Welcome to the newest edition of ID Tech’s AI Update. Here’s the latest big news on the shifting landscape of AI and identity technology:
The White House is considering a “1:1” semiconductor rule that would require chipmakers to produce as many chips in the U.S. as their customers import from abroad, a move aimed at reducing American reliance on overseas production and accelerating onshore capacity beyond CHIPS Act subsidies. It’s not an official policy yet; officials are consulting with industry leaders. “America cannot be reliant on foreign imports for the semiconductor products that are essential for our national and economic security,” a spokesperson said.
The U.S. General Services Administration (GSA) has made a deal with xAI that will allow federal agencies to procure its Grok AI models at the nominal cost of $0.42 per agency over 18 months, with access running through March 2027 and including Grok 4 and Grok 4 Fast tiers. It’s a sign of rapprochement between President Trump and xAI CEO Elon Musk, coming just days after the estranged leaders were seated together at the funeral of political activist Charlie Kirk.
Cloudflare is going to launch a stablecoin for AI agent transactions. Its proposed “NET Dollar” will be backed by the U.S. dollar, meaning that, while it is a purely digital currency, it will be “fully collateralized” and pegged to American dollars. Cloudflare, one of the world’s biggest online network services providers, says its aim is to facilitate microtransactions, pay-per-use, and fractional payments across the “agentic web”.
Google DeepMind has unveiled a new pair of AI models for robots that will allow them to plan multi-step tasks and perform fine-grained actions like sorting laundry. Gemini Robotics-ER 1.5 is a reasoning model, while Gemini Robotics 1.5 can process language and vision, and possesses spatial understanding. It sounds futuristic, but it’s already rolling out to developers via the Gemini API in Google AI Studio.
Modular has closed a $250 million Series C funding round led by U.S. Innovative Technology (USIT) and featuring DFJ Growth, GV (Google Ventures), General Catalyst, and Greylock. The firm is looking to challenge Nvidia by building AI’s “unified compute layer” – a vendor-neutral platform that lets developers run AI across heterogeneous chips without having to make the adaptations that are normally required.
China-based Moore Threads has received approval for an IPO on the Shanghai Stock Exchange’s STAR Market board. Moore Threads designs GPUs and accelerated computing software and hardware that can be used for AI training and inference, among other things, and the firm is hoping to raise more capital for its heavy R&D spending. China has been steering chip IPOs to domestic venues like the STAR Market in part to support self-reliance amid U.S. sanctions.
Capital investment keeps flooding into AI data center efforts. The Qatar Investment Authority (QIA) and Blue Owl Capital signed a strategic partnership that will start with a more than $3 billion commitment to a data center-focused compute infrastructure program, while the UK’s Nscale closed a $1.1 billion Series B funding round – the largest in European tech history – to accelerate its buildouts. In the latter deal, Norway’s Aker is committing $285 million in cash and land and taking a 9.3 percent stake, while Nvidia has pledged up to $500 million in support.
The chatbot’s take: We asked for some explanation about the whole NET Dollar idea.
–
By Alex Perala
