AI

Anthropic IPO Puts AI's Business Model Under Market Scrutiny

Conceptual illustration of connected AI computing infrastructure in a financial district atrium.
AI-generated conceptual illustration. It does not depict actual company facilities.

The Claude maker overtook SpaceX's initial offering record. The companies' supplier relationship shows how AI rivals can benefit from the same fundraising boom.

[For more news, click here]

Anthropic ended its first trading day on November 19 with its shares up 17%, giving public investors a direct stake in advanced AI's economics. Its debut puts a central industry question into sharper focus: can growing demand pay for the infrastructure needed to serve it?

The 980 million shares priced at $83 on November 18 represented an $81.3 billion offering before fees. Reuters ranked it as the world's largest IPO, surpassing SpaceX's $75 billion June listing on a base-offering basis. That comparison excludes underwriters' options to buy additional shares. Some Anthropic shares were sold by existing holders, so the total does not all become company cash.

Shares opened at $111.50, about 34% above the offer price, and reached $115.10, a gain of approximately 38.7%. They closed at $97.17. Reuters reported a closing valuation of about $1.77 trillion on a fully diluted basis, which includes potential additional shares. The Nasdaq Composite fell about 2.3%, showing that enthusiasm for the listing did not extend across the wider technology market.

Warren Gibbon, chief investment officer at BFA Family Offices, described the shift to Reuters:

"It puts a daily price on what private investors have been paying, until now, for one of the leading AI companies."

Demand grows faster than certainty

Anthropic's commercial case rests on businesses embedding Claude in their work. Its IPO fact sheet reports third-quarter revenue of roughly $17 billion, up from $11.6 billion in the preceding quarter. More than 1,900 customers generated at least $1 million each in revenue over the year through September. That spending shows adoption extending beyond small experiments.

The challenge is matching demand with capacity secured ahead of use. In its November 17 filing, Anthropic disclosed about $900.9 billion in commitments, almost entirely for computing and infrastructure. About $70.7 billion was due in the 12 months after September 30. The commitments extend as far as 2045, with conditions or termination rights applying to some agreements.

Usage-based charges supplied 79% of revenue in the first nine months of 2026. The company recorded $974 million in operating income in the second quarter, followed by a $567 million operating loss in the third. The reversal shows why a growing customer base alone cannot settle the question of durable profitability.

SpaceX keeps a customer

The consolation for SpaceX is that the new record holder also pays it for computing. On listing day, Anthropic CEO Dario Amodei told CNBC that the service was high-quality and that Anthropic enjoyed working with SpaceX. He left future contract decisions open.

Eight days earlier, Elon Musk had addressed speculation about Anthropic's computing needs. In a November 11 post on X, reported by Benzinga, SpaceX's chief executive wrote:

"We would be happy to take the compute back. No problem. But I think the opposite is going to happen."

The commercial relationship predates the change in IPO rankings. Anthropic's May agreement covered all computing capacity at SpaceX's Colossus 1 data center, including more than 220,000 NVIDIA processors. Anthropic said the added capacity helped it raise Claude usage limits.

Record-book rivalry aside, the financing strengthens a customer of SpaceX's infrastructure business. Anthropic gains resources to serve demand, while SpaceX earns revenue supplying capacity. Their competing AI products, Claude and Grok, coexist with that supplier relationship.

Suppliers also hold the shares

NVIDIA sits on both sides of this expansion. Its September 30 holdings disclosure valued its SpaceX stake at about $19.7 billion. SpaceX also announced an approximately $40 billion GPU financing commitment in October through an Apollo-managed entity. The planned lease would supply NVIDIA Vera Rubin chips, subject to closing conditions, with deliveries expected in early 2027.

Anthropic separately secured a $26 billion private placement alongside the IPO. Participants included NVIDIA and other major chip suppliers. The arrangement gives infrastructure providers an equity interest in a business whose growth supports demand for their products.

In April, Amazon announced a $5 billion investment alongside Anthropic's commitment to spend more than $100 billion on AWS technologies over ten years. Financing, computing supply and customer spending meet within the same partnership.

These connections allow several companies to benefit from the same expansion. They also spread exposure to weaker demand across customers, suppliers and shareholders. An IPO record can change hands while both businesses retain a commercial interest in each other's success.

Physical constraints remain. The International Energy Agency's October report describes strained grids, equipment bottlenecks and lengthy connection queues. Capital alone cannot remove those limits. Anthropic also enters public markets as a public benefit corporation, with a Long-Term Benefit Trust alongside shareholders. Its safety mission adds another consideration when commercial pressure favors faster deployment.

The IPO makes AI's financial interdependence easier to examine. Its lasting significance rests on whether end users find enough value to sustain the companies financing their access.

Related Articles

Abu Dhabi Launches AI Tools to Rewrite Arabic Texts for Different Reading Levels
AI

Abu Dhabi Launches AI Tools to Rewrite Arabic Texts for Different Reading Levels

Mitsubishi Heavy Industries and Sakura Internet sign MoU on sovereign AI infrastructure in Japan
AI

Mitsubishi Heavy Industries and Sakura Internet sign MoU on sovereign AI infrastructure in Japan

Nscale brings first NVIDIA Vera Rubin NVL72 systems online for AI cloud workloads
AI

Nscale brings first NVIDIA Vera Rubin NVL72 systems online for AI cloud workloads

An attendee uses a tablet at KubeCon North America 2026 in Salt Lake City.
AI

KubeCon Research Finds Higher AI Use Among Cloud Native Developers

BeyondTrust predicts AI agents will reshape cybersecurity risks in 2027
AI

BeyondTrust predicts AI agents will reshape cybersecurity risks in 2027

Etisalat
AI

Etisalat UAE Unveils AIO as e& Expands Its Sovereign AI Push