Anthropic Weighs October IPO — Early Talks with Wall Street Banks
Anthropic is considering going public as soon as October, according to Bloomberg, with early discussions underway with Goldman Sachs, JPMorgan, and Morgan Stanley for leading roles on the listing. The move would pit Anthropic directly against OpenAI, which is also pursuing an IPO this year. The timing is no coincidence: Anthropic’s revenue is reportedly approaching $20B ARR, paid subscriptions have more than doubled, and the Claude Mythos reveal — intentional or not — gives them a next-gen product narrative heading into the roadshow.
The IPO would be structured as a public benefit corporation listing, preserving Anthropic’s dual-mission structure. Analysts note that the Pentagon court victory and consumer growth surge both strengthen the IPO story, while the Mythos cybersecurity risks add a wildcard that investors will want to price in.
Pentagon Appeal Window Open — Seven Days to Challenge Injunction
Judge Rita Lin’s preliminary injunction blocking the Pentagon’s supply-chain risk designation and Trump’s federal agency ban remains in effect, but she delayed implementation by one week to give the government time to appeal. The clock started Thursday, March 26, meaning the government has until roughly April 2 to file. If no appeal is lodged, the injunction stands and Anthropic’s federal contracts remain intact.
Legal experts expect the DOJ to appeal to the Ninth Circuit, setting up a potentially landmark case on government retaliation against AI companies. Defense One reported that the ruling has already reshaped how the Pentagon approaches AI procurement conversations, with other vendors now citing Anthropic’s ethical red lines as a precedent.
Peak-Hour Limits: Developer Pushback Grows as Workarounds Emerge
Two days into the new peak-hour session limit regime, developer frustration is building. Multiple community forums and Reddit threads report that the 5-hour session windows are being consumed significantly faster during business hours (5am–11am PT / 1pm–7pm GMT), with some users saying they hit limits in under three hours. Anthropic says the overall weekly allocation is unchanged and that expanded off-peak capacity compensates — but for developers working standard business hours, that math doesn’t add up.
QWE AI Academy published a practical guide to working within the new limits, including strategies like shifting intensive Claude Code sessions to early mornings or evenings. PYMNTS called the move a signal that AI rationing is becoming a consumer reality across the industry, not just at Anthropic. The timing is awkward: throttling your fastest-growing product right before an IPO roadshow isn’t ideal optics.
The New Stack: Anthropic’s “Madcap March” by the Numbers
The New Stack published a comprehensive recap of what they’re calling Anthropic’s most consequential month ever. The tally: 14+ product launches (Claude Code auto mode, computer use, Excel/PowerPoint add-ins, voice mode, web search GA, structured outputs GA, Claude Marketplace), five service outages, a landmark court victory, Bloomberg-reported IPO plans, and a data leak revealing the company’s most powerful model. All in 29 days — and still one more to go.
The piece highlights how March 2026 compressed what would normally be a year’s worth of news into a single month, largely because the Pentagon feud amplified every product announcement into mainstream media coverage. Anthropic went from an AI company known mainly to developers to a household name practically overnight.
Claude Paid Subscriptions More Than Double — 1M Users Signing Up Every Day
TechCrunch reports that Claude paid subscriptions have more than doubled in 2026, with daily signups quadrupling from roughly 250,000 per day in January to over 1 million by early March. An independent analysis of anonymized credit card data from 28 million US consumers confirmed the surge, showing Claude gaining paid subscribers at record rates. The majority of new subscribers are on the Pro tier at $20/month.
Two factors are driving the growth: the Super Bowl ads that directly challenged OpenAI, and the explosive adoption of Claude Code and Cowork. Claude Code alone has crossed a $2.5B run rate. The Pentagon feud acted as an unexpected marketing accelerator — Claude shot to #1 on both the App Store and Play Store in early March and has stayed near the top since. Daily active users have more than tripled since January.
Mythos Cyber Risks Stir Ongoing Debate Among Security Researchers
The fallout from the Claude Mythos data leak continues to ripple through the security community. The model’s reported ability to rapidly identify vulnerabilities and assess attack surfaces has split researchers: some see it as a transformative defensive tool, while others warn the same capabilities could enable offensive exploitation at speeds defenders can’t match. Coindesk reported that software stocks briefly dipped on the news before recovering.
Neurotekai published a detailed technical breakdown of what the leaked materials reveal about Mythos’s architecture, noting the Capybara tier’s rumored 10-trillion parameter scale and its implications for cybersecurity applications. Anthropic continues to say no public release is planned in the near future, but with IPO talks underway, the commercial pressure to ship is real.
The IPO Math: Growth vs. Governance
Anthropic’s IPO calculus is unusually complex. On one hand, the numbers are spectacular: revenue approaching $20B ARR, paid subs doubling, daily signups at 1M+, Claude Code on a $2.5B run rate, and a next-gen model that reportedly outperforms everything on the market. On the other hand, they’re throttling users during peak hours because they can’t buy GPU capacity fast enough, they’re in active litigation with the US government, and their most powerful model was revealed through a security misconfiguration.
The public benefit corporation structure adds another wrinkle. Investors buying into an Anthropic IPO are buying into a company that will, by charter, sometimes choose safety over speed. That’s the entire Pentagon story in one sentence. The question is whether Wall Street will price that as a feature or a risk. Dario Amodei’s bet is that it’s a feature — that the market will reward a company that can grow this fast while maintaining ethical boundaries. October will tell us if he’s right.