In January 2025, just one day after Donald Trump’s second inauguration, AI technology leaders met in the Oval Office as part of the announcement of a massive $500 billion AI deal, dubbed “Stargate.”
OpenAI CEO Sam Altman talked about his newfound love for Trump, telling him during the event that “for AGI to be built here, we can’t do this without you, Mr. President.”
The company said that it is making $ 100 billion immediately, which sparked a heated debate about whether it has received the necessary funds and Elon Musk, the founder of the competition xAI, who was largely absent from the case.
More than a year later, OpenAI has made remarkable progress in its ambitions as reality continues to emerge. For one thing, its ambitious commitment to spend $1.4 trillion before the end of the decade on AI infrastructure was recently more than halved to $600 billion. Meanwhile, nervous executives want to cut back on “side requests” that disrupt business and coding, two areas that could generate much-needed revenue.
The rest of the company’s desire to save computing power has reached between a rock and a hard place – without more computing, the company risks falling behind the competition. On the other hand, attempts to acquire more resources could further reduce its use, which could scare off investors.
Ahead of its rumored IPO, pressure is mounting due to the company’s risky finances. Because of that, like CNBC reports, OpenAI has begun to sing a very different song compared to early 2025.
“OpenAI has realized that the market doesn’t appreciate a reckless approach to growth and spending,” Futurum Group CEO Daniel Newman said. CNBC. “The market wants to see OpenAI’s revenue continue to grow at a pace that the spending can justify. The goal was to try to show a little more fiscal responsibility.”
As the inside sources said CNBCthe company still does not have data centers directly, instead relying on purchasing cloud power from other companies such as Oracle, Microsoft, and Amazon.
Meanwhile, its ambitions to build data centers remain at a standstill. AI chipmaker Nvidia’s 100 billion investment in OpenAI, a project to install ten gigawatts of computing power using the original chips, may meet with big spirits.
That is, if it becomes a viable project. Nvidia CEO Jensen Huang rightly admitted earlier this month that the $100 billion figure was pulled out of thin air, saying his company’s recent $30 billion investment in OpenAI “may be the last” before OpenAI goes public.
Even with a reduced design, bringing data centers online can be very difficult.
Virginia Tech engineering professor Walid Saad said CNBC that it can take three to ten years to build a one gigawatt data center. OpenAI says such a data center will be ready for deployment before the end of this year.
“There are regulations, there are permits, different places have different processes,” he said. There are ways they cannot control. You never know what will happen.
Meanwhile, investors will be watching closely for any future moves by OpenAI as the company continues to burn billions of dollars on fire.
“Because of them, they’ve created an incredible growth story,” Newman said CNBC. “It’s just – the rest of the trip won’t be free.”
“And since their cost structure is very high, their profitability will be evaluated every step of the way,” he added.
More on OpenAI and data centers: Afraid OpenAI Executes Close-Up Wall Cutting Projects
#OpenAIs #Approach #Data #Centers #Runs #Troubled