
Meta claims $3.9B research tax credit on AI data centers
Meta saved $3.9 billion in 2025 by calling AI data centers research 'pilot models' under a 1981 tax credit, the New York Times reports.
The latest Policy coverage on Tech AI Wire.

Meta saved $3.9 billion in 2025 by calling AI data centers research 'pilot models' under a 1981 tax credit, the New York Times reports.

Anthropic's Claude for Government is generally available to U.S. federal and state agencies under FedRAMP High, with no seat fees and hard spending caps.

California Gov. Gavin Newsom vetoed SB-1130, which would have required recording lights on smart glasses and fined secret recording up to $1,500.

OpenAI, Google, Meta, Anthropic, Nvidia and xAI signed a voluntary White House accord to use outside auditors on frontier AI. It carries no penalties.

The FTC is investigating OpenAI, Anthropic and AI safety group METR under the FTC Act, with demands for documents and executive testimony due within weeks.

Since September 30, 2026, apps in Brazil, Indonesia, Singapore and Thailand need a Google-verified developer to install from seven app stores.

CISA added two Citrix NetScaler flaws, both scoring 9.5 of 10 on the CVSS severity scale, to its exploited-vulnerability list on September 27, giving federal agencies until September 30 to patch.

A 25% US tariff on imported chips took effect on January 15, 2026. Here is what it covers, what it exempts, and what it costs.

A Disallow AI Training toggle lets a site refuse training crawlers and stay in search. Mixed-use crawlers are 36.6% of verified crawler traffic.

Microsoft published absolute constraints for its MAI models on September 14, 2026, then said its current models are not yet trained on the document.

A 14-day comment period, named stakeholders and a merge request: GNOME has drafted a formal way to settle the decisions it now makes in chat.

The Pentagon put ChatGPT Mil and Grok for Government on its GenAI.mil portal on August 31, 2026. Both cleared Impact Level 5. Anthropic is still missing.

South Korea's science ministry named consortia led by SK Telecom, KT and Kakao to run free, unlimited AI for the whole country. Most usage must run on Korean models.

Debian developers picked the 'Responsible Use of Generative AI' option in a vote that closed on August 28, 2026. AI help is allowed, and disclosure stays optional.

Governor Greg Abbott ordered Texas agencies to stop funding Flock cameras on August 27, 2026, after officers were charged with using the network to stalk people.

A federal judge found the Pentagon's supply-chain-risk label on Anthropic violated the First and Fifth Amendments. A second case is pending, and an appeal is coming.

Meta will pay about $12 billion over 10 years, rising to $17.1 billion if rival platforms settle too. Teen accounts get a 2-hour default cap and a nightly lockout.

SourceHut will prohibit LLM-written code, tickets and emails from 10 September 2026. Enforcement is honour-based, with no automated AI detection tools at all.
One email every Friday with the week's top stories from all six desks: AI, dev tools, coding, the tech industry, startups and what's next. Free, no spam.