Artemis announces $70 million in seed funding for AI security

Artemis has emerged from stealth with $70 million in seed and Series A funding led by Felicis.

The cyber security start-up says it has built an AI-focused security operations platform designed to detect and respond to attacks in real time. First Round Capital and Brightmind increased their stakes in the Series A, while Theory VC, Two Sigma and Lockstep joined the round alongside a group of industry figures.

Founded six months ago, Artemis says it is already deployed in production. The platform is processing billions of events an hour for enterprise customers in technology, banking and financial services, according to the company.

Artemis is entering a market where security teams are under pressure to handle fast-changing attacks and vast amounts of data across cloud, identity, endpoint and network systems. In its view, many established tools still rely on static rules and manual investigation, leaving defenders behind attackers.

Its approach centres on a data model built from a customer's own telemetry. Artemis combines behavioural log data across users, machines, cloud workloads and applications with business context, then uses that information to create customer-specific detections, investigate signals and trigger automated responses when needed.

One example cited by the company involved linking a privilege escalation in Okta with unusual API activity in AWS and presenting it as a single incident rather than separate alerts. In some cases, customers can use the system to isolate a compromised identity before an attacker moves further through a network.

The founding team brings experience from established cyber security and cloud groups. Co-Founder and Chief Executive Officer Shachar Hirshberg previously worked on security operations products at Palo Alto Networks and Demisto before joining AWS to lead GuardDuty. Co-Founder and Chief Technology Officer Dan Shiebler most recently led machine learning and AI work at Abnormal AI and earlier completed a PhD in machine learning at the University of Oxford.

"We built Artemis as an AI-native defense system from the ground up," said Hirshberg. "The question isn't whether this model wins, but who builds it best. Some of the largest and fastest-growing companies in the world are among our first customers, and we're able to deliver value to them on day one. That trust matters, and we intend to earn it every day."

Cost is also part of the company's pitch. Artemis says it works alongside existing security tools and complements legacy security information and event management systems without requiring all data to be ingested and stored upfront. Instead, it retrieves data on demand from a customer's existing cloud storage and log sources through federated queries.

That model can provide full visibility at a fifth of the cost of more conventional architectures, according to Artemis. The company also says customers can create custom detections in minutes and investigate environments in natural language rather than writing complex queries.

Early customer examples suggest Artemis is targeting large organisations with sprawling environments. The company says a technology company with thousands of employees used its first scan to identify multimillion-dollar cloud spend savings as well as shadow activity missed by other tools, including over-privileged accounts, undocumented integrations and workflows calling APIs with elevated privileges.

Another customer in a highly regulated sector with tens of thousands of employees is now completing investigations in under five minutes, according to Artemis. That represents a 96% reduction in time to resolution from its previous average, the company says.

Investors backing the round described Artemis as a response to a shift in how cyber attacks are conducted. They pointed to the use of AI in offensive activity and the need for systems that can match the speed of those attacks.

"Artemis has built a truly world-class team with rare depth at the intersection of AI and cybersecurity, solving a problem that's becoming increasingly urgent as attacks grow in frequency and complexity," said Jake Storm, General Partner at Felicis. "Shachar's track record in building enterprise-scale security products and Dan's deep AI research background make them uniquely suited to tackle this problem together. Just six months after founding and while still in stealth, the team has seen enterprise inbound driven purely by word of mouth and early results. That level of demand at this stage is rare and signals the scale of the opportunity."

Brightmind also returned for the Series A after co-leading the seed round. Gur Talpaz, General Partner at Brightmind, tied Artemis's market opportunity to the shift away from older security operations architectures built around alerts and human-led workflows.

"Having led the Humio acquisition at CrowdStrike that became the basis for their Next-Gen SIEM, I've seen exactly what it takes to disrupt large markets. I've heard the demand for a modern, highly complementary protection platform first-hand from the world's largest enterprises, and have a clear view that this is where the market is heading," said Talpaz. "AI-powered attacks require a fundamentally new approach to defense with AI at the foundation, and Artemis is building exactly that. We co-led the seed and doubled down in the Series A because of our core belief that Artemis is building something that others simply can't."