For Ben Cerchio, the lesson came early—and repeatedly. At TikTok, where he led product privacy, and before that at PayPal navigating information security compliance, the pattern was unmistakable: companies wanted AI, desperately, but the data they needed to train those models was trapped behind regulatory walls and risk-averse legal departments.
"Organizations are either sacrificing privacy for AI innovation or innovation for privacy," Cerchio said when his startup, Secludy AI, disclosed a $4 million seed round on May 13, 2026. It's a familiar tension in regulated industries, one that can stretch a straightforward vendor evaluation from weeks into a months-long slog through compliance review. "We're building the infrastructure that eliminates that tradeoff."
Secludy's pitch is deceptively simple: generate synthetic data that behaves like the real thing—statistically speaking—without ever exposing the actual customer records, transaction logs, or patient files that banks and fintechs guard so jealously. The San Francisco startup keeps everything self-hosted in the customer's own cloud environment, using differential privacy techniques to produce replicas suitable for training proprietary models or running vendor proofs-of-concept.
Whether that premise is enough to carve out durable territory in a crowded and consolidating market remains an open question. But the investor lineup suggests at least some conviction: Impression Ventures, a Canadian firm with a fintech focus, led the round. Joining were LAUNCH, The Syndicate (backed by Jason Calacanis), Wedbush Ventures, Precursor Ventures, Hustle Fund, Script Capital, Mana Ventures, and Chispa VC.
The Bottleneck No One Wants to Talk About
Banks and healthcare companies sit on data goldmines. Properly leveraged, that information could power everything from fraud detection systems to personalized treatment protocols. But privacy regulations—GDPR in Europe, CCPA in California, HIPAA for health data—have turned what should be straightforward AI deployments into legal and compliance marathons.
Cerchio calls it "the biggest bottleneck to AI adoption in regulated industries." Perhaps more than the founders expected, that framing appears to resonate with investors who've watched enterprise clients wrestle with the paradox: the most valuable data is also the most restricted.
Secludy's approach sidesteps the problem by never letting sensitive information leave the perimeter. The platform is claimed by Secludy to deploy in under an hour within a customer's AWS, Azure, Google Cloud, or Databricks environment. It then applies differential privacy algorithms to generate synthetic versions of text and tabular data, which can be used for model training, fine-tuning large language models, or stress-testing new AI features without risking a regulatory headache.
The company built in leakage detection tools: canary personally identifiable information (PII) injection, membership inference attack checks, bias measurement. It's the kind of technical scaffolding that enterprises demand before they'll even consider a pilot.
Does it work as advertised? Secludy claims the synthetic datasets maintain statistical fidelity that matches or beats baseline model performance, though independent validation remains limited at this stage.
Money Trails and Market Signals
The seed round arrives roughly 20 months after Secludy's pre-seed, which was announced in September 2024 and included Script Capital, Precursor Ventures, Hustle Fund, Karman Ventures, and '43. A Form D filing with the SEC, submitted on April 24, 2026—nearly three weeks before the public announcement—shows The Syndicate structured its investment through a special purpose vehicle.
That kind of timing gap isn't unusual, but it does suggest some deliberation around when and how to surface the news publicly.
The synthetic data landscape, meanwhile, has been turbulent. Nvidia acquired Gretel in March 2025. KPMG picked up YData's intellectual property and technology in October 2025. Tonic.ai, targeting a somewhat different use case around developer testing environments, raised a $35 million Series B in March 2025. The flurry of activity signals both demand and the reality that larger players are moving into the space—sometimes by acquisition, sometimes by building their own capabilities.
For a startup still in the single-digit millions in funding, that creates both opportunity and existential risk. The window to establish a foothold may be narrower than it appears.
Who's Building This
Cerchio co-founded Secludy with Mingze He, who holds a Ph.D. in computational biology and previously built enterprise AI systems at Williams-Sonoma. It's an unusual pairing—consumer privacy veteran meets computational scientist—but one that reflects the dual challenge: understanding both the technical requirements of differential privacy and the political realities of enterprise procurement.
The company formed in Delaware and registered in California in August 2024. It remains small; LinkedIn indicates visibility of 3 employees, meaning much of the hiring is still ahead. Secludy plans to use the fresh capital for go-to-market expansion, additional hires, and extending the platform to support more enterprise AI workflows.
The initial focus is financial services: banks, payments companies, fintechs. Healthcare and life sciences are next. Both verticals are data-rich and regulation-heavy, which is precisely the point.
The Bet
What Secludy is really selling is time. The value proposition isn't just privacy—it's compressing legal and privacy reviews from months to weeks, or at least making the process less painful. In an enterprise sales cycle, that kind of friction reduction can be worth a premium.
Whether customers will pay enough of a premium to support venture-scale returns is the question every investor in the round is implicitly answering in the affirmative. The composition of the cap table—a mix of fintech-focused VCs, enterprise-savvy pre-seed funds, and a well-known syndicate—suggests the narrative around privacy-preserving AI infrastructure is resonating, at least with capital allocators who've seen the compliance gauntlet firsthand.
Still, it's early. Secludy is betting that regulated industries will increasingly treat privacy-preserving infrastructure as table stakes rather than nice-to-have. And that the market will reward specialized platforms over generic synthetic data tools.
Both assumptions could prove correct. Or the space could consolidate faster than a small startup can scale, absorbed into broader enterprise AI suites or snapped up by incumbents with distribution advantages and deeper pockets.
For now, the money is in the bank. What comes next will depend on whether Cerchio and He can translate a compelling pitch into a product that enterprises actually deploy—and pay for—at scale.
