand. Many facial recognition systems operate at a rate closer to 1 in 1 million. When both hands are scanned, the company says the false acceptance rate drops to roughly 1 in 100 trillion.

“Privacy is a human right. But deepfakes and synthetic content present weaknesses that current systems simply can’t keep up with. VeryAI is restoring trust in identity verification by replacing outdated methods with solutions that are accurate, private and frictionless,” said Zach Meltzer, founder and CEO of VeryAI. “Having helped build identity solutions for millions of crypto users, from KYC and reputation scores to ZK Protocols and credential systems, I’ve seen both their value and their limits in the face of AI-driven fraud. VeryAI is building the future of identity verification.”

Meltzer previously worked on identity infrastructure at Galxe, where the platform grew to more than 6,000 partners and 34 million users.

VeryAI plans to sell its verification system to businesses through a B2B model. Crypto exchanges, fintech platforms, and other online services can plug the palm verification system into their authentication workflows. Pricing is based on monthly verification volume.

The platform is built on Solana, which the company uses to record identity attestations on-chain. Solana co-founder Anatoly Yakovenko is listed as an angel investor in the project.

The system relies on Zero Knowledge Proof technology and the Solana Attestation Service, allowing users to confirm their identity across decentralized apps without revealing personal data. VeryAI is working with Light Protocol to compress blockchain data so identity records can be verified efficiently while keeping storage costs low.

After a verification check is completed, the platform generates a non-traceable identifier confirming that an action was performed by a verified human. The company says it does not store palm images. It retains encrypted feature representations that cannot be reconstructed into the original scan.

The startup’s scientific work is led by Chief Science Officer Hua Yang, a researcher in palm biometrics with more than 50 academic publications and patents.

Investors see the company as part of a broader push to restore trust online as generative AI spreads across social platforms, financial systems, and digital services.

“Every major platform, whether in finance, crypto, or social media, is grappling with the risks of AI-driven fraud,” said Olaf Carlson-Wee of Polychain Capital. “VeryAI’s palm verification technology closes that gap with accuracy, privacy and accessibility that no other biometric identity solution has yet to match. This is the foundation for a new standard of trust online.”

The company has added Matthew Groh, assistant professor at Northwestern University’s Kellogg School of Management and principal investigator of the Human-AI Collaboration Lab, as an advisor. VeryAI and Northwestern researchers are launching a collaboration aimed at studying how people detect synthetic media and resist deepfake manipulation.

The fresh capital will fund the next stage of the company’s Proof of Reality platform. VeryAI plans to build more tools that separate AI-generated identities from verified human users as synthetic media continues to spread across the internet.