Lean SuperIntelligence (LSI) Closes Initial Financing and Announces First Self-Improving, On-Premise, Lean Cybersecurity Models Beyond Mythos™

News provided by

Oct 09, 2026, 13:39 ET

Self-improving cybersecurity LLMs do offense–defense interplay in customizable on-prem deployments and run at massively lower cost.

PALO ALTO, Calif., Oct. 9, 2026 /PRNewswire/ -- Lean SuperIntelligence (LSI) is the first AI lab to build self-improving offense and defense security LLMs that enable enterprises to stay ahead of attackers without renting out their data to the cloud. The two models improve each other: LSI-Offense-1 LLM finds, exploits, and validates vulnerabilities and LSI-Defense-1 LLM detects, investigates, and responds within minutes, not days to weeks. Inside LSI's Cyber World Models, every attack the offense finds becomes a lesson for the defense, and every defense that holds forces the offense to find a new path, so both models keep getting better on customers' own data. Customizable and air-gapped, they beat frontier LLMs at vulnerability discovery and threat detection, and they are lean enough to run continuously, scanning every check-in software developers make. LSI's pre-seed round is backed by leading investors and angels who have led AI and security at OpenAI, CrowdStrike, SentinelOne, Zscaler, Commvault, and Gruve AI.

Outperforming frontier models on offense and defense

On independent, real-world offense security benchmarks, LSI-Offense-1 LLM significantly outperforms frontier models including Claude Mythos™, GPT-6 and Gemini-3.8-Cyber at autonomously discovering, exploiting and validating vulnerabilities. The model has already uncovered previously unknown vulnerabilities. The results show that specialized, lean cybersecurity models can exceed general-purpose frontier models.

On public defense benchmarks spanning multi-step threat detection, investigation, and response, LSI-Defense-1 LLM significantly outperforms frontier models from the same labs at a fraction of their cost, triaging alerts, correlating telemetry, and returning verdicts in seconds with far fewer false positives.

Why it matters

Attacks move in minutes. A rumoured vulnerability becomes a working exploit in under 10 minutes; defense must keep pace on every event.
Enterprises want to own, not rent. Today they send code and telemetry to frontier labs, pay hundreds of millions in inference, and own nothing they learn. That's why two-thirds of enterprises moved AI workloads back on-prem last year,[1] and sovereign cloud spending hits $80B in 2026.[2]

LSI gives enterprises self-improving security models they own.

Founded by the team behind National / Sovereign LLMs, the first AI Guardrails, AI labs at Alexa, Uber AI and Salesforce Research — and elite security experts.

Chandra Khatri, Founder and CEO of LSI, with more than 50 patents and publications, was Founding Head of AI at Krutrim, where he built India's first sovereign LLM in 2023. He co-founded Got It AI, where he built ELMAR, the first enterprise LLM, and TruthChecker, the world's first hallucination guardrails model (acquired). He led AI initiatives at Uber AI, Amazon Alexa AI, and eBay Research.
LSI's founding engineers have built sovereign AI and foundation models end to end, from pre-training data to production deployment to 10s of millions of users.
LSI's founding security team has discovered hundreds of critical CVEs for major organizations and national intelligence agencies.

Comments

"You cannot defend against an attack you have never seen. So we trained the Offense LLM first," said Chandra Khatri, Founder and CEO of LSI. "Our LSI-Offense-1 LLM is built to be a generation ahead of the attacks in the wild, and our LSI-Defense-1 LLM learns from every attack LSI-Offense-1 finds. Frontier labs give everyone the same model. We provide each enterprise its own customizable model, running on-premise and trained on its own private data."

"Chandra has spent his career building at each major wave of AI — Alexa, Uber AI, enterprise LLMs and guardrails at Got It AI, and India's first sovereign LLM. LSI is the natural next step." — Lu Zhang, Founder & Managing Partner, Fusion Fund

Availability

LSI-Offense-1 and LSI-Defense-1 LLMs are available now to a limited number of design partners (security vendors, MSSPs, and certain Fortune 500 enterprises) for air-gapped, on-prem, sovereign, or managed deployment. Request access at lsi.inc.

About LSI

Lean SuperIntelligence (LSI) is the first lab building self-improving offense and defense security LLMs, on the path to Security SuperIntelligence. Founded by Chandra Khatri (Krutrim, Got It AI, Amazon Alexa, Uber AI), LSI brings together sovereign and enterprise LLM builders and expert security researchers.

lsi.inc

Follow LSI: LinkedIn, X/Twitter

Claude Mythos is a trademark of Anthropic PBC. All other trademarks are the property of their respective owners.

References