If you are building or planning an enterprise AI system, clients will start asking whether runtime protection is built in. This is no longer an internal technical question but a requirement appearing in RFPs and sales conversations, especially with banks and regulated companies. If you do not have a clear answer about how the system is protected during actual operation, this is a gap that needs closing before it becomes a business blocker.
What happened
Oligo Security, an Israeli runtime security firm, completed a $60 million funding round and reached total funding of $140 million. The company developed security for applications, cloud workloads and AI systems, combining deep runtime visibility with real-time protection. AWS selected Oligo as its exclusive partner for AI runtime security for AWS Security Hub Extended, and the company joined Palantir’s FedStart programme to accelerate the path to FedRAMP High and DoD Impact Level 5 certification.
The company recorded 300% growth in annual recurring revenue (ARR) over the past year and doubled its valuation since Series B.
Why it matters beyond the headline
When an Israeli company wins exclusive partnership with AWS and raises $60 million, it is a clear market signal: runtime security in AI systems has moved from nice-to-have to critical requirement. Recognition from AWS and Palantir confirms that large organisations view runtime protection as an essential layer, not an add-on.
For organisations in Israel building AI systems, especially in the financial and regulated sectors, this shifts the expectation bar. Clients who see AWS choosing an exclusive partner for runtime security will start asking similar questions about the systems they commission or build internally. The question is no longer “do we have security” but “how is the system protected while it is actually running”.
The 300% ARR growth indicates that organisations are paying substantial amounts for runtime protection. This is not a feature added in the next version, it is a component affecting purchasing decisions now.
What to do this week
If you are building an enterprise AI system, check whether you have a ready answer to the question “how is the system protected at runtime”. It is not enough to say there are code reviews or access permissions. Runtime protection means monitoring and defending what actually happens when the system runs, including API calls, data access, and model behaviour.
If you are working with an external vendor, ask them directly: what visibility do you have over what happens at runtime, and how do you respond to anomalies. If the answer is vague, that is a warning sign. Organisations building systems under banking or regulatory security standards need to plan the runtime layer as part of the architecture from the start, not as an afterthought.
If you are considering building a custom system, speak with someone who builds these systems under regulated standards. AI security for business systems requires early planning, not retrospective fixes.
Frequently asked
Any business processing sensitive data or working with regulated clients needs runtime protection, regardless of size. The question is not whether you are large enough, but whether you have data you cannot afford to lose or expose. If the answer is yes, runtime security is a requirement.
Regular security testing checks code before it goes live. Runtime security monitors what actually happens when the system runs, detects anomalous behaviour and responds in real time. It is like the difference between checking a car at the garage and a system that monitors driving on the road.
A short call, no commitment, and we will tell you whether it is relevant to your business or not.