A KV cache, or key-value cache, speeds up autoregressive transformer inference by storing attention information calculated for tokens that have already been processed. Without the cache, the model would recompute those keys and values each time it generates another token. Reusing them reduces repeated work and improves response latency, especially for long prompts and outputs. The tradeoff is memory: cache size grows with sequence length, batch size, layer count, and model dimensions. Production systems manage this cost with paging, quantization, eviction, prompt caching, and request scheduling. KV caching changes inference efficiency, not the model’s learned knowledge, and cached data must be protected when prompts contain sensitive information.
Open-source AI agent OpenClaw released its largest update yet, simplifying installation and adding shared cloud sessions that let multiple people collaborate on one agent.
OpenAI says ChatGPT Ads reached a $1 billion annualized revenue run rate in under 200 days, expanding self-serve ad buying to India, Europe and the Middle East.
Anthropic agreed to a $35 billion, six-year computing deal with Nvidia-backed Lambda, its latest megadeal in a compute-buying spree that has topped $150 billion this year.
A federal judge ruled the Pentagon’s designation of Anthropic as a supply chain risk was illegal retaliation for the company’s public stance on military AI use.
Nvidia has reportedly agreed to acquire open-source AI platform Hugging Face for $12.9 billion, though no signed deal is confirmed and talks could still fall apart.
Anthropic opened a research preview of the Model Hardware Standard, letting AI agents operate lab and factory equipment like microscopes and robotic arms in its first physical-world product.
Nvidia is building hardware optimizations for Chinese AI models like DeepSeek and Qwen, even as it warns regulators that US restrictions on such models could hurt its business.
Google launched Gemini 3.5 Transcribe, a speech-to-text model supporting 85+ languages that cleans up filler words and self-corrections, ranking fifth on an independent accuracy leaderboard.
Sam Altman told TIME that OpenAI expects an internal system it would call AGI by year’s end, built on its unreleased Astra model, though the claim rests on a contested definition.
Bill Gates published a 6,000-word essay warning that the AI transition could be one of history’s most disruptive periods, urging new institutions and a plan before it’s too late.