Two more AI developer platforms disclose stolen customer keys by year-end
Why I called it
Wiz counted nine AI developer platforms breached in October, all with attackers hunting stored keys for OpenAI, Anthropic and Google models, and two vendors said the goal was . The keys convert straight into compute, the attackers used to find them, and nothing about the incentive ended on October 31st.
The call, in full. At least two developer platforms not on Wiz's October list disclose, between November 1st and December 31st 2026, a breach in which customer-stored secrets or model-provider were accessed.
Scoring criterion. RESOLVES CORRECT if between 2026-11-01 and 2026-12-31 23:59 UTC at least two developer hosting, deployment or agent platforms other than Fireworks AI, Andromeda, Zeabur, Lightning AI, OpenCode, Railway, Warp, CommandCode and Northflank publicly disclose (via their own post-mortem, status page or security advisory, or reporting by an established outlet) an intrusion in which customer-stored , secrets or model-provider API keys were accessed. RESOLVES WRONG otherwise.
The criterion is the machine-checkable version: a prediction that cannot be settled by a third party against a public source fails the build before it reaches this page.
Related
- Two more lab spin-outs of the Discovery Loop shape · 75%, open
- Palantir hits its raised FY2026 guidance · 80%, open
- A major breach names prompt injection as a root cause · 75%, open