Records challenge Rhombus Power's predictive AI claims
Retired US defense officials helped Rhombus Power pitch predictive AI tools to military leaders even as records and assessments questioned whether the products met its claims, according to an investigation by the Project On Government Oversight. The company secured classified work, a contract worth up to $200 million and a multiyear foreign military sale financed by US taxpayers.
CEO Anshu Roy said the company's Unblinking Eye platform forecast Russia's invasion of Ukraine months in advance, almost to the day. In a letter to Ukraine's defense minister, Roy said the forecast was accurate within 12 hours. Rhombus also claimed its software predicted North Korean missile launches, Iranian attacks and conflicts along the China and India border.
Emails show that WestExec Advisors helped Rhombus reach senior military officials. Records also show that military officials warned Rhombus personnel against bypassing formal contracting channels, while staff at one command believed the company was using senior relationships to compensate for product failures.
A former Rhombus vice president separately accused the company in a lawsuit of failing to report foreign contacts and mishandling classified information. Rhombus denied the allegations and said an investigation by an outside law firm found no evidence of compromised classified material. US Special Operations Command said no Rhombus tools are currently in use there after its contract ended in September 2025.
We hope you enjoyed this article
Consider subscribing to one of our newsletters like AI Policy Brief, Defense AI Brief or Daily AI Brief.
Also, consider following us on social media:
More from Regulation
Oct 29 Canadian Privacy Regulators Call for AI Rules in Health Care Oct 29 House Democrats seek Trump's AI chatbot records Oct 29 UK Business AI Adoption Doubles to 32% in 2026 Oct 29 New York Bill Would Let State Limit Compute for Risky AI Models Oct 29 UK AI Revenue Rises 41% to £33.6 Billion in 2025More from Defense
Oct 29 Foundry Robotics Raises $100 Million for Robotic Assembly Factories Oct 29 Leonardo Selects Shield AI Hivemind for Proteus Rotorcraft Oct 28 BigBear.ai Revenue Rises 31% and 2026 Guidance Increases Oct 28 Pentagon AI Projects Fall Short of Responsible Use Guidelines Oct 27 Swarmer Buys Vataga Drone Technology and Hires 15 StaffAI Policy Brief
Weekly report on AI regulations, safety standards, government policies, and compliance requirements worldwide.
Whitepaper
Tensordyne Napier: What If One Rack Could Do the Work of Nine?
Tensordyne
This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.
Read moreYou may also like
Former US officials back Anthropic in Pentagon blacklist appeal
Bridge Defense Wins Software Agreement With Up to $63 Million in Initial Capacity
Anthropic Expands Cyber Verification Program With Three Access Tiers
Anthropic Gives METR Employee Level Access for Safety Reviews
Pentagon AI Projects Fall Short of Responsible Use Guidelines
Daily AI Brief: the AI news that matters, in your inbox.