CEO: Funding Will Develop Exploit Blocking Throughout Cloud-Native and Agentic Apps
A startup led by a former Israeli Army Intelligence captain raised $60 million to increase runtime safety to agentic purposes as enterprises quickly undertake synthetic intelligence.
See Additionally: Why Conventional DLP Cannot Hold Up With AI Information Development
The funding – led by Ballistic Ventures, Canon Capital, Greenfield Companions, Lightspeed Enterprise Companions, Purple Dot Capital and TLV Companions – will assist Oligo determine exploitation because it happens and block it earlier than an attacker causes harm, stated co-founder and CEO Nadav Czerninski. He stated Oligo had developed the know-how earlier than agentic AI grew to become a serious enterprise safety concern.
“We’re constructing the main runtime safety platform to guard cloud-native and agentic purposes, and we’re truly the one resolution that may block assaults with out taking down the appliance in runtime,” Czerninski advised ISMG. “We try this utilizing our know-how that we name the deep app inspection, which principally permits us to see reside code execution in actual time in a non-intrusive approach.”
Oligo, based in 2022, employs 129 folks and has raised $140 million, having beforehand accomplished a $50 million Collection B funding spherical in January 2025 led by Greenfield Companions. The corporate has been led since its inception by Czerninski, who beforehand spent greater than six years main the Israeli Army’s cyber analysis crew (see: Oligo Safety Raises $50M to Deal with App Detection, Response).
How Securing Agentic Apps, Conventional Apps Differs
The normal interval between vulnerability disclosure and exploitation is disappearing, with attackers reverse-engineering a patch and exploiting the vulnerability earlier than organizations have had a practical alternative to deploy the repair. Czerninski stated these adjustments make safety approaches constructed round prolonged detection, investigation and remediation cycles more and more insufficient.
“We’re seeing the exploit window being quicker than ever,” Czerninski stated. “Prior to now, as soon as there was a patch out, you had time to patch it. Now we’re seeing a adverse quantity of days, which means as soon as the patch is out, attackers already leverage it. Organizations actually must have one thing to guard themselves.”
Agentic purposes introduce safety issues that typical purposes don’t, he stated, which means safety groups should perceive prompts, fashions, device calls, perform calls and the actions these elements in the end provoke. The objective is to hint exercise from the unique immediate by way of the agent’s choices and power calls right down to the system calls occurring on the working system, he stated.
“We added the flexibility to see not simply the motion that occurs on the working system, however we see agentic purposes and all the things that occurs in them from the immediate itself to these device calls and performance calls right down to the precise system name detect on the working system,” Czerninski stated. “The agentic purposes require us so as to add one other degree of deep visibility to observe from immediate to motion.”
Czerninski stated Oligo’s deep utility inspection already observes the appliance name stack and the exercise that in the end produces system calls. The first distinction is what Oligo does with the knowledge after gathering it. For agentic environments, he stated the corporate has developed detections and analytical capabilities that determine fashions and instruments and study how they behave at runtime.
“The principle distinction is how we productize round that information,” Czerninski stated. “We acquire the identical information, very deep, the deepest information that anybody can get from runtime, after which based mostly on that, the risk detections that we’re including and the other ways to extract the info to grasp what fashions are getting used, what instruments are getting used after which how they behave at runtime.”
How Many AI Apps Are in Manufacturing Environments?
Deep runtime visibility can deal with vulnerability prioritization and assault prevention. For prioritization, companies can decide whether or not an utility is definitely executing the weak perform related to a flaw relatively than treating each detected vulnerability as equally pressing. For assault prevention, the platform can decide whether or not a specific library ought to be making a particular system name, he stated.
“We are able to see what the precise features which are getting used are,” Czerninski stated. “And lots of instances, 90% of the vulnerabilities aren’t exploitable as a result of they aren’t working the weak perform.”
Oligo continues to see typical cloud-native purposes dominate manufacturing environments, however Czerninski stated it is also seeing rising use of purposes constructed round LLMs and brokers able to making device calls. Czerninski stated all of Oligo’s prospects have already adopted AI in some capability, and the variety of brokers working in enterprise environments continues to extend.
“We’re seeing an increasing number of the usage of AI purposes with the usual LLMs, but in addition equivalent purposes with device calls,” Czerninski stated. “And lots of instances we’re seeing our prospects shocked with what number of purposes that they’ve, what number of brokers they’ve working of their atmosphere with out them understanding.”
Czerninski stated Oligo intends to make use of the funding to broaden the protection and deployment choices for its exploit-blocking capabilities. One use case is persistent safety in opposition to unknown vulnerabilities in software program libraries that repeatedly produce safety flaws. One other is digital patching when a brand new CVE is disclosed. Oligo can situation a devoted safety designed to dam exploitation of that vulnerability.
“Relying simply on hygiene and counting on post-exploit detection of that’s simply not related anymore,” Czerninski stated. “In in the present day’s world, you must begin with truly blocking the exploit, and solely then patching and doing the routine hygiene that’s all the time vital.”
