NVIDIA is buying Hugging Face, the platform where millions of developers share AI models, for nearly $13 billion. The deal promises to keep Hugging Face open and independent while scaling its infrastructure.

• Hugging Face has 18 million users sharing 3 million models and serves 200,000 companies; NVIDIA says it won't require its own compute to build or deploy on the platform.
• NVIDIA is already the largest contributor of open models to Hugging Face (500+ models, 250+ datasets) and frames the acquisition as strengthening open-source AI rather than locking it down.
• The deal hinges on a commitment to multi-cloud, multi-accelerator support—meaning developers won't be forced to use NVIDIA hardware even after the acquisition.

Nvidia's dominance is shifting from raw GPU competition to controlling the entire data center infrastructure around compute. As AI systems scale to gigawatt levels, the company's specialized hardware for data orchestration—CPUs, networking, storage—is becoming harder to replicate than the chips themselves.

• Nvidia's Vera CPU and supporting hardware deliver 3x performance improvements by optimizing data movement to GPUs, addressing a critical bottleneck as companies optimize for tokens-per-watt efficiency.
• Competitors like OpenAI are tackling the same data movement problem differently (integrated chips like Jalapeño), but the underlying challenge shows the competition has shifted from GPU design to full-system efficiency.
• Operating megascale data centers at peak efficiency is still incredibly difficult, creating a new competitive layer where system integration matters more than individual chip superiority.

This snippet shows how to call NVIDIA’s integrate API to run the moonshotai/kimi-k2.6 chat model. It covers setting headers, payload fields (model, tokens, temperature) and handling both JSON and stream responses.

• Sample code calls NVIDIA's integrate API (https://integrate.api.nvidia.com/v1/chat/completions) to run moonshotai/kimi-k2.6, requiring only an NVIDIA_API_KEY swapped into the Authorization header.
• Default payload uses temperature 1, top_p 1, a 16,384 token cap, and a fixed seed of 0 for reproducible outputs.
• Same code handles both streaming (line-by-line printing) and non-streaming (full JSON dump) responses by toggling the stream flag and Accept header.
• Snippet is portable and meant to be adapted to any endpoint provider, not just NVIDIA's hosted API.

Claudia Ng outlines how to run the open-source Hermes agent for free by pointing it at NVIDIA’s free API tier. She walks through installing the Hermes desktop app, configuring the NVIDIA API key and model (switching from GLM-5.1 to DeepSeek V4 Flash), and enabling web search tools, while noting latency and rate-limit issues on the free tier.

• Local Mac models couldn't hit the 64K-token context window Hermes needs for tool loops, forcing a switch to NVIDIA's free API tier.
• Setup was just a few terminal commands: install the desktop app, run `hermes model` to point at NVIDIA (49 models available), and `hermes tools enable web` for search.
• GLM-5.1 burned through the free tier fast due to timeouts and heavy reasoning overhead, so switching to DeepSeek V4 Flash gave faster responses with shorter reasoning traces.
• Every tool call adds a round trip and grows the prompt, so free-tier sessions can stall once rate limits hit.

The article argues that enterprises should measure AI infrastructure economics by cost per token rather than raw compute metrics like FLOPS per dollar. It shows how maximizing delivered tokens—through hardware, software and system optimizations—drives down real-world cost and boosts revenue, citing NVIDIA Blackwell’s 35× lower token cost versus Hopper.

• Cost per token (total infra cost ÷ tokens generated), not FLOPS/dollar or GPU hourly rate, is the real measure of AI infrastructure efficiency.
• Blackwell GB300 NVL72 costs almost 2x more per GPU-hour than Hopper H200 ($2.65 vs $1.41), but delivers 65x the tokens/sec per GPU (6,000 vs 90).
• That throughput gap translates to 50x more tokens per megawatt and a 35x lower cost per million tokens ($0.12 vs $4.20).
• Techniques like FP4 precision, speculative decoding, KV-cache offloading, and disaggregated serving are necessary, not optional, to actually achieve these lower token costs.

Nvidia Corp. will invest $1 billion over five years to establish a new artificial intelligence laboratory in collaboration with Eli Lilly & Co. The facility, located in Silicon Valley, aims to accelerate the application of AI in the pharmaceutical industry by leveraging Lilly's lab expertise.

• Nvidia will invest $1 billion over five years to build an AI drug discovery lab with Eli Lilly, located in Silicon Valley
• The lab will use Lilly's lab expertise combined with Nvidia's computational tools to predict how compounds interact with biological systems
• The goal is to speed up and cut costs in drug discovery while improving precision in targeting specific diseases
• The deal signals Nvidia's strategic push into healthcare as a growth area beyond its core computing and gaming business

The article explores the parallels between the film "Lord of War" and the current AI compute market, focusing on Jensen Huang of Nvidia as the central figure in the AI arms race. It details an experiment where multiple AI image generation models were tested to recreate a parody poster, "Lord of Tokens," using advanced prompts that challenge the models' capabilities. The results highlight varying levels of success in achieving the desired artistic and technical details.

• The "Lord of Tokens" poster parody frames Jensen Huang as the central arms dealer of the AI compute boom, mirroring Nicolas Cage's character in "Lord of War"
• Multiple leading AI image generation models were torture-tested with a complex, design-agency-grade prompt to recreate this parody poster
• The models showed significantly varying success rates in nailing the specific artistic composition and technical details demanded by the prompt

Nvidia and Uber are collaborating to launch a fleet of 100,000 fully autonomous robotaxis by 2027, utilizing Nvidia's Drive AGX Hyperion 10 technology for level-4 automation. While Uber will not manufacture the vehicles, it will partner with automotive companies like Stellantis and Mercedes-Benz to build the fleet, aiming to establish a significant presence in the autonomous ride-hailing market.

• Nvidia and Uber plan to deploy 100,000 fully autonomous robotaxis by 2027 using Nvidia's Drive AGX Hyperion 10 platform for level-4 automation
• Uber won't manufacture vehicles itself, instead partnering with automakers like Stellantis and Mercedes-Benz to build the fleet