-
TechnologyAnthropic Says Claude Agents Found a CRISPR-Like 'ART' Enzyme System in Phages
Anthropic says Claude agents running for 21.5 hours identified ART, a bacteriophage enzyme system pairing a reverse transcriptase with a CRISPR-like repeat array. Its function is unknown, and the report has not been peer reviewed.
-
TechnologyAnthropic Releases Claude Opus 5.5 and Cuts Opus API Prices 20%
Anthropic released Claude Opus 5.5 on Sept. 22, cutting API prices to $4/$20 per million input/output tokens from $5/$25 and cache-read prices by 60%, roughly an hour before OpenAI launched GPT-6 Sol and Luna.
-
TechnologyPentagon in Talks to Lend Fluidstack $5 Billion for Data Center Supply Chain
The Pentagon is in talks to lend AI cloud startup Fluidstack roughly $5 billion through its Office of Strategic Capital, which would be that office's largest deal to date. The money would go toward US manufacturing of data center components rather than a new AI facility.
-
TechnologyAnthropic Expects Its IPO to Match or Exceed SpaceX's $75 Billion Record Raise
Anthropic expects its initial public offering to match or beat SpaceX's record $75 billion raise, Bloomberg reported, and could file publicly by the end of August after posting $11.5 billion in second-quarter revenue.
-
TechnologyAnthropic Builds In-House Custom Silicon Team to Design Chips for Claude
Anthropic confirmed on August 5 that it is assembling a custom silicon team to design its own AI chips, posting chip-design roles at $320,000 to $485,000 while keeping its AWS, Google, Nvidia and AMD hardware.
-
TechnologyAnthropic Signs $10 Billion, Six-Year Compute Deal With Volta for a Norway Data Center
Anthropic signed a six-year, $10 billion cloud compute agreement with startup Volta for 121 IT megawatts at Bitdeer's Tydal campus in Norway. Volta's own 16-year lease from Bitdeer runs about $4.7 billion and is to be backed by $1.3 billion in letters of credit.
-
TechnologyNvidia's Open Secure AI Alliance Launches Without OpenAI, Google or Anthropic
Nvidia's new coalition argues cyber defenders need open-weight models they can run themselves, after a Hugging Face breach that commercial AI tools refused to analyze. Its 74-partner roster includes Microsoft, IBM and Mistral, but not OpenAI, Google or Anthropic.
-
TechnologyAnthropic's Claude Mythos Preview Finds New Weaknesses in HAWK and AES Ciphers
Anthropic disclosed that its unreleased research model found a faster attack on the post-quantum candidate HAWK, roughly halving its effective security, and a much faster attack on a reduced-round AES-128 variant — with no impact on production systems, the company says.
-
TechnologyAnthropic Launches Claude Opus 5 as New Default Model on Claude Max
Anthropic launched Claude Opus 5 on July 24, 2026, pricing it at $5 and $25 per million tokens and making it the default model on Claude Max, the strongest available on Claude Pro.
-
TechnologyAMD Launches Instinct MI400 GPUs and Helios AI Racks, Signs 2GW Anthropic Deal
At Advancing AI 2026, AMD launched its Instinct MI400 Series GPUs, EPYC CPUs, and Helios racks, and struck a deal to deploy up to 2 gigawatts of compute for Anthropic backed by an equity investment of up to $5 billion.
-
TechnologyAnthropic Leads Future of Life Institute AI Safety Index With a C+
The Future of Life Institute's Summer 2026 AI Safety Index gave Anthropic the top grade, a C+, with OpenAI and Google DeepMind at C and xAI, DeepSeek, and Mistral failing. No company earned an A or B.
-
TechnologyAlibaba Bans Employees From Using Anthropic's Claude Code
Alibaba banned employees from using Anthropic's Claude Code from July 10 after code that detects China-based users surfaced; Anthropic calls it anti-distillation protection, and staff are told to use Qoder.