Susana SchwartzAnthropic closes the gap on OpenAI as anticipated public debuts near
Anthropic has for the first time surpassed OpenAI in quarterly revenue, bringing in $11.6 billion in Q2 2026 compared to OpenAI’s $6.7 billion. The company is on track to generate annualized revenue of more than $65 billion – 7x greater than what it was generating end of last year.
With both Anthropic and OpenAI eyeing public offerings, it’s significant that Anthropic has rapidly gone from “underdog” status to Wall Street darling, touting a lean-efficiency position that increasingly contrasts OpenAI’s operating losses.
With B2B adoption and high-margin, sticky developer APIs and workplace tools, Anthropic has captured 34.4% of businesses, as compared to OpenAI’s 32.3%. Also, OpenAI’s consumer retail focus and massive non-paying base contributed to its $12.3 billion in operating losses this quarter.
As a result of this change in position, some institutional backers and investors are predicting an “historic” public market debut of $2 trillion for Anthropic. To make that come to fruition, Anthropic is working with Morgan Stanley, Goldman Sachs, JPMorgan and other core banks, asking for top-tier commitments of $1.25 billion and lower-tier allocations of $1 billion.
Whether Anthropic will get enough to truly unseat SpaceX’s IPO as the biggest ever remains to be seen, but it achieved a historic milestone by turning a small adjusted operating profit – something frontier AI companies are not known for.
As it files listing documents for its IPO, Anthropic is also riding high on the $15 billion in financing it received for a massive 2,900-acre AI Nexus data center campus in Hubbard, TX. For that project, the company received financial backing from Google, which is guaranteeing lease and power obligations, in addition toa $1.3 billion private credit loan led by Eagle Point Credit Management (as part of a broader $16 billion financing package led by banks like Morgan Stanley).
Susana Schwartz
Technology Editor
RCRTech
AI Infrastructure Top Stories
Google-AMD hybrid chip?: Google’s TPU v10 could bring AMD x86 cores onboard, which would make Google’s 10th-gen TPU a hybrid AI ASIC, putting general-purpose CPU cores directly on the accelerator package.
Nvidia Feynman GPU: Nvidia is reportedly designing its next-gen “Feynman” GPUs to debut on TSMC’s A16 node – a 1.6nm-class process – rather than the standard 2nm (N2) generation. That would make Nvidia the first major customer for A16.
AI Today: What You Need to Know
SpaceX Phase 1 of 100M-sq.ft. Terafab: SpaceX will invest $16.8 billion for phase 1 of the Grimes, TX, Terafab for edge computing and inference chips to be used in Tesla robots, self-driving Cybercabs, and space-based data centers.
No more cold plates for heat-transfer?: Molex is investing in CAEPlus, whose BoundaryCool is a departure from conventional passive cold plates, instead using an active cooling architecture to enhance heat removal at the chip level.
PA Governor restricts DC development: In a major shift, Gov. Josh Shapiro is asking DC developers seeking state funding to voluntarily comply to the Governor’s Responsible Infrastructure Development (GRID) standards.
Former printing press for DC: The North Loop Neighborhood Association is holding community workshops and online meetings to discuss Legacy Investing’s plan to turn the former Minnesota Star Tribune printing plant into a data center.
AI bolsters factory supply chain: Generac will spend $250 million to help several factories beef up its backyard generators for data center use. The company already has an order backlog of $1.6 billion, for which it has added 1,000 workers.
Saragon to build 10 DCs: Saragon launched as new U.S. Data Center company purpose-built for AI Inference and Edge Colocation, and backed by up to $1 billion from I Squared Capital. Saragon will launch with 10 DCs, with 53 MW of power.
