Trump says Mark Zuckerberg showed him a ‘Manhattan-sized’ AI data center



President Donald Trump makes unexpected statement Yuan CEO Mark Zuckerberg after speaking at the World Economic Forum on Wednesday. In a Q&A after the speech, Trump recounted the moment Zuckerberg showed off a map of proposed AI data center facilities that covered Manhattan. Trump noted that the project’s footprint appears to engulf the entire island.

“Mark Zuckerberg showed me a factory and he put it on a map of Manhattan and it was basically the size of Manhattan,” Trump told a crowd of global elites. “I said, ‘You’ve got to be kidding me.’ It’s miles long, miles wide, and very tall. It actually covers most of the island.”

Meta shares rose 1.6% after Trump’s statement, before rising to 0.9% at press time.

While Trump calls the project a “big factory,” he may be referring to Meta’s Richland Parish “Hyperion” artificial intelligence Data center campus It is currently going up in Louisiana and is expected to open in 2030. This large industrial park covers 2,250 acres, equivalent to approximately 1,700 football fields. Once Meta completes construction, the campus will house 4 million square feet of data center space, making it one of the largest silicon clusters on the planet. However, those 4 million miles are far smaller than Manhattan’s 22 million square miles; raising questions about whether Trump was referring to a larger project Zuckerberg is building.

Trump also marveled at the project’s $50 billion price tag, contrasting the staggering investment with traditional real estate and recalling his past life as a developer. This $50 billion is comparable to the estimated cost of building a Hyperion data center campus.

“If you spend $500 million, you can build a good mall,” Trump said. “But how do you spend $50 billion? When I saw this thing, I understood why.”

Meta, which is currently transforming from a social media company into an artificial intelligence infrastructure giant, raised its capital expenditure guidance for fiscal 2025 to nearly $72 billion, a 70% increase from the previous year. Looking ahead to 2026, Zuckerberg warns market analyst spending will increase ‘significantly’ projection Annual spending could exceed $100 billion.

The president said artificial intelligence is “huge” and driving rapid demand for energy that requires a shift in policy. Recognizing that the aging U.S. grid cannot support the two to five gigawatts of power required by these “megaclusters,” Trump ordered the administration to step aside in his speech. He told the Davos audience that he had empowered artificial intelligence companies Acts as your own private utility.

So far, under this new framework, it looks like tech giants will build their own on-site power plants using natural gas, coal or oil. Trump promised to cut the bureaucratic red tape that often bogs down such projects, delaying them for four to five years. Instead, he promised the federal government would greenlight private energy plants in just two weeks as long as the companies built their “own power plants.”

“You’re smart. You’ve got a lot of money,” Trump told the room of tech executives, urging them to start building their own factories.

The industrial blitz is a major bargaining chip ahead of the president’s planned state visit to China in April. Trump believes that U.S. leadership in artificial intelligence depends on physical infrastructure, not just software, and says Chinese President Xi Jinping respects such rapid industrial execution.

“My relationship with President Xi has always been very good,” Trump said. “But we are way ahead of the world in artificial intelligence because I allow these big companies to build their own power. We create more energy than any country in the world.”

It is widely believed that China and the United States are evenly matched in the AI ​​race, with both countries emphasizing their own advantages to defeat the other. A recent report showed that the United States is ahead of China in AI talent, investment and access to semiconductors, while China beats the United States in AI-related infrastructure. Morgan Stanley Report. CEO Demis Hassabis says China is only “a few months” behind the U.S. in model quality Google deep thinking Tell CNBC.

On the same day that Trump spoke in Davos, NVIDIA CEO Jen-Hsun Huang was fired Worry about an AI bubble when discussing China’s progress. He particularly emphasized that the release of DeepSeek in 2023 is a “major breakthrough.” Huang called China’s progress in this area “a big deal for most industries, most companies in the world, because it was the world’s first open inference model.” Since then, many open inference models have emerged, he explained, allowing researchers to create domain-specific or something that specifically meets their needs. In the United States, Mark Zuckerberg stands out for his commitment to open source AI models, Specifically citing DeepSeek as a model.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *