VFF - The signal in the noise
NewsTrending

NSA Uses Anthropic Model Despite DoD Blacklist

Read original
Share
NSA Uses Anthropic Model Despite DoD Blacklist

The NSA is using Anthropic's Mythos Preview model despite the Department of Defense attempting to blacklist the AI startup from government contracts, according to Axios. The report does not specify how the NSA is deploying Mythos or the scope of its use. This creates a notable tension between DoD policy and actual NSA operations, raising questions about enforcement and coordination across defense agencies.

  • NSA is actively using Anthropic's Mythos Preview model
  • DoD has sought to blacklist Anthropic from government contracts
  • Axios reported the discrepancy but did not detail specific use cases
  • Suggests potential misalignment between defense agency policies and operational decisions

This reveals a gap between stated government policy and actual agency behavior regarding AI vendor relationships. It signals that despite official restrictions, Anthropic's models are finding their way into sensitive government operations, which has implications for how AI governance and procurement rules are enforced across the federal government.

For AI founders and operators, this demonstrates that official blacklists or restrictions may not fully prevent adoption of advanced models by government agencies. It also highlights the complexity of navigating government relationships when different agencies have conflicting positions on the same vendor.

  • DoD blacklist policies may lack enforcement mechanisms or coordination across agencies
  • Anthropic's models are sufficiently capable that agencies find workarounds to use them despite restrictions
  • Government AI procurement and policy decisions are fragmented across different agencies with different priorities

Monitor whether DoD escalates enforcement of its blacklist or whether other agencies follow the NSA's lead in using Anthropic models. Also track any official statements from DoD or NSA clarifying the policy discrepancy and whether this leads to broader changes in how government AI vendor restrictions are implemented.

OneUpAI
OneUp Your Business. Get More Done. OneUp Your Business. Get More Done. OneUp Your Business. Get More Done.
Learn More
Share

Subscribe to the newsletter

The latest stories and analysis, delivered to your inbox.

Free. No spam. Unsubscribe any time.

Related stories

Anthropic IPO Investors Push for AI-Specific Metrics
TrendingNews

Anthropic IPO Investors Push for AI-Specific Metrics

Anthropic is preparing to file its IPO prospectus after Labor Day 2026, marking the first time public investors will see the company's full financial picture. Beyond standard regulatory disclosures, large institutional investors are pressing Anthropic to provide AI-specific metrics including token economics, revenue per gigawatt of compute, and net revenue retention, though the company has not committed to releasing these figures.

by Cory Weinberg· The Information
Nscale Touts $103B Contracted Revenue Ahead of IPO
TrendingNews

Nscale Touts $103B Contracted Revenue Ahead of IPO

Nscale is marketing approximately $103 billion in total contracted revenue to prospective investors, including a newly signed $45 billion computing deal with Anthropic. The neocloud infrastructure company is preparing for an initial public offering that could occur as soon as this month. The contracted revenue figure represents a significant milestone as the company moves toward going public.

by Cory Weinberg· The Information
Anthropic cuts agent costs 75%, adds enterprise safeguards
TrendingModel Release

Anthropic cuts agent costs 75%, adds enterprise safeguards

Anthropic released Claude Fable 5.1 and Mythos 5.1, its latest large language models, alongside a 75% cost reduction for cached context reads and a new Enterprise Frontier Safeguards security architecture. The release targets enterprise deployment of persistent agents capable of multi-hour problem-solving tasks. Fable 5.1 shows significant benchmark improvements across scientific research, coding, and business workflow tasks, though results are vendor-reported rather than independently verified.

by carl.franzen@venturebeat.com (Carl Franzen)· VentureBeat AI
Anthropic Locks $35B Compute Deal With Nvidia-Backed Lambda

Anthropic Locks $35B Compute Deal With Nvidia-Backed Lambda

Anthropic has signed a $35 billion compute capacity deal with Lambda Labs, an Nvidia-backed cloud provider. Nvidia will supply the chips for the data center providing the capacity and hold a stake in the arrangement. The deal underscores the critical role of GPU supply and cloud infrastructure partnerships in scaling large language model development.

by Tiffany Li· The Information