Oligo Security Deploys NVIDIA AI to Rewire Vulnerability Intelligence at Runtime

Tel Aviv builds companies that think in threat models. In 2022, Oligo Security entered the arena with 3 former elite IDF cyber officers who had watched a single open source flaw ripple across production systems and stain global brands. Nadav Czerninski, Co-founder and CEO. Gal Elbaz, Co founder and CTO. Avshalom Hilu, Co-founder and CPO. They did not set out to scan code. They set out to interrogate what actually runs.

By February 2023, Oligo Security emerged from stealth with $28M across Seed and Series A, backed by Lightspeed Venture Partners, Ballistic Ventures, TLV Partners, and operators including Shlomo Kramer, Eyal Waldman, Adi Sharabani, and Eyal Manor. In January 2025, a $50M Series B led by Greenfield Partners, with Red Dot Capital Partners and Strait Capital, pushed total funding to $80M in under 2 years. Fortune 500 security teams adopted the platform. Industry recognition followed. In startup news, capital is common. Sustained traction is not.

On February 17, 2026, Oligo Security detailed a deeper shift. The company rebuilt its vulnerability intelligence engine using NVIDIA AI, deploying Nemotron Nano 9B v2 models through Amazon Bedrock and integrating the NVIDIA AI Blueprint for Vulnerability Analysis. This is infrastructure work, not branding. A Named Entity Recognition pipeline designed to parse CVEs at function level, separating abstract exposure from executable risk inside live environments.

The performance claims are sharp. A reported 55x cost reduction versus the prior Claude 4.5 Sonnet configuration. Response times 2.5x faster. CVE enrichment throughput 4x faster. Roughly 20% broader CVE coverage through enhanced preprocessing, including GitHub Security Advisories. Oligo Security states it can process about 50,000 CVEs in under 1 hour and identify 1100% more vulnerable functions than commercial advisories surface. In a market drowning in alerts, precision becomes margin.

Application Detection and Response only works if telemetry meets context. A library sitting idle is different from a function executing in production. Oligo Security’s wager is that AI tuned for security semantics, not general conversation, becomes the filter that lets security teams prioritize what is reachable, not theoretical. That is a product thesis with operational consequences.

Nadav Czerninski has argued that legacy application security cannot keep pace with modern development velocity. The NVIDIA-backed pipeline operationalizes that stance. Faster signal, lower compute cost, deeper function-level visibility. In startup news, partnerships are announced daily. Infrastructure rewrites that change unit economics and detection depth are rarer.

For investors tracking startup news in cybersecurity, this is less about model selection and more about control of the vulnerability stack. Whoever reduces noise while expanding coverage owns the decision surface. Oligo Security is positioning itself where AI acceleration meets runtime truth, and that intersection is where modern security budgets are quietly consolidating.