AI news, 19 November: Anthropic opens trading after record $81bn IPO
Anthropic prices record IPO and begins Nasdaq trading. Anthropic priced its initial public offering at $83 per share on 18 November, above the marketed $70–$78 range, and sold about 980 million shares to raise $81.3 billion—the largest IPO on record, surpassing SpaceX. The deal values the Claude maker at roughly $1.5 trillion on a fully diluted basis. Shares begin trading today on Nasdaq under the ticker ANT, with CEO Dario Amodei and co-founders set to ring the opening bell. According to CNBC, TechCrunch and Bloomberg, demand was heavily oversubscribed; a separate private placement with firms including Google, Microsoft, Nvidia and AMD adds further capital. The listing marks the first major public debut of a frontier AI lab and is seen as a key test of investor confidence in the sector.
Nvidia doubles revenue in bumper quarter. Nvidia reported third-quarter fiscal 2027 revenue of $115.8 billion, up 103% from a year earlier and well above estimates, driven by surging demand for AI data-centre chips. Data-centre sales reached $107.6 billion. Net profit more than doubled to about $65 billion. The company guided fourth-quarter revenue to around $130 billion and noted early shipments of its Vera Rubin platform. Interim CEO Colette Kress highlighted strong momentum; the board has confirmed it will not reinstate former CEO Jensen Huang. Results were widely covered by Reuters and other outlets as evidence that AI infrastructure spending remains robust despite recent industry turbulence.
OpenAI models deliver chip-design gains with AMD. OpenAI published results showing its models improved power, performance and area (PPA) by up to 14% on AMD’s flagship GPU intellectual property against already heavily optimised baselines—described as roughly a year of progress in under five months. Some model-suggested circuit changes have already been included in a design AMD sent for manufacturing. Similar techniques also reduced area in parts of OpenAI’s own Jalapeño inference chip, with gains verified on real silicon. OpenAI is working with Synopsys to extend the approach. The announcement underscores how AI is beginning to accelerate the design of the very chips that power it.
Google rolls out Gemini 4 Neon and Genie 4. Google introduced Gemini 4 Neon, a fast frontier-tier model aimed at coding, knowledge work and science tasks, ranking highly on independent indexes while offering lower cost and a 1-million-token context window. It is rolling out via the Gemini API, app and other products for paid users. Separately, Google DeepMind began making its Genie 4 world model available to eligible Google AI Ultra subscribers inside Project Genie, enabling real-time editing of interactive AI-generated worlds with native audio and a new creation agent. Both moves expand access to Google’s latest generative capabilities.
Other notable developments. Microsoft continued its Ignite announcements with new in-house models and Copilot agent features for enterprise customers. Earlier stories, including OpenAI’s GPT-6.2 release and California’s lawsuit over rogue agents, remain in the background without major fresh developments reported in the past day.