infrastructure — AI Digest

7 мая Anthropic Signs SpaceX Colossus Compute Deal, Doubles Claude Code Rate Limits Anthropic tools
15 мая MinT: Managed Infrastructure for Training and Serving Millions of LLMs Mind Lab research
13 июл vLLM v0.25.0: Model Runner V2 Default, PagedAttention Retired, Transformers Backend Parity tools
23 июл AMD and Anthropic announce 2-gigawatt Instinct MI450 compute deal with up to $5B AMD investment Anthropic industry
17 авг LLMRouter: Unified Infrastructure for Developing, Evaluating, and Deploying LLM Routers University of Illinois at Urbana-Champaign research
24 июн Prime Intellect Releases prime-rl v0.6.0 for Agentic RL on Trillion-Parameter MoE Models Prime Intellect research
25 сен Google's Project Suncatcher to fly TPUs on SpaceX rideshare mission Google DeepMind research
19 мая LongLive-2.0: NVFP4 Parallel Infrastructure for Long Video Generation (NVIDIA, 1,220 HF upvotes) NVIDIA research
23 июл SLAI T-Rex: Full-Parameter Post-training of the DeepSeek-V4 Family on Ascend SuperPOD SLAI research
10 сен Miles v0.1: a production-level RL post-training system RadixArk research
18 июн Yandex Open-Sources YaFF Data Format, Saving Up to 20% Server Capacity Yandex tools
3 июл Meta Announces "Meta Compute" Cloud Business to Monetize Surplus AI Infrastructure Meta AI industry
31 июл Meta raises 2026 AI capex guidance to $130-145B as free cash flow collapses Meta industry
15 авг LLMRouter: Unified Infrastructure for Developing, Evaluating, and Deploying LLM Routers University of Illinois Urbana-Champaign research
9 мая Anthropic Signs $1.8 Billion Seven-Year Cloud Computing Deal with Akamai Anthropic industry
1 сен Anthropic signs $35 billion cloud deal with Nvidia-backed Lambda Anthropic industry
6 сен DeepSeek plans a 160,000-chip Huawei Ascend cluster in Inner Mongolia DeepSeek industry
6 сен ByteDance secures record $29.6 billion loan from ~30 banks for AI buildout ByteDance industry
13 сен Z.AI launches $5 billion Hong Kong raise: $2B share placement plus $3B convertible bonds Zhipu AI (Z.AI) industry
12 мая vLLM v0.21.0rc1: Python 3.14, CUDA 13.0, and Transformers v5 Compatibility tools
3 июл ELDR: Expert-Locality-Aware Routing Cuts MoE Serving Latency by up to 14% Microsoft Research research
18 июл LongStraw: Long-Context RL Beyond 2M Tokens under a Fixed GPU Budget Mind Lab research
19 авг Dharma-AI lifts GPU-cluster utilization 53.6% → 87.0% by encoding physical constraints into the scheduler Dharma-AI tools
12 сен OpenAI details Habitat, the storage platform serving 70M requests per second OpenAI industry
24 сен Google advances Private AI Compute with secure, server-side memory Google DeepMind research
19 июл Meta and Anthropic reportedly in early talks over $10 billion compute lease deal Anthropic industry
12 июн VK Tech Reduces VK Data Platform Infrastructure Requirements 2.5× for AI Deployments VK AI tools