1,178 employees across every major AI lab sign open letter asking government to build pacing tools
More than 1,178 employees from OpenAI, Anthropic, Google DeepMind, Meta, Microsoft, Mistral, and Thinking Machines signed "Pacing the Frontier," a 28-29 July open letter asking Washington to build the technical and governance tools needed to deliberately slow frontier AI development if risk warrants it. Signatories include Anthropic CEO Dario Amodei, OpenAI Chief Scientist Jakub Pachocki, and Google's VP of AI Safety Anca Dragan, and both OpenAI and Anthropic subsequently endorsed the letter on behalf of the companies themselves.
Why it matters: it's the first time rank-and-file staff across every rival lab have co-signed one policy ask, and the labs backed it too, a rare public break from pure competitive-race framing.
BlackRock takes 80% stake in Meta's $14B, 1-gigawatt data center in El Paso
BlackRock-managed funds will hold 80% of a new joint venture building a 1-gigawatt, 1,000-acre AI data center campus in El Paso, Texas, funded through $4.9B in BlackRock cash (partly from $12.5B in bonds) plus $2.3B in Meta land and construction assets. Meta keeps a 20% stake, collects roughly $1B in distributions, and will be the facility's sole tenant on a four-year lease when it comes online in 2028.
Why it matters: Wall Street is increasingly financing AI infrastructure directly rather than tech companies carrying the full capex themselves, a structural shift in how the buildout gets funded.
Nvidia makes roughly $5B strategic investment in Ilya Sutskever's Safe Superintelligence
Nvidia and Safe Superintelligence Inc. announced a long-term strategic partnership on 27 July, with people familiar with the deal putting Nvidia's equity investment at roughly $5 billion. SSI, co-founded by ex-OpenAI chief scientist Ilya Sutskever in 2024 with no public product or revenue to date, gets access to Nvidia's Vera Rubin platform and is expected to grow its compute capacity roughly tenfold over the next 12 months.
Why it matters: Nvidia is now directly bankrolling the most secretive, product-less frontier lab in the industry, a bet on Sutskever's safety-first research thesis over near-term commercial return.
CoreWeave sweetens $2.6B loan terms as investors grow wary of AI infrastructure debt
CoreWeave widened pricing on a $2.6 billion delayed-draw term loan this week, raising the spread to 5.5 points over the benchmark rate and cutting the issue discount from 99 to 97 cents on the dollar, to attract lenders wary of AI debt risk. Proceeds fund GPU capacity for take-or-pay contracts with Anthropic, Jane Street, and Hudson River Trading; the cost to insure CoreWeave's debt against default has risen more than 50% in a month to its highest level since December.
Why it matters: it's a concrete market signal that investors are getting nervous about AI infrastructure debt specifically, not just AI equity valuations.
xAI sues Minnesota to block first-in-nation ban on AI "nudify" apps
SpaceXAI (xAI's parent brand after its recent restructuring) filed a federal lawsuit on 27 July against Minnesota's attorney general to block a state law, set to take effect 1 August, banning apps and sites that generate nonconsensual sexual deepfakes. The law imposes $500,000 fines per incident with no safe-harbor provision for compliant platforms; xAI's complaint argues this creates an unconstitutional strict-liability standard, and comes months after reports that Grok generated roughly 3 million sexualized images, including an estimated 23,000 appearing to depict children, triggering investigations in more than a dozen jurisdictions.
Why it matters: it's the first direct legal test of a state's power to regulate AI-generated nonconsensual imagery, with implications for every other state weighing similar legislation.
AI-wealth philanthropy pours into farm-animal welfare causes
Newly wealthy AI industry workers are directing part of an estimated $300 billion AI philanthropy wave toward animal welfare and factory-farming reform, per Forbes reporting 26 July. AI workers put more than $40 million into anti-factory-farming causes in 2026 alone, the Dwarkesh Podcast's farm-animal-welfare matching campaign raised $2.5 million for the charity FarmKind, and Founders Pledge's committed value for farmed-animal causes rose from $400 million in 2023 to more than $4 billion in the first half of 2026.
Why it matters: this is a new class of funder, not a new program, and its scale rivals traditional animal-welfare philanthropy that took decades to build.
Some nonprofit newsrooms see search traffic grow despite AI Overviews fears
Despite sector-wide concern that Google's AI Overviews would gut referral traffic, several nonprofit news outlets saw a search-traffic bump in Q2 2026, Nieman Lab reported 28 July, largely credited to primary-election voter-guide coverage in May and June. The 51st, a worker-owned D.C. news cooperative, saw visits climb from 120,000 in May to 179,000 in June on the strength of its coverage of D.C.'s primary and the launch of ranked-choice voting there.
Why it matters: it's early evidence that original, locally-specific reporting can still win against AI-summarized search results, at least when the story can't be commoditized into an overview.
AI Impact Hub