In theCUBE Pod: Machine data poses a cybersecurity threat – SiliconANGLE

Warning bells are ringing for cybersecurity, such as machine identification fill the system faster than companies can store their data.

At the annual Nvidia GTC event, the CEO of Nvidia Corp. Jensen Huang introduced a series of open tools based on OpenClawan open source AI agent. The original tool, NemoClaw, was designed to protect OpenClaw’s results, but security experts are still worried.

“Everything that Jensen announced last week at GTC increases the range and area of ​​attack and risk,” he said. Dave Vellante (right), senior analyst at theCUBE Research. “That’s all we talked about [RSAC 2026]. Machine identification now outnumbers human identification as much as 80,000 to one. Google … shared with me that they discovered more than 800 OpenClaw exploits that can be downloaded by malware.

It is open the latest installment of theCUBE PodVellante and John Furrier (left), chief analyst for CUBE Research, discussed why the cybersecurity sector needs to change dramatically. They also explored the proliferation of machine learning, as well as the relationship between cloud-based technology and AI.

RSAC was the canary in the coal mine for old AI

The outlook for RSAC, the annual cybersecurity conference, was negative, according to Vellante, who called it “the canary in the coal mine.” Agent AI, if not addressed in time, can be a threat to cybersecurity, leading to the proliferation of machine identities that are difficult to track and secure.

“Shadow AI was probably the first or second topic at RSAC this year,” Vellante said. “CISOs are really worried about it. There’s no doubt that the fear is very high. One thing was that there are a lot of unknown unknowns. What’s the first line of defense now? Is it data? Is it agents? It’s changing.”

Companies will soon need a layer of security across their systems to combat AI-driven attacks. Business is already worried about getting a return on investment for AI, Vellante and Furrier pointed out, and the threat posed by agent AI is not encouraging.

“Everybody last year was drinking the Kool-Aid,” Furrier said. “Three years ago it was: Agents are bullshit. Last year, ‘Oh, agents are good. We can do the last job.’ This year, it’s like, ‘Oh shit, bad reps.’ A full circle. That’s right [enterprise] from humans to automated systems and from implementation to real-time machine speed.”

Machine information is the new Wild West

Machine identity is an uncharted territory for cybersecurity professionals. Instead of people talking to people, or agents talking to people, security now involves agents talking to agents. And if hygiene AI is the cause of the problem, it may have to be part of the solution, according to Furrier.

He explained: “People think irrationally. “But criminals use agents, and they are not people, and they have no choice. There is a complete structural change. This is a game-changer because it’s decades, generations of security principles being fully upheld.”

Everyone knows that agent AI is dangerous to cybersecurity but opinions are divided on what to do about it. Many companies have not yet fully embraced AI in their portfolios and implementing machine learning faster than it can be controlled can be a source of disaster, Vellante believes.

“Start with the government,” he said. “If you don’t have a strong governance structure and a strong culture of security and governance and enforcement, what’s going to happen? You’re going to get shadow AI just like we have shadow IT, just like we had shadow big data, except at this very dangerous time.”

Cloud-native meets AI-native

KubeCon 2026 in Amsterdam shed light on how cloud-native professionals are responding AI production requirements. Cloud-native infrastructure has become the substrate for AI-native technology, providing the governance needed to maintain models and agents.

“You’re starting to see the creation of this new architecture where cloud-native mojo comes to the table, but not in a leadership way to drive AI-native,” Furrier said. “It’s an old position standing on the shoulders of giant ideas, which is good because open source is growing.”

The AI ​​sector still needs a structure or framework to manage the growing amount of machine information, according to Furrier, and no clear leader has emerged. As the “system of intelligence” built around machine identification grows, businesses will need to know where they are in relation to their data, Vellante added.

“If you’re going to have a digital image of your business, and agents take action, you have to have transactions,” he said. “The place of this intelligence system or state of mind is really important. Businesses will work … “semantic layer, the transaction layer, all of that will be done around the data.”

Watch the full podcast below to find out why these industry experts were mentioned:

Jim Kavanaugh, CEO of World Wide Technology
Michael Dell, chairman and CEO of Dell Technologies
Peter Tuchman, stock trader, the Einstein of Wall Street
Bob Pisani, former CNBC reporter
Jensen Huang, president, co-founder and CEO of NVidia
Lena Smart, Representative – AIUC-1
Jon Oltsik, principal cybersecurity analyst at SiliconANGLE Media
Adi Shamir, Israeli journalist and inventor
Bipul Sinha, co-founder, chairman and chief executive of Rubrik
Pat Gelsinger, former CEO of Intel
David Floyer, featured analyst at TheCUBE Research
George Gilbert, principal analyst at CUBE Research
Danny Brickman, co-founder and CEO of Oasis Security
Ali Ghodsi, co-founder and chief executive officer of Databricks
Ben Horowitz, general partner at Andreessen Horowitz
Pete Hegseth, United States Secretary of War
Nir Zuk, founder and CTO of Palo Alto Networks

Here’s the full episode of this week’s theCUBE Pod:

Don’t miss the latest “theCUBE Pod” episodes. Contact us by registering with us RSS feed. You can also listen to us at Apple Podcasts or to continue Spotify. And for those who like to watch, check us out YouTube playlist. Listen now, and be part of the ongoing conversation.

Photo: SiliconANGLE