The Signal — May 12, 2026
The machine-learning toolkit that was supposed to stay on the defensive just scored its first confirmed kill on offense, and the biggest AI chip challenger in years is about to test whether Wall Street's appetite for alternatives to Nvidia is real.
Google Detects First Confirmed AI-Built Zero-Day Exploit
Google's Threat Intelligence Group has identified what it says is the first confirmed case of a criminal hacking group using AI to discover and weaponize a zero-day vulnerability. The exploit targeted a two-factor authentication bypass in a widely used open-source web administration tool and was built for mass deployment.
The detection came with an unusual twist: Google's analysts spotted telltale AI signatures in the exploit code itself. The code contained a hallucinated CVSS severity score, a fabricated rating that doesn't match any real vulnerability database entry, along with LLM-characteristic formatting patterns. Those artifacts helped confirm the code's origin.
Google coordinated with the affected vendor to patch the vulnerability before the exploit could be used at scale. GTIG described the discovery as "the tip of the iceberg," warning that AI-powered offensive capabilities are advancing faster than many defenders have anticipated. The group noted that while AI-assisted vulnerability research has been theorized for years, this is the first documented instance of criminals completing the full chain from discovery to weaponization using AI tools.
Sources: The Verge · SecurityWeek · CyberScoop · SiliconAngle
Cerebras Upsizes IPO to $4.8 Billion Amid 20x Oversubscription
AI chipmaker Cerebras Systems raised its IPO price range to $150–$160 per share, up from $115–$125, and increased the share offering from 28 million to 30 million shares. At the top of the new range, the company would raise $4.8 billion and carry a market valuation of roughly $34.4 billion.
The upsize reflects extraordinary demand. The deal is reportedly 20x oversubscribed, with total orders exceeding $10 billion. Cerebras is expected to price the offering around May 13, with trading beginning May 14, positioning it as one of the largest tech IPOs in recent years.
Cerebras designs wafer-scale processors purpose-built for AI training and inference, competing with Nvidia in a market where supply constraints have made alternative chip architectures increasingly attractive to hyperscalers. The oversubscription signals that investors see room for more than one winner in AI silicon, even as Nvidia continues to dominate. Whether that appetite survives contact with public-market scrutiny of Cerebras's revenue concentration and margin profile remains an open question.
Sources: Reuters · Bloomberg · Wall Street Journal · Investor's Business Daily
On the Editor's Desk
Thin day. A Wall Street Journal report on OpenAI employees who flagged a ChatGPT user's violent conversations before the Tumbler Ridge mass shooting circulated widely on aggregator sites this morning, but the underlying investigation was published in February and the related lawsuits were filed in April. We held it rather than frame stale reporting as breaking news. The EU-Anthropic Mythos access dispute is real, but it overlaps heavily with recent Mythos coverage on May 9 and May 10, so we held it rather than keep circling the same model-access story. Jensen Huang's CMU commencement speech generated quotes but no hard news. METR's latest productivity survey was cut after recent coverage of their evaluation work. DeepSeek's Huawei chip optimization was covered May 7. Two strong stories cleared the bar today; we'd rather give them room than pad the count.
