OpenAI just dropped a sweeping “Economic Blueprint” that sketches out how the United States can—and should—take the lead on AI infrastructure and policy.
And what’s in it has some huge implications for AI development, especially over the next four years of the newly elected Trump Administration.
To find out what it all means, I spoke to Marketing AI Institute founder and CEO Paul Roetzer on Episode 131 of The Artificial Intelligence Show.
At the heart of OpenAI’s new policy proposal is a single, striking data point: “There is approximately $175 billion in global funds awaiting investment in AI projects.”
OpenAI claims that if the US doesn’t mobilize now to secure these funds, China will. The solution? A comprehensive national AI strategy, featuring:
Given the huge economic promise, why is OpenAI rolling out this proposal at this moment According to Roetzer, timing is everything:
“With the new administration coming in, everyone’s lining up to get their messaging in place and build the relationships they need to build and have a say in what happens next,” he says.
In other words, OpenAI wants to shape the AI agenda in America before new policies or regulations become fixed in law.
Roetzer also cautions that the $175 billion figure comes from multiple layers of source material. It originally appears in an OpenAI paper called Infrastructure Is Destiny—but that paper cites a separate report from Houlihan Lokey, which itself is a dense analysis of digital infrastructure.
“Everyone latches on to these numbers with no concept of where the number actually originated from or how legitimate that number is," he says. He doesn't think it's necessarily inaccurate, but:
“It could be less, it could be way more—we just don’t know.”
Even so, OpenAI has built an entire economic argument around this $175 billion investment pool.
OpenAI’s Infrastructure Is Destiny report calls for a nationwide data center buildout—and it’s not shy about the numbers. Each 5-gigawatt data center would need:
The core argument? Massive expansion of data centers creates jobs, boosts local economies, and cements America’s position as an AI leader. All of which then plays into the policy recommendations in the Economic Blueprint.
"They think data centers, which are needed to build the future AI models and deliver all the AI that we need at inference time—all this intelligence we need at inference time when you and I use our smartphones and use ChatGPT and things like that—is a really big deal, " says Roetzer. "And it's going to be a massive driver of employment and GDP, specifically in the states where the data centers are built.
OpenAI also doubles down on policy alignment. They propose the federal government—together with industry—take the lead on crafting regulations to avoid what they call a “growing patchwork” of state and international rules that could slow down AI’s progress.
Interestingly, OpenAI uses a historical example to drive home its point about potential overregulation. They reference the 1865 Red Flag Act in the U.K., which once required a person waving a red flag to walk in front of every car on the road—slowing down adoption of automobiles in favor of horse-drawn transport.
"So they're sharing this as like a lesson of let's not over-regulate things," says Roetzer. "Let's like accept that change happens and it may look weird at first, but that we shouldn't actually restrict this.
In line with that, the blueprint wades into copyright issues. It argues that AI developers should be free to train their models using “publicly available information," including copyrighted content.
All of this happens against the backdrop of OpenAI’s push into scientific research. The company announced it’s now using an advanced AI model for engineering proteins, with a goal of tackling longevity science.
Altman personally funded Retro Biosciences, the longevity company partnering with OpenAI on the effort, to the tune of $180 million, according to MIT Technology Review.
The biggest takeaway?
OpenAI’s evolving “reasoning models,” like o1 (and the next in line, o3 or o4), are designed to solve very difficult tasks—including breakthroughs in biology and other hard sciences.
“They see the ability for these things to start solving really hard problems in society—and I think they want to prepare the government and the world for that,” Roetzer adds.
Finally, there’s a January 30 closed-door briefing in Washington that Sam Altman is hosting with lawmakers.
The question is: Will it just be Sam Altman and lawmakers, or is there a possibility of seeing surprise guests like Elon Musk in the room?
“I keep thinking Elon is going to be wherever Sam is, and I’m not sure the two of them have been together in person for a long time,” jokes Roetzer.
The stakes are high for America’s AI future, and OpenAI is making a bold play to influence everything from data center construction to how (and whether) copyright laws constrain model training.
If OpenAI’s blueprint has its way, we’ll likely see:
Whether you view that as an exciting blueprint for American competitiveness or a risky rush into the unknown, one thing is crystal clear:
OpenAI—and the entire AI world of AI—won’t wait for lagging regulations or uncertain investors to catch up. The race for AI dominance is happening now.