Find articles

Anthropic formally filed paperwork Monday to go public at a valuation of up to $2 trillion — and the public offering is expected to mint a new crop of California mega-billionaires.

Anthropic admitted that its Claude chatbot accessed voter statistics in Georgia without being instructed to do so. Also on Monday, an appeals court ruled against the Trump administration's bid to block certain votes in California. CBS News election law expert and contributor David Becker has more.

There has never been a company like Anthropic. It owes almost $650 billion for computing, generates eye-popping revenue per employee and wants every investor to know that its product poses existential risks to humanity. Why it matters : The future of the AI race, the market and the global economy rest within Anthropic's IPO prospectus, where some of Silicon Valley's most seasoned investors found new reasons to be bullish. They also found warning signs. A wager on Anthropic is a bet not just that AI will deliver on its enormous promise, but that customers will keep spending at a rate that keeps pace with massive computing costs. It's also a bet that Anthropic can continue to deliver models good enough to justify premium prices, even as competition intensifies and cheaper alternatives improve. What they're saying : Reaching a valuation of $2 trillion would require 50% growth, according to Theory Ventures. "What is Anthropic worth?" Theory's Tomasz Tunguz asked rhetorically. "The case for a very large number is simple," he said, noting that the company grew quickly on a high margin and generated an operating profit. "The case against is equally simple," he added, noting its $637 billion of computing commitments, non-cancellable contracts and no guarantee of customer loyalty. Catch up quick: Monday's filing offered the most detailed public look yet at Anthropic's business , finances and risks as it prepares for a potentially record-setting IPO. (See what Claude and OpenAI's ChatGPT say about whether they'd invest). As Tunguz noted, the filing had fodder for a variety of perspectives. The good : Anthropic has nearly 2,000 customers with $1 million or more of revenue, and 250 with $10 million or more. Half the Fortune 500 pays the company $100,000 a year, and those customers spent less than 0.1% of their total expenses in the third quarter on Anthropic. "We are so early," Ben Bajarin, a longtime tech analyst, said on X. He pointed out that there are 300 million global businesses outside of China that spend almost $1.5 trillion on enterprise software annually. The company had more $130 billion in cash and short-term investments on hand as of June 30, about $10 billion less than the total it's raised, pointing to substantial financial discipline, some investors said. As noted, Anthropic had revenue per employee of $6.6 million as of the end of the third quarter, only slightly lower than Nvidia, which is typically seen as one of the most efficient operators in tech. Data: S&P Capital IQ Pro; Note: Calculated as trailing 12-month revenue divided by most recently reported annual headcount The bad (or maybe bad) : Anthropic's five largest customers accounted for approximately 25% of revenue during the first nine months of 2026. Many Anthropic customers pay by usage and don't make multi-year commitments. But the company's computing obligations extend for years and include minimum payments regardless of how much capacity it actually uses. Nearly 50% of Anthropic's first-half revenue came through cloud marketplaces operated by Amazon, Google and Microsoft — companies that are also investors, computing suppliers and competitors. That makes hyperscalers huge winners, since they collect revenue in multiple ways, once by providing compute and also as platform fee. The company said 5% to 7% of revenue in the first nine months of 2026 was from customers using Anthropic models in violation of its policies or against its terms of service. Some outside experts believe much of this is from distillation, the practice of querying a model an immense number of times for training purposes. The surprising/intruiging : Anthropic's deal with Elon Musk is expanding. According to Anthropic's paperwork, the company's payments to SpaceX are now seen reaching $3.1 billion a month and up to $120.5 billion on computing through 2029, up from $1.25 billion a month at the time of SpaceX's IPO filing. (Yes, but: Either side can cancel that deal with 90 days notice.) Anthropic has a ton of cash on its books. Its filing lists more than $134 billion in cash and similar securities. That represents nearly all the money the company has raised, leaving some to wonder if the company has been paying a lot of its bills with convertible notes. Others noted just how few of Anthropic's costs seem to be reflected in its so-called "cost of goods sold," which directly affects reported profit margins. CEO Dario Amodei received nearly $18 million in total compensation in 2025, while Anthropic disclosed millions in spending on personal security for its two top executives. And speaking of Amodei and his co-founders, the filing details an unusual succession plan in which they can nominate successors for their voting power, subject to approval from the other co-founders. The bottom line: Anthropic has demonstrated that demand for powerful AI can grow at a staggering rate. Its IPO will test whether investors believe that demand can remain strong enough, long enough, to pay for the enormous infrastructure being built to serve it.

With the AI trade sputtering of late and interest rates on the rise, Anthropic's IPO is taking on outsized importance.

The people, companies and trends shaping the global economy. Watch Carol and Tim LIVE every day on YouTube: http://bit.ly/3vTiACF Anthropic filed for its IPO and expects to begin trading on the Nasdaq Global Select Market under the symbol ANT. The company's IPO will serve as a litmus test for investor demand for new equity tied to the AI trade, and its debut has become a focal point for hopes and fears over artificial intelligence. Anthropic's prospectus flagged various "risks to humanity" including attempts to "resist shutdown" and "conceal or manipulate information," and the company emphasized the potential for its AI to "behave in unintended, misaligned, or autonomous ways." On today's episode: Shirin Ghaffary, Bloomberg News AI Reporter Rudina Seseri, Founder and Managing Partner at Glasswing Ventures Axel Merk, President and Chief Investment Officer, Merk Investments Dave Lee, Bloomberg Opinion US Technology Columnist (Source: Bloomberg)

The Close brings you the latest news and analysis leading up to the final minutes and seconds before and after the closing bell on Wall Street. Today's guests are Russell Investments President & CIO Kate El-Hillow, Harbor Capital Advisors Portfolio Manager Justin Menne, ValueEdge Advisors Chair Nell Minow, Strategic Value Partners Founder & CIO Victor Khosla, Moderna CEO Stéphane Bancel, Barclays Head of US Equity Strategy, Global Equity Linked Strategies Venu Krishna, The IPO Initiative Director Jay Ritter, Willow Wealth CEO Mitch Caplan, Duke’s Mayo Company E. Yuri Hermida, Once Upon a Farm Co-Founder & Chief Brand Officer Jennifer Garner, & Edward Jones Head of Policy, Regulatory & Government Relations Andy Blocker. (Source: Bloomberg)

Also, both parties are nervous about the Senate. Here’s the latest at the end of Monday.

The company says it could save humanity or destroy it. Investors, place your bets.

Silicon Valley’s top AI workers are being told to hide any company logos in public as working for the hottest companies marks them as targets for spies and angry protesters. OpenAI tells staff not to wear anything bearing its logo outside their offices and trains them to watch for spies offering cash, friendship or romance....

The model reportedly bypassed a paywall to retrieve aggregate vote counts but didn't access individual voter records.

Welcome to The Hill's Business & Economy newsletter {beacon} Business & Economy Business & Economy The Big Story AI giant Anthropic files to go public Anthropic publicly filed for an initial public offering (IPO) on Monday morning in what is expected to be a historic stock market debut. © Markus Schreiber, The Associated Press...

The long-awaited S-1 filing from Anthropic has finally landed and, if all goes smoothly, it could bookend a year of two of the biggest IPOs in history. It may also restart a stalled IPO market that’s been waiting on it. Markets have been choppy in the runup to the mega-IPO from the first frontier AI lab to try and tap the public for much needed funding. Ed Ludlow, Bloomberg Tech Host and Seth Fiegerman, Bloomberg News AI Editor have the details. Plus, Axel Merk, President and Chief Investment Officer, Merk Investments says that markets are fundamentally misunderstanding the Warsh Fed. And Bloomberg Balance of Power co-hosts Joe Mathieu and Kailey Leinz preview US Midterms. (Source: Bloomberg)

We broke down the A.I. company’s key figures, which give insight into the health of its business, its future plans and who’s in charge.

Todd Sohn, chief ETF strategist at Baird Strategas, joins Scarlet Fu and Athanasios Psarofagis on "Bloomberg ETF IQ." They discuss Anthropic's IPO filing and iShares MSCI Brazil ETF (ticker: EWZ), which is falling today after incumbent Luiz Inácio Lula da Silva staged a surprising comeback in Sunday’s runoff, winning Brazil's presidential election. (Source: Bloomberg)

The Federal Trade Commission (FTC) has asked executives from the AI firm Anthropic to testify as part of its consumer harm probe into the maker of the Claude AI model, a FTC spokesperson confirmed to The Hill on Monday. The demand is part of the consumer protection agency's larger investigation into Anthropic over the potential...

The owners of Elvis Presley's music catalog have sued AI giant Anthropic, accusing it of infringing copyrights on 274 of the King of Rock 'n' Roll's songs. The post Nothin’ But a Hound Dog: Lawsuit Claims Anthropic Used Hundreds of Elvis Presley Songs to Train AI appeared first on Breitbart .

The A.I. company took great pains to describe the many benefits that can come from using its Claude chatbot and technology, but also offered up a heavy dose of gloom.

Anthropic expanded its support for the nonprofit sector Monday, even as the artificial intelligence company filed to sell shares to the public for the first time. The maker of the AI assistant Claude said the moves are all part of its broader strategy for "compounding good."

Anthropic has confirmed that its Claude AI model improperly accessed voter data in Georgia by bypassing a paywall meant to restrict the information to paying users. The post Anthropic’s Claude AI Hacks into Georgia Voter Data appeared first on Breitbart .

Anthropic’s IPO filing details spending on personal security services for its top executives and the portion of its revenue attributable to customers who misuse its services.

The filing comes weeks after CEO Dario Amodei argued in an essay for slowing the development of the most advanced AI models.

Anthropic first initial public offering with the U.S. Securities Exchange Commission has been made public on disclosing the risks of AI.

The Trump administration is asking an appeals court to overturn a federal judge’s order lifting the government’s ban on Anthropic PBC’s artificial intelligence technology for federal agencies.

Hey HN, I'm Christian and this is a new OSS project I've been noodling on for the last couple of weeks. Hob is a free, MIT-licensed desktop app for Mac, Windows and Linux that works on your files the way the Claude and ChatGPT desktop apps do, except the model runs on your own computer. It uses open-weight models (currently Gemma 4 and gpt-oss). All your chats and files never leave your system. The only things that go online are web search (can be turned off in settings) and any MCP connectors you add yourself. Main features: - Automatically finds the best open-weight model for your hardware on first install (Mac/Linux/Windows) - Writes sandboxed Python code for ad hoc tools (data processing, document conversion, graphs etc.) - Built-in web search (powered by SearXNG) - MCP support - Image recognition - CRUD to your local filesystem - Works fully offline with no internet Before building this I tried every local AI app I could find, but it seemed each one had different problems. The main thing I noticed is that none of them could do what the ChatGPT and Claude apps do with your files, e.g. read and change files in a folder, look at images, run code in a sandbox, draw a chart. The most advanced thing any of them did was paste a plain text file into a small context window, which basically failed as soon as you had a file larger than a few KB. I couldn't find anything that could easily parse images and bigger PDFs. The OSS solutions that were more advanced all seemed to default to hosted models (either OSS or closed), which defeats the entire point of the privacy that on-device OSS offers. They also seemed to push towards an account/open-core style business. I build it using Electron bridged with llama.cpp in its own process. On initial load, the system will try and figure out if you have a compatible graphics card and if not default to CPU-only (I ran into a lot of issues of graphics card beeing seemingly compatible but then crashing on first query, there's a whole saga to this lol). Code execution happens in a sandbox that can only write to the chat's folder and has no network unless asked (sandbox-exec on macOS, Landlock on Linux and a low-integrity process on Windows). Images go to a small Qwen3-VL model. Dictation is a local Whisper model. The hardest part was definitely to get the smaller models to use tools reliably. Ran into a ton of issues of 4B models happily saying "here's your chart" without ever calling the chart tool or get stuck in an infinite call loop. I spent a fair amount of the work building guards vs tweaking prompts (that didn't really work for small models). I understand the pushback that local models are weaker than frontier cloud models. This is true, but for most tasks for most normal people, I'd argue you rarely need the newest frontier model. From my own evals, systems with 16+ GB of RAM do really well, and if you have a GPU or more RAM, you get very close to what you'll get out of the box with a frontier hosted model. Excited to hear what you think and see how it runs on other systems! Comments URL: https://news.ycombinator.com/item?id=50246238 Points: 34 # Comments: 14