- The U.S. needs an AI-era equivalent of the National Broadband Plan — but lacks focus and urgency
- China is out-executing the U.S. through infrastructure, financing and global strategy
- Public distrust of AI and hyperscalers is becoming a major constraint on future policy
America used to build the future; now it holds hearings, shouts at allies, lets hyperscalers drink the water and wonders why China is winning.
In this conversation, Blair Levin, chief architect of the 2010 National Broadband Plan, reflects on what the U.S. got right, what it missed and why it risks falling behind in the next phase of the global digital economy. From AI infrastructure to Chinese soft power, hyperscaler dominance to Roosevelt-era ambition, Levin offers a candid assessment of where America stands — and what it needs to do next.
The lesson from Levin’s career, and this conversation, is clear: America doesn’t just need better technology. It needs a return to strategic thinking, public purpose and the willingness to build.
Without that, the future won’t be something the U.S. designs. It will be something it reacts to.
‘It’s always a mixed bag’
Steve Saunders: You were the chief architect of the National Broadband Plan in 2010. What has the output of that been? Have we fulfilled its obligations?
Blair Levin: Let me start by saying that, like every effort in government, it's always a team effort. I played a role, but so did the 250 people who worked on it and members of Congress.
The second thing I would say is — it’s a mixed bag. There are definitely some successes, and there are definitely things we did that I wish we had done better. Some things we didn’t see at the time. It’s important to be honest about that.
But the most important thing is this: in 1996, Congress had a bipartisan consensus. We wanted faster, better, cheaper communications for all Americans, and the way to do that was through competition. The core service was voice.
The National Broadband Plan reflected the same consensus, but updated: faster, better, cheaper communications, now centered on broadband. It was both a continuation and a new chapter.
Saunders: We’ve moved from voice to broadband. Now we’re moving into an AI-enabled economy. Does the U.S. need an AI plan?
Levin: I think so. What the 1996 Act tells me is that we need a similar consensus now.
The FCC could play a valuable role in addressing key questions: what networks do we really need to fully utilize AI? What do we do about universal service? Cybersecurity? Protecting children?
One of my biggest regrets about the 2010 plan is we didn’t talk about the potential negative impact on children. To be fair, we didn’t see it at the time. But we had rules in the broadcast era to protect children — we should have done more in the social media era, and we should not make that mistake again with AI.
Sadly, the FCC today is focused on a number of things where you just go, “Are you kidding me?” They’re examining whether to revoke Disney’s licenses without evidence — something widely seen as a political response to Donald Trump. My point is they’re focused on things that are not going to matter in 10 years.
‘China is absolutely killing us’
Saunders: I look at the full stack required for the AI economy — networks, AI, broadband, but also energy, water, governance. We don’t have a plan for any of that. We’ve stopped investing in infrastructure the way we used to, while countries like Sweden, Singapore, Japan and China move forward. We even have a certain amount of autocracy, but we don’t get any of the benefits of it.
Levin: I agree, but slightly disagree on one point. There was an infrastructure bill under Biden. It pushed clean energy, broadband and other areas.
But what we’re discovering with data centers is we need far more energy, particularly clean energy. And China is absolutely killing us in that area. They’re going to be the car companies of the future.
Meanwhile, our current administration is blocking wind and solar projects already in progress. You can have different views on climate change, but why stop projects unless your priority is serving the carbon industry?
There are a lot of contradictions in policy. But what will it take to get out of this? Obviously, a change in political leadership.
Saunders: America used to shape global markets through institutions like the World Bank, insisting on free-market approaches. Now China shows up with financing, infrastructure, and execution.
Levin: After the broadband plan, I was sent to various countries to promote U.S. broadband policy. It was very obvious that the Chinese were already there and they were offering a better deal.
It wasn’t just traditional infrastructure. In Costa Rica, they built a football stadium for free. The Costa Ricans were grateful. That kind of engagement has consequences.
We didn’t have a strategy like that. And now, as America proclaims, “America First,” the rest of the world responds, “If that’s your strategy, we say America alone.” Traditional allies are rethinking their relationships. America no longer has the same trusted voice it once did.
‘Audience of two: Trump and Musk’
Levin: Brendan Carr is not a dumb guy. He could be doing positive things — on spectrum, satellites, copper retirement. But his focus is serving an audience of two: Trump and Musk.
That doesn’t mean Musk wins everything — but he wins a lot. And there’s no holistic effort to anticipate problems and solve them ahead of time. That’s what we should be doing with AI. We’re not.
Saunders: We still have a siloed system of governance. And now we have this face-sucking vampire of hyperscalers sitting on top of it doing whatever they want.
Also, I don’t have many good things I can say about Brendan Carr.
Levin: He is very good at using the levers of power, but that is very different than using the levers of negotiation and diplomacy. I negotiated deals for years — you don’t destroy the other side. You find a way where both sides get what they need.
But their view is: we control everything, you bow to us or else.
I was on a call with investors who asked why there are so many protests about data centers. I told them something they didn’t want to hear. The public looks at what hyperscalers are doing and sees this: tech gets rich; you pay more for water and electricity; your kids may not have jobs.
And you’re surprised that 85% of the public doesn’t like that deal? They’re not wrong.
AI is very unpopular; possibly more unpopular than Congress. Because these companies aren’t engaging with the public. They think they can just ram it through. That makes it very difficult to build a coherent national plan.
The spirit of Roosevelt
Saunders: Roosevelt talked about bold, persistent experimentation and universal service. He also brought kindness into the equation. We seem to have lost that. Roosevelt, for me, is what America should stand for. Bold, persistent experimentation but also empathy. I hope we get both back.
Levin: I hope so, too. Roosevelt brought a sense of universalism: we bring electricity to everyone, telephone service to everyone. We weren’t fighting tribes.
That spirit informed the National Broadband Plan. The most important thing we did was try to solve problems before people realized they were problems. We weren’t there to win elections. We were there so that 10 years later, there wouldn’t be a pothole on the information highway.
That’s what we should be doing with AI now. We’re not. And those potholes are going to emerge.
Some of the most important outcomes weren’t even in the plan. Comcast created Internet Essentials to address affordability. Google built fiber, which forced incumbents to invest. We caused a defection and that changed everything.