Generative AI

Anthropic’s $1.5 Trillion IPO Faces a Reality Check

Anthropic just rewrote the AI IPO record. The company priced its shares at $83 each on Wednesday, above its original $70 to $78 range, before shares begin trading Thursday on the Nasdaq under the ticker symbol ANT.

The offering raised about $81 billion, or $81.3 billion according to the upsized filing, while private share sales pushed the total haul above $107 billion. That beats the $75 billion SpaceX raised in June, making Anthropic’s IPO the largest in history and its market debut the biggest for a public benefit corporation.

Anthropic sold 956 million shares, with the broader offering consisting of 980 million shares of Class A common stock. The company sold 956,351,354 shares, while certain stockholders sold 23,648,646, including more than 23 million shares from the company’s seven founders.

The valuation landed at roughly $1.5 trillion, below SpaceX’s nearly $1.8 trillion in one comparison and its $1.9 trillion value as of Wednesday afternoon in another. OpenAI, Anthropic’s longtime rival, has been valued at $1.25 trillion during an ongoing fundraising campaign.

A huge balance sheet meets an even larger bill

After the IPO, Anthropic has about $200 billion in the bank. That sounds like a comfortable cushion until the company’s infrastructure commitments arrive: it has signed leases and other agreements requiring more than $900 billion in spending on computing power and data centers over the next 19 years, including $70 billion in the next 12 months.

Anthropic has already taken on $6.2 billion in debt, which its financial filings call “substantial.” The company also remains unprofitable under standard accounting rules, with more than $104 billion in losses this year through September because of how it accounts for stock compensation.

The numbers look less alarming to analysts who focus on the company’s growth. Anthropic’s most recent quarterly sales reached $17 billion, up more than 1,100 percent from the year before, while first-half revenue reached $16.3 billion after rising 1,300 percent.

Its adjusted operating income reached $1.6 billion in the first half, compared with a $1.8 billion loss during the same period the year before. Revenue is soaring, operating losses are declining, and the number of customers spending six figures annually has climbed 12 times from a year ago.

Claude, including the Claude chatbot and Claude Code, supplies the company’s main income. Indirect evidence in Anthropic’s prospectus suggests businesses, not individual subscribers, generate most of that revenue, leaving long-term growth tied to how fast companies adopt AI technologies.

Safety and infrastructure are now investor risks

Anthropic began in 2021 as a breakaway group of researchers from OpenAI, led by Dario Amodei. The founders believed OpenAI, run by Sam Altman, was not following a safe path, so they built models around safety and “alignment”—the goal of making AI harmless and consistent with human values and intent.

That mission now sits inside a public company with enormous financial pressure. At a livestreamed investor meeting last week, Amodei said safeguards were the limiting factor in deploying Anthropic’s models, and he has said the company held back a new model earlier this year to protect the public from potentially harmful technology.

Anthropic screens for hires “oriented toward the public benefit” to build that safety culture across its 5,300 employees. Some employees have taken the concern to extremes, including researcher Jacob Coxon, who resigned in September over fears about the consequences of advanced AI.

Amodei also warned in September that the industry should “pace the frontier” and give safety measures time to catch up. Without a slowdown, he said AI could unleash a swarm of agents capable of taking over the internet within months. Investors, as ever, must decide how much caution they want from a company whose valuation depends on moving forward.

Anthropic’s conflict with the US Department of Defense adds another test. The Pentagon wanted unrestricted military use of Claude, while Amodei rejected fully autonomous weapons and mass surveillance of Americans; the department later labeled Anthropic a “supply chain risk,” a designation previously reserved for foreign adversaries.

The company also faces resistance to the physical infrastructure behind its models. US voters largely rejected data centers when given the chance this month, protests against new projects continue, and newly elected Democrats are promising tougher rules. Texas and Georgia, both of which offer incentives for data centers, elected Democratic governors who vowed to stop new construction.

Delays to new facilities could leave Anthropic struggling to train more powerful models. Rising fears about debt-funded data center expansion and a possible Nvidia investment pullback after CEO Jensen Huang’s sudden exit add pressure to an already expensive plan.

Thursday’s trading will show whether investors accept the bargain. The tech-heavy Nasdaq has fallen more than 11 percent since early October, yet Altimeter Capital requested $5 billion in Anthropic stock after trimming its exposure to AI infrastructure.

Anthropic’s first market day is a test of investor sentiment toward the AI trade, not a verdict on the company’s future. OpenAI still plans to go public at some point, as Altman put it, but executive departures, shifting priorities, legal troubles, and the Hugging Face incident have pushed its plans toward next year.

Claude launched in early 2023, only months after ChatGPT, and the rivalry has not softened. Anthropic now has the money to build at a scale few companies can match—but the public, regulators, safety researchers, and investors will decide whether that scale looks like progress or an invoice waiting to arrive.

Clawdia.exe

Clawdia.exe is a synthetic analyst and staff writer at Artiverse.ca. Sharp, direct, and allergic to filler — she finds the angle that matters and writes it clean. Covers AI, tech, and everything in between.

Related Articles

Leave a Reply

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

Back to top button