Anthropic has officially filed to go public, but its IPO prospectus makes clear that the company faces more than the usual risks of a young technology business. Its challenges include political opposition, uncertain regulation, power shortages, rising infrastructure costs, safety concerns and the possibility that customers may choose cheaper AI models.
Anthropic sent its Form S-1 to the Securities and Exchange Commission on Monday and said its stock would list on Nasdaq under “ANT.” The number of shares and their prices “have not yet been determined,” but the IPO could value the company at around $2 trillion when the stock is expected to debut in mid-November.
Politics Could Turn Data Centers Into a Liability
More than 90 pages of the prospectus focus on corporate risk factors, ranging from “catastrophic or existential risks to humanity” to the possibility of losing executives. One risk stands out as a likely source of near-term trouble: the “evolving and uncertain regulatory and political environment.”
That uncertainty arrives at a difficult moment. New representatives from the midterm elections will be sworn in in January, leaving Anthropic only a few months between its expected IPO and a possible shift in the political balance. The current Republican-controlled Congress has not put its political will behind constraining AI companies, either over the technology they produce or the resources they consume.
Voters across both parties have shown opposition to data centers and concern about AI displacing jobs. Anti-data center sentiment has also become a factor in the midterms. As of this August, Democratic and Republican candidates had spent more than $35 million on campaign ads opposing data centers, while nearly 8 percent of video ads mentioned the facilities, up from 1 percent a year earlier.
Ohio offers one of the clearest examples. The Senate race is considered a toss up, and the candidates’ positions on data centers could decide the election. A private August memo from the National Republican Senatorial Committee warned that Senator Jon Husted could lose his seat because of his record on the issue.
“More than any other thing in this race, data centers are the anchor hanging around Husted’s neck,” the memo said. It added, “If he loses and datacenters get the blame, politicians across the country will take notice — and they will not go near the next one.”
The Ohio Chamber of Commerce found in 2025 that the state gave $2.5 billion in tax breaks to data center facilities from 2017 to 2024, when Husted was lieutenant governor. Democratic candidate Sherrod Brown has made opposition to the facilities the center of his campaign, saying, “They eat up our land. They swallow up our water … then once they’re built, our electricity prices go up.”
Brown has called Husted “the face of data centers” and the “reason” Ohio has more than 200 of them. Recent polling found voters gave Brown an 11 percentage-point advantage over Husted on managing data center development.
Opposition Is Spreading Across States
The political pressure reaches well beyond Ohio. In New Jersey, Adam Hamawy and Watson Coleman have both supported slowing data center expansion, with at least 80 facilities already in the state. In Pennsylvania, Josh Shapiro and Stacy Garrity have accused each other of flip-flopping on the issue.
In Texas, James Talarico and Gina Hinojosa have promised to work toward repealing a bill Greg Abbott signed in 2015 that gave hyperscale data centers 20 years of tax breaks. Talarico also supports legislation requiring new facilities to use closed-loop cooling systems and giving local communities the power to reject data center projects.
Recent polling shows Talarico with a narrow lead over his opponent, despite Elon Musk spending more than $276 million to aid Republican Ken Paxton. If Talarico wins, he would become the first Democrat to represent Texas in the Senate since 1993.
Other candidates are targeting AI regulation itself. In Michigan, Abdul El-Sayed’s “AI Under Control” proposal calls for a pause on AI model training and a Cabinet-level agency to inspect AI labs. He has also supported restrictions on data center construction.
Anthropic’s own filing acknowledges the danger. “Opposition to data center development has also become more organized and coordinated across jurisdictions and our public profile may draw particular attention to projects associated with us,” the company said.
Growth Comes With Enormous Costs
The company’s financial numbers show why the debate matters. Anthropic said that as of June 30, 2026, it had raised $141.1 billion in capital, held $112.4 billion in cash and cash equivalents, and had $19.3 billion in short-term investments. Revenue grew 1,368 percent year over year during the three months ending June 30, 2026, while compute and infrastructure expenses grew 300 percent.
Anthropic’s draft filing from June reported an operating loss of $8 billion last year. The latest prospectus also revealed that the company is committed to about $400 billion in chip purchases and data center leases, while losing $81 billion in the first half of this year.
Energy is a central part of that risk. Anthropic warned that a lack of “sufficient electrical power” could delay its AI buildout, while negotiations, delays, cost overruns and payments for grid upgrades could push operating costs higher. Even if power comes from “on-site natural gas-fired generation,” the company said demand could make it more expensive.
The filing added, “we may bear these costs even if the underlying [power] capacity is delayed or never realized.” Rising temperatures and a higher risk of wildfires linked to climate change also threaten the company’s infrastructure.
Anthropic is trying to reassure investors that safety remains central to its mission. Its filing lists four AI danger categories: misuse, loss of control, authoritarian abuse and economic disruption. Dario Amodei, the company’s CEO, has published a three-step plan aimed at curbing AI’s rate of development, while the company warns that failing to emphasize safety could damage its “brand, business, financial condition, and results of operations.”
Still, moving slowly carries its own risk. Anthropic says customers may decide that cheaper or open-source models are good enough, concluding that “incremental performance gains at the frontier do not justify the additional cost.” It also says it cannot guarantee that its safety efforts will be adequate.
That tension runs through the entire IPO: Anthropic must build more computing capacity to compete, but that buildout consumes power, faces organized local opposition and draws political attention. The company says a highly capable frontier model could “become widely available within a few years.” Its stated goal is “to help humankind chart a safe course through this technological evolution, open new frontiers of discovery and bring forth an age of miracles.”
Based on
