The lead
Meta brings BlackRock into the financing of its AI compute build-out
Meta and BlackRock have formed a venture to develop and own a data-centre campus in El Paso, Texas. Meta frames the deal as a way to pair its operating expertise with outside infrastructure capital as it scales Meta Compute for its long-term AI ambitions. The concrete implication is financial as much as technical: the next wave of AI capacity is being structured as an infrastructure asset, letting a hyperscaler add sites without funding every build solely from its own balance sheet. Read Meta’s announcement on the El Paso venture →
What else matters
NVIDIA launches an open security alliance for AI tools
NVIDIA and founding members have created the Open Secure AI Alliance to build and share open tools for responsible AI use and security. The group builds on the Linux Foundation’s Akrites initiative and OpenSSF work, with a stated focus on vulnerability remediation and disclosure. For security teams, the test is whether shared harnesses become practical defensive infrastructure rather than another industry badge. Read NVIDIA’s Open Secure AI Alliance announcement →
Anthropic rejects a blanket ban on open-weight models
Anthropic chief executive Dario Amodei says the company has not advocated banning open-weight models, including Chinese models, while arguing that capability and safety risks should determine policy. The distinction matters because the debate is shifting from open versus closed as an ideology to what controls are appropriate for systems with dangerous capabilities. Read Anthropic’s position on open-weight models →
Cognizant expands Claude deployment across its enterprise platforms
Anthropic says Cognizant will embed Claude in Flowsource, Neuro AI Engineering and Neuro IT Ops, expand its Claude-certified workforce and become a Global Premier Partner. More than 30,000 Cognizant associates have completed Claude training, according to the announcement. The practical signal is that large service firms are productising model access inside their own delivery systems, not simply reselling a chatbot licence. Read the Cognizant–Anthropic partnership update →
Meta funds 30 AI-glasses projects across accessibility and work
Meta has named 30 recipients of its AI Glasses Impact Grants across 18 US states, from hands-free coaching for construction trainees to daily-routine support for people with dementia. The nearly $2 million programme is not evidence that smart glasses have reached mass adoption, but it does put funded, specific workplace and accessibility use cases in the field. Read Meta’s AI Glasses Impact Grant announcement →
