Sam Altman, CEO of OpenAI Inc., speaks at the BlackRock Infrastructure Summit 2026 in Washington, DC, U.S., Wednesday, March 11, 2026.
Daniel Heuer | Bloomberg | Getty Images
When OpenAI CEO Sam Altman After speaking at BlackRock’s US Infrastructure Summit earlier this month, he acknowledged that his company faces a harsh reality: data centers are tough.
“On this scale, it feels like so many things are going wrong,” Altman said during a fireside chat at the conference in Washington, DC.
Altman gave the example of a severe weather event on a data center campus in Abilene, Texaswhich temporarily “brought things down.” The facility serves as OpenAI’s flagship site, Oracle and SoftBank’s $500 billion Stargate Project. Altman said his company also faces supply chain challenges and pressure to meet tight deadlines.
The stakes for Altman are rising as he aims to transform OpenAI, valued at $730 billion in a year. record fundraising last month, from a private market darling to an investable asset for a more discerning class of public market money managers. That means abandoning some big spending plans, shelving some ambitious projects, and accepting OpenAI’s role as a buyer of massive amounts of cloud capacity rather than a builder of massive data centers.
“OpenAI realized that the market doesn’t necessarily appreciate the reckless approach to growth and spending,” Daniel Newman, CEO of Futurum Group, told CNBC in an interview. “The market wants to see OpenAI’s revenue grow at a rate that can justify the spending. The pivot, in my view, has been to try to demonstrate a little more fiscal responsibility.”
This strategic shift means OpenAI may have to settle for doing less while trying to compete with Anthropic, Google and a host of other companies developing AI models, applications and capabilities. OpenAI trains and runs AI models that require enormous amounts of computing resources, including chips, processing power, memory and energy. Altman and other OpenAI executives have pointed out for years that compute is a major bottleneck for the company, which has raised astronomical sums of cash, including $110 billion earlier this year, including $50 billion from Amazon.
In a post on In November, Altman wrote that OpenAI and other companies “must limit our products and not offer new features and models because we face a very severe computational constraint.”
Until then, the big OpenAI story last year was the extreme lengths Altman went to secure his capacity. The company has signed a series of multi-billion dollar infrastructure deals with companies such as Nvidia, Advanced microdevices And Broadcom. Altman said in his November post that OpenAI was considering commitments of around $1.4 trillion over the next eight years.
The offers public markets shakenhas raised fears about a potential AI bubble and led many investors to wonder how OpenAI could afford to make such spectacular commitments with $13.1 billion in turnover for the year.
OpenAI’s most notable announcement came with Nvidia. The chipmaker, which is also the world’s most valuable company, agreed in September to invest up to 100 billion dollars in the startup for several years, the distribution of capital being linked to the development of OpenAI and the use of Nvidia technology. OpenAI has announced plans to deploy at least 10 gigawatts of Nvidia systems, with the first $10 billion in investment arriving alongside the completion of the first gigawatt, a unit of power roughly comparable to the electricity consumption of a mid-sized city.
THE press release said the partnership “enables OpenAI to build and deploy at least 10 gigawatts of AI data centers.”
Analysts told CNBC At the time, the deal was reminiscent of the seller financing that fueled the dot-com bubble in the late 1990s. Altman repeatedly worries swept away on OpenAI’s ambitious infrastructure plans, suggesting revenues would reach hundreds of billions by 2030.
But in recent months, as the company prepares for a potential IPO Later this year, OpenAI tempered expectations and set a more measured strategy. OpenAI told investors in February, it is now targeting around $600 billion in total computing spending by 2030, a figure expected to be more directly linked to its expected revenue growth.
The company also emphasizes discipline in other areas of its business. In December, OpenAI declared a “code red” to focus on improving its ChatGPT chatbot in the face of growing competition from Google and Anthropic.
A ransom for SimoCEO of OpenAI Applications, held a meeting for all with staff earlier this month about the company’s business, and said the company is “moving aggressively” toward high-productivity use cases.
“What really matters to us right now is staying focused and executing extremely well,” Simo said, according to a partial transcript of the meeting reviewed by CNBC.
“It’s the race”The Stargate AI data center in Abilene, Texas, United States on Wednesday, September 24, 2025.
Kyle Grillot | Bloomberg | Getty Images
OpenAI currently has no data centers, and may not for the foreseeable future, according to people familiar with the matter who asked to remain anonymous because they were not authorized to speak publicly.
Instead, it chose to rely heavily on partners like Oracle, Microsoft and Amazon, trying to cobble together as much capacity as possible.
A year ago, things were very different for OpenAI. In January 2025, the president Donald Trump revealed Project Stargate alongside Altman, SoftBank CEO Masayoshi Son, and Oracle Chairman. Larry Ellison during an event at the White House. The companies have committed to deploying $500 billion over four years to build new AI infrastructure in the United States.
OpenAI would be responsible for the project’s operations, while SoftBank would be in charge of finances, according to a blog post at the time. Oracle and Nvidia have been named as key initial technology partners.
“Oracle, Nvidia and OpenAI will collaborate closely to build and operate this computing system,” the statement said.
As Stargate was getting off the ground, OpenAI was set to develop much of the project itself and aimed to lease or directly own some data center campuses, according to a report from The information. But after the company faced practical construction problems and struggled to gain support from lenders, it pivoted.
Oracle leases the Stargate data center campus in Abilene and finances its construction by investing tens of billions of dollars in debt.
OpenAI and Nvidia said in their September statement that the first gigawatt of Nvidia systems would be deployed in the second half of 2026. Experts said the timeline would be tight at best.
Walid Saad, an engineering professor at Virginia Tech, said building a 1-gigawatt data center from start to finish could take three to 10 years. Challenges can arise at every step of the process: from finding a site, obtaining the appropriate permits and approvals, accessing electricity, building the physical structure, delivering the equipment, all the way to bringing it online.
“There are regulations, there are permits, different places have different processes,” Saad said. “There are processes they can’t control. You never know what’s going on.”
These obstacles have become very real for OpenAI, Arun Chandrasekaran, AI analyst at Gartnertold CNBC in an interview.
“They’re starting to say, ‘You know what, let’s try to get what capacity we can from suppliers who are willing to give us that capacity now,'” Chandrasekaran said.
OpenAI did not provide comment for this story.
As part of During OpenAI’s $110 billion funding announcement last month, the company agreed to consume about 2 gigawatts of Trainium capacity through Amazon Web Services infrastructure. Trainium is AWS’s custom AI chip. Amazon announced the latest version, Trainium3, in December.
Nvidia also contributed to OpenAI’s funding round, investing $30 billion. OpenAI said it expanded its collaboration with Nvidia as part of the deal and agreed to use 3 gigawatts of dedicated inference capacity and 2 gigawatts of training capacity on Nvidia’s upcoming Vera Rubin systems.
“OpenAI is doing what it needs to do, which is accessing computing at scale,” said Futurum Group’s Newman, adding that MetaAnthropic and Google do the same. “It’s racing.”
Nvidia’s investment landed after months of speculation on the progress of the major infrastructure deal announced by the companies in September. The chipmaker revealed in a quarterly filing in November that the $100 billion deal might not go through, and the Wall Street Journal reported in January that the deal was “on the ice”.
Nvidia noted in a February filing that there was “no guarantee” that the company would enter into an “investment and partnership agreement with OpenAI or that any transaction would be finalized.”
At a conference earlier this month, Nvidia CEO Jensen Huang controlled expectations even further, and said the opportunity to invest $100 billion in OpenAI was probably “not an option.”
The latest investment is not tied to any deployment milestone and is distinct from the deal structure the companies announced six months ago. Huang said “this may be the last time” Nvidia invests in OpenAI before its IPO.
“It must be recognized that they have built an incredible growth story. It is j “But the rest of the journey won’t be free,” Newman said of OpenAI. “And because their cost structure is so high, their path to profitability will be scrutinized every step of the way.”
–CNBC’s Kate Rooney contributed to this report
WATCH: OpenAI renews focus on company in IPO all-hands meeting
