Frame Security raises $50M to fight deepfakes

Frame Security announced a $50 million debut financing round led by Index Ventures, Team8, and Picture Capital, with participation from Wiz CEO Assaf Rappaport and investor Elad Gil, according to reporting by Globes, The Jerusalem Post, Fortune, and SiliconANGLE. The company launched publicly on May 11, 2026, unveiling an AI-driven platform it frames as addressing "human risk security," which the coverage describes as protecting organizations from AI-powered social engineering and deepfake attacks. Multiple outlets report Frame's founders are Tal Shlomo (CEO) and Sharon Shmueli (CTO), both alumni of Israel's Unit 8200, and that the startup already counts enterprise customers, reported as "tens" by SiliconANGLE and as roughly 20 to 30 by Fortune, including AlphaSense and Louis Dreyfus Co.

Per press coverage, Frame's platform uses generative AI to produce realistic attack simulations across email, chat, voice and video, combined with hyper-personalized, role-based training and on-the-spot guidance for employees. SiliconANGLE and VentureBurn report the system continuously analyzes employee behavior and organizational communication patterns to surface likely threat vectors and speed the creation of targeted exercises when new attack types emerge. Fortune and Globes published direct quotes from CEO Tal Shlomo describing AI-enabled deepfakes as a growing attack vector and the motivation for the product.

Companies and practitioners have documented a clear rise in AI-enabled social engineering, with Frame citing Gartner data that security leaders encountered deepfake audio and video attacks in 2025. Observed patterns in similar vendor offerings show two converging trends: first, using generative models to automate realistic phishing and impersonation scenarios; second, shifting training from periodic, generic modules toward continuous, contextualized interventions tied to user behavior. For practitioners, this raises data and privacy tradeoffs around instrumenting corporate communication channels for behavior signals and the need to validate synthetic-attack realism against adversarial evolution.

Editorial analysis: The round and product launch place Frame within a growing field often called human-risk or security-awareness automation, where startups apply generative models to both attack simulation and defensive training. From an industry perspective, a $50 million Series A-caliber round led by major investors suggests investor conviction about the market opportunity and the scaling requirements for model development and enterprise sales. For security teams, the practical importance is twofold: defenders need more realistic testing to match attacker tooling, and vendors will be measured on how well simulations translate into measurable reductions in risky behavior without creating excessive false positives or privacy exposure.

Editorial analysis: Observers should track three indicators: customer outcomes data (phishing click-rate reductions or incident metrics) reported by vendors or customers; the vendor's approach to data minimization and privacy controls when ingesting communication signals; and how detection and simulation fidelity evolve as adversaries adopt more advanced generative models. Also monitor product integrations with existing security stacks, and whether competitors or incumbent security vendors accelerate similar offerings. Finally, watch for third-party audits or red-team results that evaluate simulation realism and potential regulatory scrutiny of employee monitoring approaches.