From November 9 to 10, Anthropic kicked off its IPO roadshow, Google and OpenAI each made moves on the same day, and a security incident deserves attention. All 6 items below happened within the past 24 hours.
Anthropic kicks off IPO roadshow
Anthropic launched its IPO roadshow on November 10, offering 825 million Class A shares at $70–$78 and seeking a Nasdaq listing under the ticker ANT. The night before, a document circulated to banks was withdrawn over incorrect pricing; Reuters reported it showed a $70–$80 range and up to $69.3 billion in proceeds. Details: How the IPO pricing was pulled.
Google unveils TPU 8tx and 8i Max
Google introduced TPU 8tx and 8i Max on November 9, bringing its eighth-generation TPU lineup to four systems in a year. The 8tx targets confidential inference, keeping data and model weights protected while in use, with Google claiming latency half that of the 8t; the 8i Max packs two 8i chips in one package, doubling compute, on-chip memory and memory bandwidth. Neither system has a published price or launch date. Details: Google's new TPUs.
OpenAI launches Control Room
OpenAI introduced Control Room on November 9, a central authorization layer covering all research inference workloads: a job registers and gets a unique ID, and after a stop instruction it is refused further model output within 35 seconds at most. All frontier reinforcement-learning jobs are already covered, with the rest to follow within days; the system follows agent overreach incidents in July and September. Details: Control Room debuts.
ChatGPT Enterprise joins AWS
OpenAI announced on November 9 that ChatGPT Enterprise is available on AWS Marketplace, letting companies buy through existing AWS purchasing and billing, with eligible spending counting toward AWS commitments. The bundle includes ChatGPT, ChatGPT Work, Codex and dots, with Healthcare and Financial Services editions available through the same channel; features match the direct version, with no public price.
Fireworks AI widens breach disclosure
Fireworks AI updated its security bulletin on November 9: the October incident was worse than first disclosed, with attackers entering development machines via a third-party cloud provider and reaching some customers' billing information, secrets, datasets and unfinished model weights. Inference, the Training SDK and Nexus were unaffected; the company has notified law enforcement and brought in Palo Alto Networks Unit 42 for forensics.
NVIDIA fallout: short seller flagged it six months ago
TechCrunch reported on November 9 that short seller Culper Research had already questioned Jensen Huang's relationship with Alice Huang, a Megaspeed executive in Singapore, in a May 13 report, calling it "uncomfortably close." NVIDIA's board has opened a formal investigation, Huang has admitted an extramarital affair, and U.S. intelligence officials believe Alice Huang was recruited by the Chinese government.
