Safety & Security

Google Projects Fully Autonomous AI Cyberattacks Within 6 Months

Google Threat Intelligence and Mandiant project fully autonomous agentic cyberattacks in three to six months, days before Palo Alto Networks reports earnings it needs to defend a $21.2 billion backlog.

Google Says Fully Autonomous AI Cyberattacks Are Likely Within 6 Months
Google Says Fully Autonomous AI Cyberattacks Are Likely Within 6 MonthsSwidaAlba / pixabay
By James Calloway5 min read

Why it matters

  • Google Threat Intelligence Group and Mandiant projected fully autonomous agentic attacks within three to six months in a Nov. 16 report.
  • A system called exploit_agent ran exploits 'without a human in the loop for the first time,' targeting already-patched flaws.
  • Zero-day exploits rose from an average of 8 per month in 2025 to 11 in 2026, hitting 22 in August.
  • Palo Alto Networks reports fiscal Q1 2027 results after the close on Nov. 18, with guidance of $3.300–$3.310 billion in quarterly revenue.
  • Palo Alto closed at $309.29 on Nov. 17, down 26.3% from its Oct. 6 peak of $419.91.

Google's Threat Intelligence Group, working with Mandiant, projected on Nov. 16 that sophisticated, fully autonomous agentic cyberattacks are likely within three to six months. The forecast landed two days before Palo Alto Networks (NASDAQ:PANW), one of Google's largest security rivals, reports fiscal first-quarter 2027 results after the close on Wednesday, Nov. 18.

The projection is not hypothetical hand-waving. According to the report, a system called exploit_agent — found on an exposed staging server — compared code before and after patches, wrote exploits, and ran them "without a human in the loop for the first time." That is the line the industry has been watching for, and Google says attackers have crossed it. The caveat matters: both the multi-agent and fully autonomous stages remain projections, not documented observations.

What did Google actually document?

The report places current attacker activity at a stage where AI tools run several steps of an attack with minimal human involvement. From there, it projects two further stages. Multi-agent, simple attacks with little to no human interaction are expected within one to three months. The fully autonomous stage follows within three to six months.

The documented tooling is specific. Threat actors used open-source AI penetration-testing frameworks named ARTEX, CAIRN, CyberStrikeAI, and PentAGI. One of them, CAIRN, targeted the U.S. aerospace sector.

The exploit_agent system went after flaws that already had patches. That detail carries the practical weight of the entire report: patching speed becomes the main defense, and patching speed favors automated security platforms over manual operations teams. In plain words, the attack side automated first, and the defense side now has to match it.

Why do these numbers reach security budgets?

The report arrived against a backdrop of rising attack volume, and the figures are concrete. Vulnerability disclosures and exploits nearly doubled between January and August 2026, according to the report's data. Zero-day exploits — attacks that hit flaws before any patch exists — averaged 8 per month in 2025, rose to an average of 11 per month in 2026, and reached 22 in August alone.

The financial losses are equally specific. In LLM-jacking cases, where attackers take over a company's cloud and AI accounts, single-day fraudulent charges exceeded $100,000. A direct loss of that size makes the spending case straightforward for a chief information officer.

Google's own report contains a counterweight to the alarm. The company says these AI tools give up stealth for speed and leave tracks, which helps detection businesses such as Palo Alto's Cortex and Unit 42. The takeaway in plain words: faster attacks are also noisier attacks, and the winners may be the vendors that detect the noise.

Is Google selling the fix it is forecasting?

The tension for shareholders is that Google is forecasting the threat and selling the defense at the same time. Also on Nov. 16, Google promoted Google AI Threat Defense, a product combining its Gemini models with Wiz, its CodeMender patching agent, and Mandiant. Palo Alto's own filing already lists Alphabet as a competitor that builds security into its products.

The competitive asymmetry is stark. For Alphabet (NASDAQ:GOOG, NASDAQ:GOOGL), valued at about $3.49 trillion, security is a side business — Google Cloud revenue grew 103% last quarter. Security is all Palo Alto does, and its Next-Generation Security annual recurring revenue rose 63% to $9.10 billion. In plain words, Alphabet can subsidize a security war; Palo Alto has to win it.

What happened to the stocks?

Palo Alto fell 2.15% to $309.29 on Tuesday, Nov. 17, its fourth straight decline and its lowest close since June 26, 2026. That left the stock 26.3% below its $419.91 closing peak of Oct. 6. It traded at $315.49 after hours at 7:59 p.m. ET. Alphabet slipped 1.24% to $280.20. No source ties either move to the report.

The rest of the security sector split on the day. Fortinet (NASDAQ:FTNT) rose 1.00%, while CrowdStrike (NASDAQ:CRWD) fell 1.63% and Zscaler (NASDAQ:ZS) fell 2.96%. Palo Alto carries the widest exposure of the group because it covers network, cloud, identity, and security operations — the same chain of steps that multi-agent attacks move through. That breadth cuts both ways: it positions Palo Alto to capture AI-threat spending, and it exposes the company to every link in the attack chain Google described.

What does Palo Alto's backlog show before earnings?

Palo Alto's fiscal 2027 guidance calls for revenue of $14.10 billion to $14.20 billion. Its remaining performance obligations — contracted revenue not yet booked — reached $21.2 billion. With the threat count rising and the stock well off its peak, the market now values that backlog lower than it did in October.

The pattern around earnings reports is the immediate risk. Palo Alto beat estimates in seven straight reported periods, yet the stock averaged -4.53% on report days. An early drop after Wednesday's numbers would match its own history, regardless of the underlying results. In plain words, good numbers have not protected this stock on earnings days lately.

What should investors watch on the Nov. 18 call?

Two points will carry the weight of the evening. First, whether management speaks directly to AI-driven attacks — the forecast from its rival hands them the opening. Second, whether quarterly revenue lands within the company's $3.300 billion to $3.310 billion guidance. If both check out, the case that AI-driven threats are lifting demand holds together. Revenue below that range would undercut it.

Google's three-to-six-month projection puts a clock on the question. If autonomous attacks materialize on that timeline, security budgets follow; if they do not, Palo Alto still has to defend a $21.2 billion backlog at a valuation the market has already marked down by a quarter.

Source: 247wallst.com

Share this article:

More from James Calloway

James Calloway

Show full bio

News editor covering people, launches and day-to-day practice in artificial intelligence and machine learning at AI In Context.

749 articles

Related articles

  1. Google Reports First Autonomous AI Attack That Finds and Exploits Flaws
  2. Google finds fully autonomous exploit pipelines in the wild
  3. Google Cloud warns attackers will turn from data theft to AI decision manipulation
  4. Google Threat Intel: AI Agents Now Turn Security Patches Into Exploits
  5. Google says fully automated cyberattacks have arrived

« Previous articleNext article »