Anthropic's $2T IPO looms as Fable 5 adoption lags; OpenAI launches Ultrafast GPT-5.6 Sol

Anthropic is preparing for a blockbuster IPO that could raise over $100 billion at a $2 trillion valuation, potentially surpassing SpaceX's record, but its flagship Fable 5 model is struggling with corporate adoption, capturing only 6-11% of tokens as customers opt for cheaper alternatives like Opus 5. Meanwhile, OpenAI launched an 'Ultrafast' mode for GPT-5.6 Sol offering up to 14x faster speeds following a 20%+ price cut, and Google released Gemini 3.7 Flash with a 1M token context window at competitive prices. Nvidia is entering the AI model race with a $6 billion deal to license Poolside's technology and hire its engineers to build open-weight Nemotron models.

Written by AI from the sources it cites; it can be wrong — follow any citation to confirm.

13 sources2 min read

Anthropic's IPO could top $100B raise at $2T valuation, beating SpaceX Anthropic's bankers have told investors it could raise more than $100 billion in its IPO, potentially valuing the company at $2 trillion and surpassing SpaceX's record $85.7 billion raise. The company filed to go public in June and could list shares this autumn, potentially beating OpenAI to market. 1 2 3

Anthropic's Fable 5 lags as customers switch to cheaper Opus 5 Anthropic's flagship Fable 5 model has captured just 6% of purchased tokens and 11.4% of revenue in its first two months, as customers flock to the half-price Opus 5 which matches its performance on many benchmarks. Analysts say most enterprises don't need frontier models, raising questions about Anthropic's aggressive spending ahead of its IPO. 4 5 6

OpenAI launches Ultrafast mode for GPT-5.6 Sol at 14x speed OpenAI introduced Ultrafast mode offering GPT-5.6 Sol at up to 14x faster speeds, following a price cut of over 20% that brought pricing to $4 per million input tokens and $20 per million output tokens. The move comes amid intensifying competition from Anthropic and emerging Chinese AI models. 7 8

Google releases Gemini 3.7 Flash with 1M token context window Google launched Gemini 3.7 Flash, offering a 1 million token context window at $0.38 per million input tokens and $1.88 per million output tokens. The competitive pricing could influence enterprise adoption and pressure other LLM providers on cost. 9

Nvidia enters AI model race with $6B Poolside deal for open-weight models Nvidia will pay $6 billion to license Poolside's technology and hire most of its engineers to work on Nemotron, its family of open-weight AI models, with plans to invest another $1 billion at a $12 billion pre-money valuation. The move aims to challenge OpenAI and DeepSeek in the open-weight model space. 10 11

OpenAI's active agent users hit 20M as enterprise revenue surges 50% OpenAI disclosed at an all-hands meeting that its AI coding and agent features reached 20 million weekly active users as of August 19, representing about 2% penetration of its 1 billion active ChatGPT users. Overall revenue run rate climbed 35% quarter-to-date while enterprise revenue grew more than 50%. 12

Anthropic expands Mythos 5 access to defenders, unveils $35M open source fund Anthropic is broadening access to cybersecurity capabilities of its advanced AI models through partner integrations, an updated Claude Security offering, and a new $35 million open source funding program. The move builds on Project Glasswing and expands the Cyber Verification Program while keeping guardrails around direct model interaction. 13