OpenAI was issued warrants in power and data-center infrastructure developer SB Energy that were estimated to be worth about US$5.5 billion, according to The Wall Street Journal, which cited draft documents prepared for SB Energy’s planned initial public offering.
The reported arrangement is significant—but the headline number requires an important qualification.
Warrants are not the same as cash. They generally give their holder the right to acquire shares under predetermined conditions. Their ultimate value can therefore depend on factors including the eventual share price, exercise terms and the success of a potential IPO.
The US$5.5 billion figure is consequently an estimated value, rather than US$5.5 billion that SB Energy simply transferred into OpenAI’s bank account.
Reuters said on August 31 that it had been unable to independently verify the Wall Street Journal report. OpenAI and SB Energy had also not immediately responded to Reuters’ requests for comment outside regular business hours.
Why would SB Energy give OpenAI such valuable warrants?
The answer appears to lie in one of the most valuable commodities in the AI industry today: a giant anchor customer.
The Wall Street Journal reported that the warrants were offered as part of SB Energy’s effort to secure OpenAI as a major data-center tenant before SB Energy goes public.
For an infrastructure developer attempting to convince investors that billions of dollars of future construction can generate dependable returns, having one of the world’s largest AI companies committed to leasing its facilities could dramatically strengthen its story.
That makes OpenAI more than simply another customer.
Its presence could potentially help SB Energy attract financing, justify new power projects and support the valuation investors are willing to give the business once its shares reach public markets.
OpenAI already invested $500 million in SB Energy
The relationship did not begin with the warrants.
In January 2026, OpenAI and SoftBank Group announced that they would each invest US$500 million in SB Energy, injecting a combined US$1 billion into the infrastructure developer.
Under that agreement, SB Energy was selected to build and operate OpenAI’s previously announced 1.2-gigawatt data-center site in Milam County, Texas.
SB Energy also agreed to become an OpenAI enterprise customer, using OpenAI APIs and deploying ChatGPT within its own workforce.
The partnership was tied directly to Stargate, the massive multi-year effort to expand artificial-intelligence computing infrastructure in the United States.
The result is an unusually interconnected arrangement.
OpenAI invests in SB Energy.
SB Energy builds infrastructure for OpenAI.
OpenAI leases that infrastructure.
SB Energy also buys OpenAI technology.
And now OpenAI reportedly holds warrants whose value could increase substantially if SB Energy’s valuation rises.
That circular-looking network of commercial and financial relationships is becoming one of the central questions surrounding the enormous amount of capital flowing into AI infrastructure.
Nvidia has now entered the equation—with far bigger numbers
The financial web widened further in August.
Nvidia agreed to invest US$1.5 billion in SB Energy and provide guarantees of as much as US$105 billion connected to OpenAI’s planned data-center development in Pike County, Ohio.
Reuters reported that the Nvidia guarantee is designed to help support OpenAI’s long-term lease at the SB Energy-developed campus.
Nvidia will also serve as the exclusive AI computing infrastructure provider at the site.
The proposed Ohio campus is enormous.
The PORTS-Pike Technology Campus is planned to reach as much as 8 gigawatts of IT capacity, with OpenAI expected to occupy the site under a 20-year lease.
OpenAI has said the project could support roughly 35,000 construction jobs during its buildout through 2032 and around 2,500 permanent operating jobs.
The company has also announced a US$40 million community grant commitment, alongside an earlier SB Energy commitment, as the companies try to demonstrate that the extraordinary infrastructure investment will deliver benefits beyond the AI industry itself.
The financial stakes are enormous
The numbers involved help explain why SB Energy’s coming IPO has attracted so much attention.
Reuters reported in May that SoftBank had hired JPMorgan, Goldman Sachs, Morgan Stanley, Citi and Mizuho to prepare a U.S. listing for SB Energy.
People familiar with those preparations said the IPO could arrive as early as September 2026, with SB Energy potentially seeking a valuation of more than US$50 billion.
That potential valuation gives added importance to OpenAI’s reported warrants.
If SB Energy succeeds in becoming one of the dominant infrastructure providers behind the AI boom, those warrants could represent an extremely valuable strategic asset for OpenAI.
But the structure also means the fortunes of the companies are becoming increasingly linked.
And that is where investors may start asking harder questions
The AI infrastructure boom increasingly involves companies acting simultaneously as suppliers, customers, investors and financial backers.
Nvidia sells the computing hardware.
It also invests in infrastructure companies that buy or host that hardware.
SB Energy develops the data centers.
OpenAI invests in SB Energy while simultaneously becoming one of its biggest tenants.
SoftBank backs both the infrastructure buildout and OpenAI itself.
These arrangements can make enormous projects possible much faster than traditional financing alone.
But they also create a harder question for investors:
How much of today’s extraordinary AI demand is being financed by companies whose financial success increasingly depends on one another?
Reuters noted that Nvidia’s infrastructure financing strategy has already triggered scrutiny over what some investors describe as potentially circular funding flows.
Nvidia CEO Jensen Huang has rejected that characterization, arguing that the company is using its scale and visibility into future computing demand to secure infrastructure that can support successive generations of Nvidia technology.
Power may ultimately matter as much as chips
There is another reason companies are willing to structure deals this aggressively.
The next phase of the AI race is increasingly constrained not only by access to advanced processors, but by land, power generation, grid connections, cooling capacity and construction speed.
Reuters reported that electricity availability and grid constraints have become major obstacles for new U.S. data centers as AI companies attempt to deploy ever-larger computing clusters.
For the Ohio project alone, SoftBank and SB Energy plan at least 10 gigawatts of new power generation, while approximately US$4.2 billion in regional grid infrastructure is planned through a partnership with AEP Ohio.
That means the AI race is rapidly becoming an energy race as well.
The companies that can secure electricity and build massive computing campuses quickly may gain an advantage that software alone cannot deliver.
The real story behind the $5.5 billion number
At first glance, the reported OpenAI warrants look like another breathtaking figure from an AI boom already filled with billion-dollar announcements.
But the more important story may be the structure forming underneath them.
OpenAI needs extraordinary amounts of computing capacity.
SB Energy needs enormous customers to support extraordinary infrastructure investments.
Nvidia wants those new facilities filled with Nvidia computing systems.
SoftBank wants to build an ecosystem capable of turning AI demand into infrastructure, energy and investment returns.
Each company therefore has a powerful reason to financially support the others.
If AI demand continues accelerating, that structure could prove extraordinarily valuable.
But if the economics of massive AI data centers eventually fall short of today’s expectations, the same web of investments, leases, guarantees and warrants could expose just how closely the industry’s biggest bets have become connected.
For now, the US$5.5 billion warrant estimate is grabbing the headlines.
The bigger question is whether it represents the value being created by the AI infrastructure boom—or just how much money the industry is now willing to put on the line to keep that boom accelerating.
Leave a Reply