Cisco Systems Inc. said Monday it’s finalizing the acquisition of an artificial intelligence-focused security startup called Robust Intelligence Inc. for an undisclosed price.

Robust Intelligence has created a platform that’s designed to protect AI models and the data they feed on throughout the entire development-to-production lifecycle. The company first got the attention of Cisco a couple of years ago, when the networking giant invested in it through its Cisco Investments arm.

The startup’s expertise in terms of AI model security and governance is complementary to Cisco’s own security offerings, Tom Gillis, senior vice president and general manager of Cisco’s security business, wrote in a blog post.

“The combination of Cisco and Robust Intelligence means that we can deliver advanced AI security processing seamlessly into the existing data flows, by inserting it into Cisco security and networking products,” he promised. Cisco believes that this integration will result in superior visibility in terms of AI traffic, so its customers will be able to build and deploy AI applications in a more secure way, Gillis added.

In addition, Cisco believes Robust Intelligence’s tools can aid in threat protection for AI applications and models. The company said it will help to mitigate risks such as prompt injection attacks, which refers to incidents when malicious users try to prompt AI models to behave badly. It can also help to safeguard against data poisoning, jailbreaking and unintentional model outcomes, Gillis said. Other benefits for users include simplified configuration and user experiences, attack prediction and adaptive policy enforcement.

The company said it plans to bring Robust Intelligence’s automated model assessment and risk mitigation capabilities within the umbrella of Cisco Security Cloud, so customers will be able to secure their network, AI applications, models and users in the same place.

Gillis said customers are increasingly prioritizing information technology modernization and cybersecurity to safeguard their AI investments. He cites Cisco’s 2024 AI Readiness Index as proof of that claim. That study found that 95% of enterprises think AI will “significantly impact IT infrastructure and security.”

As a result, enterprises will require more robust security measures, Gillis said. “As organizations increasingly rely on AI applications for innovation and efficiency, safeguarding these applications and their underlying models becomes critical,” Gillis wrote. “Traditional security tools often fall short in addressing the unique challenges posed by AI, creating a pressing need for specialized solutions.”

The acquisition of Robust Intelligence accelerates Cisco’s pivot into the security business, as part of an ongoing effort to diversify its revenue stream and bring in more recurring sales. The company, once the world’s most valuable, has struggled in recent years with declining revenue from its core networking hardware business, which is focused on selling things like Ethernet switches and routers for data centers. Its customers are steadily migrating workloads to the public cloud, so they’re investing less in their own hardware.

Cisco has responded by focusing more on networking security, notably paying $28 billion to acquire the cybersecurity and observability firm Splunk Inc. Meanwhile, the AI boom has given it a sense of greater urgency, as enterprises are still struggling to come to grips with how to protect their AI investments.

To capitalize on this opportunity, Cisco earlier this year announced a $1 billion fund to invest in AI startups, and has already made significant bets on AI model makers such as Cohere Inc., Mistral and Scale AI Inc.

Image: SiliconANGLE/Microsoft Designer