Corma, the Tel Aviv-based startup, raised a $125M Series A led by NEA at a $1B valuation three months after leaving stealth to scale its defensive AI model.
Key Takeaways
- Corma raised $125M in a Series A led by NEA at a $1B valuation, announced November 16, 2026.
- The round follows a $60M seed in August, bringing total funding to $185M.
- Server capacity grew 24-fold since the August launch, and the company reports Fortune 500 adoption.
Lead
Corma, an Israeli-founded cybersecurity startup building its own foundation model for defense, announced a $125M Series A on November 16, 2026. NEA led the round at a $1B valuation. Sequoia Capital, Khosla Ventures and Coatue also took part. The company emerged from stealth in August 2026 with a $60M seed led by Sequoia Capital, so it has raised $185M in roughly three months of public existence.
What Does Corma Actually Build?
Corma trains proprietary models specifically for defensive security, rather than wrapping general-purpose models from OpenAI or Anthropic. The models ingest security telemetry such as system events, audit logs and network traffic. They look for anomaly patterns over long time spans, which matters for multi-stage intrusions that unfold over days or weeks.
On top of the model, the company deploys autonomous agents that work inside a customer's existing security tooling and alongside human analysts. Corma says the agents investigate threats and handle defensive tasks without human intervention, and that they keep learning from each enterprise environment.
The company's pitch rests on an asymmetry. At launch it said that models from OpenAI and Anthropic could carry out end-to-end attacks in its tests but failed to defend against the same attacks. That is Corma's own claim, and no independent benchmark has been published.
Who Is Behind the Company?
CEO Alon Pluda founded Corma in 2025. He served in Israel's Unit 8200 until 2022 and previously co-founded Miggo Security. The team combines researchers from Google DeepMind with security specialists from 8200. The company is headquartered in Tel Aviv and San Francisco, and it has about 20 employees, according to Israeli business press.
In the funding announcement, Pluda described cybersecurity as "a battle of intelligence against intelligence." He has also argued that general-purpose language models are good at tasks like software development but remain insufficiently developed for defending organizations.
Why Did the Valuation Jump So Fast?
The valuation reached $1B because investors are paying for a scarce asset: a security-specific model with early enterprise traction. The Series A cheque is more than double the $60M seed, and the company went from launch to unicorn status in a single quarter. The same investors from the seed returned, which signals that insiders saw usage data they liked.
Corma says adoption among Fortune 100, Fortune 500 and Fortune Global 500 companies surged after launch, across healthcare, financial services, energy, critical infrastructure, retail and telecommunications. It named no customers. Israeli business press reported annual recurring revenue in the tens of millions, a figure the company has not confirmed in detail.
A skeptical reading is straightforward. A $1B price on a roughly 20-person company that is three months out of stealth values the model and the team more than a revenue base. Pilot deals with large enterprises can convert slowly or not at all, and security buyers tend to test new tools in parallel with incumbents before committing.
How Does the 24-Fold Server Expansion Matter?
The 24-fold increase in server capacity since August shows that Corma's costs are tied to compute, not just headcount. Training and serving a frontier-scale model for continuous log analysis is expensive, and much of the $125M will likely go to infrastructure. The company has not published a use-of-funds breakdown.
That cost structure separates Corma from the earlier generation of security startups that layered detection rules on top of cloud data. It resembles an AI lab more than a software vendor, with gross margins that depend on inference efficiency.
Competitive Context
Corma enters a crowded field. Established security platforms are adding AI analysts to their products, and several well-funded startups are pursuing autonomous security operations. Israeli founders are also prominent in the sector, and Israeli-founded companies took nearly half of Fortune's 2027 Cyber 60 list.
Corma's differentiator is owning the model. If a purpose-built model defends measurably better than a general one, that is a defensible position. If frontier labs close the gap with their own security fine-tunes, the advantage narrows.
Outlook
Corma now holds $185M in total funding, a $1B valuation and a customer base it describes as large enterprises, all inside one quarter. The next test is whether pilots become named, multi-year contracts and whether independent evaluations back its claim that its model defends better than general-purpose ones. Compute costs and competition from both security incumbents and AI labs will shape how quickly the valuation is earned.



