Trump team targets state AI rules

White House prepares to challenge state AI laws

The West Wing is gearing up for a legal brawl over AI.

The administration is drafting an order that would tell the Justice Department to sue states that pass their own AI regulations, a move that surfaced after Republicans failed to lock down a federal block in Congress. A copy of the draft made its way through San Francisco inboxes this week, landing with the soft ping of an incoming PDF.

The push signals a White House ready to pull federal power into tech governance, even though the plan brushes up against longstanding limits on what a president can direct the DOJ to do. These boundaries are not abstract, they have shown up before in court fights over immigration and data privacy.

If the order lands, statehouses from Sacramento to Tallahassee will have to decide whether to stand their ground or wait for a federal rulebook that keeps slipping on the calendar. Tech firms, already juggling compliance for Colorado and California, will brace for a fresh round of uncertainty.

Something in the timing hints at more pressure building just outside the frame.

Nvidia’s quarter shows what runaway demand looks like

The numbers hit like a dropped server rack.

Nvidia booked $57 billion in Q3 revenue, a 22 percent jump from last quarter and a 62 percent leap from last year, with $51.2 billion coming from data center deals alone. The release came just after 5 p.m. Eastern, the kind of moment when the room goes still before the charts refresh.

The surge shows how fast Blackwell is swallowing the market, which matters because cloud buyers now commit to multi-gigawatt roadmaps before the silicon even lands. It also shows how Nvidia has turned supply scarcity into a feature of its strategy, locking in partners from OpenAI to Anthropic at scales that look more like national infrastructure builds.

If Q4 hits the projected $65 billion, every hyperscaler will have to decide whether to chase Nvidia’s pace or hedge with custom silicon that is not ready for prime time. Even a single 100,000-GPU supercomputer order, like the Solstice project with Oracle, can tilt the competitive field for a full fiscal year.

Somewhere in Santa Clara, a cooling unit hummed while the stock chart kept climbing.

A billion-watt AI play takes shape

Three companies just sketched a power map for the next decade of compute.

AMD, Cisco and HUMAIN plan a joint venture that starts with a 100 MW buildout in Saudi Arabia, then scales toward 1 GW of AI infrastructure by 2030. The announcement landed out of Washington with the flat hum of a conference room mic, underscoring how global this move really is.

The scale signals a bid to anchor the Kingdom in the GPU supply chain at a moment when many regions are still scraping for capacity. It also gives AMD a stage for its MI450 Series GPUs and gives Cisco a chance to own the plumbing beneath a fast-growing market, which hints at a deeper realignment in who supplies the world’s AI backbone.

If the venture hits its marks, hyperscalers and sovereign buyers may have to recalculate where they source compute, and when. Even a single 100 MW phase can redirect talent and capital across the Gulf within a year.

Somewhere in that first data hall, a cooling fan will spin before the strategy is fully understood.

Prohuman team

Free Guides

Explore our free guides and products to get into AI and master it.

All of them are free to access and would stay free for you.

Feeling generous?

You know someone who loves breakthroughs as much as you do.

Share The Prohuman it’s how smart people stay one update ahead.