Good morning. Here is what mattered in AI today. Anthropic priced its IPO at $83 a share and raised about $81.3 billion, the largest offering on record, ahead of its Nasdaq debut. Google released Gemini 4 Neon, a mid-priced model that slots between Hydrogen and Carbon. Palo Alto Networks agreed to buy Applied Compute for $4.5 billion in stock. For builders, we cover a Transformers technique that runs a 180-billion-parameter model on a 32GB laptop and Hugging Face’s move to route inference through OpenRouter. We also profile Colette Kress, the CFO now running Nvidia as interim CEO.
Daily Digest: Anthropic's record IPO, Gemini 4 Neon, a $4.5B deal
The AI stories that mattered on November 19, 2026: Anthropic's $81.3B IPO, Google's Gemini 4 Neon and Palo Alto Networks buying Applied Compute.
Today in 6 stories · November 19, 2026
- 01Markets & Economy
Anthropic prices IPO at $83 a share, raising a record $81.3B
Anthropic sold 980 million shares above its $70-$78 range, the largest IPO on record. Shares start trading on the Nasdaq under the ticker ANT on November 19.
- 02Model Releases
Gemini 4 Neon launches at $3/$15 between Hydrogen and Carbon
Google's new mid-tier model has a 1M-token context window. Google says it ranks third on the Artificial Analysis Intelligence Index; we compare its price with the rest of the Gemini 4 family and GPT-6.2 Sol.
- 03Startups & Funding
Palo Alto Networks to buy Applied Compute in $4.5B stock deal
The security company is buying a startup that trains and serves open-weight models, less than three weeks after Applied Compute was valued at $3.25 billion.
- 04Local & On-device AI
Hugging Face streams MoE experts to run 180B model on a 32GB Mac
A Transformers pull request loads mixture-of-experts weights from SSD on demand. We explain the trade-offs and how it compares with llama.cpp and Ollama.
- 05People / Who's Who
Colette Kress: who is Nvidia's interim CEO after Jensen Huang
Nvidia's CFO since 2013 now also runs the company. Her career, her dual role and what she has said publicly since taking over.
- 06APIs & SDKs
Hugging Face moves Inference Providers routing to OpenRouter
Chat-model calls on Hugging Face now route through OpenRouter. Endpoints and prices stay the same, but bring-your-own provider keys stop working for language models.
Sources
This story was reported and written in our own words from the sources below. Primary sources are preferred.
- Anthropic prices initial public offering — anthropic.com
- Introducing Gemini 4 Neon — blog.google
- 180B Parameters, 32GB of RAM: Pushing MoE Inference to the Limit with Transformers — huggingface.co