In partnership with
Good morning, {{first_name | AI enthusiast}}.
Nvidia just made its boldest move yet outside chips: a reported $12.9 billion bid for Hugging Face, the open-source hub where much of the AI world stores and shares its models. If the deal closes, the company that already controls AI’s hardware layer would also own the layer where open models actually live.
Hugging Face built its reputation on staying neutral in a market full of competing interests. Can it hold onto that trust once its landlord is the company with the biggest financial stake in who wins the AI race?
Today in AI Brief:
Nvidia’s reported $12.9B bid for Hugging Face
OpenAI’s Jalapeño chip outguns Nvidia on watts
Alibaba’s Wan 3.0 turns documents into video
Stop Paying for 10 Tools. One AI Does It All.
Most e-commerce sellers are running their store across 6 to 10 separate tools — and spending more time managing software than growing their business. StoreClaw replaces your entire stack with one autonomous AI engine that monitors competitors, optimizes listings, automates marketing, and tracks real profit across Shopify, Amazon, and beyond.
It doesn't wait for you to ask. It runs 24/7 in the background, so you wake up to a full dashboard instead of a list of things you forgot to check.
Connect your store, and StoreClaw gets to work — no prompts, no complex setup, no six-app stack.
Free to start. No credit card required.
Nvidia’s Reported $12.9 Billion Bid for Hugging Face
In Brief: Nvidia has agreed to buy Hugging Face, the open-source hub where developers host and share AI models, for $12.9 billion, according to a report from The Information. Neither company has confirmed the deal, and it could still fall apart before signing.
The Details:
Nvidia’s offer marks a sharp jump from the roughly $7 billion valuation it proposed for a minority stake just months ago — Hugging Face was worth $4.5 billion as recently as 2023.
The deal would hand Nvidia a foothold in the distribution layer of open-source AI, on top of the $40+ billion it has already poured into AI equity deals this year, including its $30 billion OpenAI stake.
It follows reports that Nvidia has been circling Hugging Face for months, with talks accelerating in the past week.
Take Away:
Hugging Face has stayed useful precisely because it’s seen as neutral ground in a market full of competing interests. If Nvidia — the company profiting most from the AI infrastructure race — ends up owning where the models live, that neutrality becomes the first casualty.
OpenAI’s Jalapeño Chip Takes Aim at Nvidia
In Brief: OpenAI published its first benchmark results for Jalapeño, the custom AI inference chip it co-developed with Broadcom, claiming major efficiency gains over Nvidia’s hardware. The chip is built to answer queries faster while burning less power at scale.
The Details:
Running at 700 watts versus a 1,200-watt Nvidia comparison chip, Jalapeño delivered up to 1.9x more AI work per watt and as much as 3.6x lower latency on models like DeepSeek R1 and Kimi K2.5.
The chip took nine months to go from design to manufacturing-ready, and OpenAI says two more generations are already in development.
OpenAI’s VP of Hardware said the company will keep buying Nvidia chips for training — Jalapeño is built for inference only, deploying later this year with production ramping through 2027.
Take Away:
The benchmarks came from OpenAI itself, not an independent lab, and critics note Jalapeño uses newer memory than the Nvidia chip it was tested against. Still, a foundation model company shipping its own silicon is exactly the kind of vertical integration that should worry Nvidia’s margins.
SPONSORED BY LEVANTA
Build a Holiday Creator Affiliate Program in 90 Days
Creators lock in holiday content calendars 90 days out, before brands figure out commissions. Waiting too long to launch an affiliate program means less runway to build demand and a missed shot at the best partnerships.
The 90-Day Holiday Sprint covers commissions, recruiting, and scaling a program at Day 30, 60, and 90.
Alibaba’s Wan 3.0 Turns Documents Into Video
In Brief: Alibaba launched Wan 3.0 in public beta, an AI video generator that doubles clip length to 30 seconds and can build videos directly from documents, spreadsheets, and slides — not just text and images. The release lands the same week Alibaba raised $10.2 billion in new shares to fund its AI push.
The Details:
Wan 3.0’s Omni-Reference feature accepts text, image, audio, video, PDFs, and webpages as source material, then generates synchronized facial expressions and multilingual voice output.
Clips now render at up to 1080p with better instruction-following and shot-to-shot consistency than the 15-second-max Wan 2.7.
Alibaba’s stock dropped after the $10.2 billion share placement, even as the company directed the proceeds entirely toward AI investment.
Take Away:
Turning a slide deck or spreadsheet into a video clip is the kind of practical, try-it-yourself feature that drives real usage, not just headlines. It’s also a reminder that China’s AI labs are shipping consumer-facing tools as fast as anyone in Silicon Valley.
Everything else in AI
Apple unveiled a refreshed Mac mini with its first 2nm chip, the M6, running on-device AI workloads up to 4x faster starting at $899.
Anthropic is reportedly pitching investors a $30 trillion total addressable market ahead of an IPO targeting a $2 trillion valuation.
Two researchers walked away from Jeff Bezos’s Project Prometheus to launch Accelerated Understanding, a startup building physics-first AI models instead of chatbots.
Perplexity launched Portable Computer, a local AI agent that runs entirely on Nvidia RTX hardware with zero per-token costs.
OpenAI rolled out website login support for ChatGPT Work, letting the agent access password-protected pages without ever seeing your credentials.