Free Intensive Nov 21–22Free 2-day hands-on Vibecoding Intensive on Nov 21–22 at 16:00 MSK (Sat–Sun) · Live Cohort 3 starts Nov 23Join Free →Cohort 3 (Nov 23)
IndustryInsight · On RadarThe Verge→ original

Anthropic Goes Public on Nasdaq: Shares Rise from $83 to $111, but Can Safety Hold Up?

Anthropic has become the first pure-play AI lab to list on a stock exchange. Shares were priced at $83 and reached $111 on the first day of trading. But the company now has to pay compute bills of roughly $900 billion while showing investors revenue growth. Analyst Steve Rattner estimates that Anthropic would need to grow revenue about 17-fold over ten years. The question to watch is whether it will be the first to cut expensive safety measures.

AI-processed from The Verge; edited by Hamidun News
Anthropic Goes Public on Nasdaq: Shares Rise from $83 to $111, but Can Safety Hold Up?
Source: The Verge. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceThe Verge ↗
Editor-in-Chief
Shareholders cannot change managementpower rests with the founders and a special trust
Key metric #2OpenAI is in no hurry to go public and is watching how the market values its rival
Key metric #3Small businesses decide for themselves which tasks go to the expensive model and which to the cheaper one

Anthropic, the company behind the Claude models, has listed on the Nasdaq exchange. It is the first pure-play AI lab to go public. The share price was set at $83, and on the day of listing it climbed to $111. For founders and employees, this is a celebration: according to Hayden Field, senior AI reporter at The Verge, the IPO will create thousands of new millionaires and hundreds of people with fortunes in the tens of millions of dollars. For those building products on Claude, something else matters more: what happens now to a company for which safety was the central marketing argument.

Money: bills are growing faster than revenue

The most uncomfortable fact is compute. According to regulatory filings cited by The Verge, Anthropic's obligations for computing capacity in the coming years total about $900 billion. That money has to be earned by selling tokens: the company rents capacity from cloud giants, pays for electricity and GPUs, and then must sell the model's answers for more than they cost it.

Analyst Steve Rattner estimates that to reach profitability comparable to other technology companies, Anthropic would need to grow revenue roughly 17-fold over ten years. Field stresses that investors want either profit or convincing signs that it will arrive soon. The corporate governance experts she spoke with expect that under such pressure some things will be pushed to the back burner, possibly including costly safety projects.

There is also counter-pressure on price. Satya Nadella, speaking at the Microsoft Ignite conference as the host recounted, said Microsoft will lower token prices and make model providers interchangeable within Copilot: the system will choose the best model for the task and the cheapest of the suitable ones. For Anthropic, which needs high margins on every token, this is bad news.

How Anthropic plans to keep customers

Field sees three levers. The first is raising prices, which has already caused a scandal: ordinary users complained that they were paying more for the same or worse quality, the dispute reached a class action, and the company partially rolled back the changes. The second is tying customers to its tools so that switching to another provider is too much of a hassle. The third is large enterprise and government contracts, where every penny does not have to be counted.

But large customers have their own conditions: data sovereignty and the ability to switch providers quickly. That is why, Field argues, Anthropic leans so heavily on the image of safe and reliable AI, including cybersecurity tools. This is what may keep a client when price and flexibility argue against it.

Small and mid-sized businesses are already looking for workarounds. Law firms and other companies the journalist spoke with are building internal benchmarks: which tasks really require Anthropic's expensive model, and which can go to an open or very cheap one. The goal is to send only a small share of tasks to the costly model, for example around 10%.

Governance and the OpenAI picture

Anthropic's governance structure is unusual. Shareholders, experts explain, cannot vote to replace leadership and have almost no influence over anything: real power rests with the founders and a special trust. Field adds that the founders' control is even greater than it appears. But, according to those she spoke with, the issue is not formal votes: market pressure will still force the company to earn revenue and buy compute in order to catch up with OpenAI.

OpenAI itself, in Field's assessment, is content that Anthropic went public first. Sam Altman was reportedly once obsessed with overtaking the competitor, but after a string of setbacks and public hostility, OpenAI decided to wait. Now it can point to the precedent and see how the market views the revenue of AI labs.

What this means in practice

For developers and product managers, the takeaway is down-to-earth. First, expect Claude's prices and plans to change: the company needs to grow revenue. Second, it is worth building task routing between models in advance, so you are not tied to a single provider and use the expensive model only where it justifies its price. Finally, watch how Anthropic behaves on safety. For now these are more the assumptions of journalists and experts than established facts, and only time will test them.

Story

Key Questions Answered

What are the key details on Money: bills are growing faster than revenue?

The most uncomfortable fact is compute. According to regulatory filings cited by The Verge, Anthropic's obligations for computing capacity in the coming years total about $900 billion. That money has to be earned by selling tokens: the company rents capacity from cloud giants, pays for electricity and GPUs, and then must sell the model's answers for more than they cost it.

How Anthropic plans to keep customers?

Field sees three levers. The first is raising prices, which has already caused a scandal: ordinary users complained that they were paying more for the same or worse quality, the dispute reached a class action, and the company partially rolled back the changes. The second is tying customers to its tools so that switching to another provider is too much of a hassle. The third is large enterprise and government contracts, where every penny does not have to be counted.

What are the key details on Governance and the OpenAI picture?

Anthropic's governance structure is unusual. Shareholders, experts explain, cannot vote to replace leadership and have almost no influence over anything: real power rests with the founders and a special trust. Field adds that the founders' control is even greater than it appears. But, according to those she spoke with, the issue is not formal votes: market pressure will still force the company to earn revenue and buy compute in order to catch up with OpenAI.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Unified AI Workspace · Nomad

Test Frontier LLMs & Autonomous Agents in One Workspace

Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.

Try Nomad AI →Launch Nomad Agent
What do you think?
Loading comments…