Skip to main content

AI & Tech Brief: Cantwell signals openness on AI framework

March 27, 2026

The Lead Brief

We have our first credible sign that bipartisan negotiations on a national AI framework are possible.

In a statement that I obtained exclusively, the top Democrat on the Senate Commerce Committee, Sen. Maria Cantwell of Washington, is signaling that she is open to engaging on the issue and “committed” to developing a federal standard.

“What the United States needs is a federal standard that sustains innovation while giving consumers, businesses, and communities genuine assurance that AI is safe to use, reliable when it matters most, respectful of people’s work, and will not endanger our national security,” Cantwell said in the statement when asked for her views on the White House AI framework.

“That’s an important conversation to get right. The White House framework identifies key areas to address, and I am committed to developing a federal standard that works.”

This is important: Cantwell is the highest-ranking Democrat on the Senate committee with jurisdiction over AI policy. The Washington lawmaker has also acted as the lead Democratic negotiator on AI when Republicans twice attempted to enact a “preemption” of state AI regulations last year — interfacing with House Majority Leader Steve Scalise (R-Louisiana) in the lower chamber.

Cantwell also said she met with White House Office of Science and Technology Policy Director Michael Kratsios several weeks ago to discuss the administration’s latest effort to implement a uniform national AI standard.

The Legislative Chessboard

There will be several key nodes of influence if negotiations do kick off in earnest.

One name you should know: While Cantwell’s staff director Lila Helms is the highest-ranking Democratic staffer on the Senate Commerce, it’s Ali Nouri, Cantwell’s senior policy director, is expected to lead negotiations on AI issues. Nouri was formerly a deputy director for the Office of Legislative Affairs under the Biden-Harris administration and also worked in the Office of Science and Technology Policy.

There will be other Democratic negotiators who will be equally important if there’s a deal to be had. They include House Democratic Commission on AI co-chairs Reps. Ted Lieu of California, Josh Gottheimer of New Jersey and Valerie Foushee of North Carolina, plus their staffs.

Importantly, the top Democrat on the House Energy and Commerce Committee, Frank Pallone Jr. of New Jersey, will be a critical player in any negotiations — especially given a high level of skepticism of the framework from House Democratic leadership and House Minority Leader Hakeem Jeffries of New York.

On the Republican side: When I broke news two weeks ago that the White House framework was around the corner, I noted that the policy leads on the House Republican side have been Cyrus Artz, the senior policy adviser for House Speaker Mike Johnson (R-Louisiana), and Bijan “BJ” Koohmaraie, Scalise’s policy director.

But I also have a new name for you: The lead staffer on the administration side negotiating AI policy is Ryan Baasch, the deputy director for Trump’s National Economic Council. Baasch works intimately with Kratsios and David Sacks on AI policy, I’m told.

Sacks said yesterday that he was stepping out of his role as White House AI and crypto czar, and would continue to advise the administration as the new co-chair for the President’s Council of Advisors on Science and Technology. But don’t expect Sacks’s influence to diminish simply because his title changed.

What it all means: Cantwell’s public statement is significant, particularly because the Senate remains the central bottleneck for moving an AI national framework. While technically legislation containing national standards would only need a majority in the House to pass, Republicans would need to bring a minimum of seven Senate Democrats on board to pass something in the upper chamber.

Still, Cantwell stops short of endorsing provisions in the framework — and significant policy differences remain between Democrats and Republicans on this issue. The framework will also likely need House Democratic buy-in, which is looking quite difficult at this juncture. And Congress doesn’t have much time left before the midterm elections.

With progressives campaigning hard against data centers and for safety features for the AI frontier models, it also appears that there could be polarization on AI policy between the far left and moderates who might want regulatory clarity for businesses.

I’ll have more in Monday’s edition on a look-ahead on the congressional dynamics here.

Chip Smuggling

The Department of Justice announced charges on Wednesday against two American citizens and a Chinese citizen for attempting to smuggle millions of dollars’ worth of advanced chips subject to export controls to China.

According to the affidavit filed for the case, much of the evidence that allowed prosecutors to charge the suspects with the crimes came to light from a phone and laptop seizure from one of the defendants, Matthew Kelly. Kelly was going through customs in February on his way back from Rome on an international flight (border agents can seize phones without warrants at the border and ports of entry).

But there’s a deeper story to be told about how the chip-and-server producers first detected and headed off the smuggling operation, which initially involved an order of $170 million worth of computer servers from San Jose-based IT company Supermicro. Of the 750 requested servers, 600 contained advanced chips, including Nvidia’s H100 chips, that were at the time subject to export controls. Here’s what we know:

  • According to people familiar with the investigation, who spoke on the condition of anonymity to share confidential details, Nvidia raised flags in the first transaction proposed by one of the defendants in October 2023.
  • The first red flag was that the deal came from a business originator purporting to represent a buyer in Thailand. But Nvidia had never heard of this originator, according to one of the sources.
  • According to that source, Nvidia then sent its dealers in the Asia-Pacific region to check out this Thai company but couldn’t verify the legitimacy of the buyer. Nvidia shut the deal down.
  • The defendants then tried to obtain the chips through a different originator in the summer of 2024, but Nvidia raised red flags again (after local investigative efforts in Bangkok) and put brakes on the second proposed deal.
  • The chip company received grand jury subpoenas in June 2024 and May 2025 seeking documents on the transactions, and Nvidia employees were later interviewed for the case in May 2025.

“The complaint shows that our due diligence process works well. Despite several efforts, the would-be smugglers failed to clear our diligence process and did not receive GPUs from us,” a spokesperson for Nvidia said when asked for comment.

Why it matters: The incident shows that, in the wake of stricter export controls, Nvidia has trained its personnel to be adept at investigating noncompliance. Advanced chipmakers don’t just rely on intermediaries to ensure compliance with exports. Nvidia requires end-user certification even when the transaction happens further down the supply chain.

Expect Nvidia and other advanced chipmakers to point out their effective compliance systems as the House Committee on Foreign Affairs continues to push bipartisan legislation aimed at further constricting chip exports. Yesterday, the committee passed the Chip Security Act, which would enforce mandatory reporting for chipmakers and require local verification that they’re not being diverted.

The Strategic View

Anthropic clinched a big win on Thursday when a San Francisco court issued a preliminary injunction on the government’s efforts to blacklist the AI company.

Tim Hwang, general counsel for the Foundation for American Innovation, told me in an interview that the initial ruling is encouraging. The government has the authority to designate “supply chain risks” but must do so using the procedures outlined by Congress, according to an amicus brief filed by Hwang, with help from FAI Research Fellow Samuel Roland, in the D.C. circuity case.

“Everybody sees this definitely as a promising sign, but I think the right way to say it is we’re not out of the woods yet,” said Hwang, pointing to the fact that the circuit case is still playing out, along with the California case. “There’s two routes through which this is all being asserted.”