Fetching from the wire…
Infra2026-07-10 · source-backed
TechCrunch argues that by proving compute's value, Nvidia created a market where neoclouds and power providers capture rents on the periphery while Nvidia carries the R&D burden. Commoditizing the sale of GPU-hours decouples chip margin from compute margin. The practical read for builders: inference price competition has more room to run than chip pricing implies. Plan your unit economics on prices falling further.
Each link below shares sources, entities, or timing with this story.
NVIDIA's Blackwell successor is in production ahead of schedule. The NVL72 rack (72 GPUs) delivers 3.6 exaFLOPS for inference, with 288GB HBM4 per GPU. NVIDIA claims 10x lower cost-per-token versus Blackwell. The Rubin CPX variant — purpose-built for million-token inference —...
The drop follows NVIDIA licensing Groq's inference technology and hiring founder Jonathan Ross along with much of the senior team. Groq is now positioning as a data center operator rather than a chip designer, planning to exceed 200MW of inference capacity by 2027. TechCrunch...
AMD unveiled its first rack-scale system to directly contest Nvidia at the rack level, with engineering samples in H2 2026 and mass production targeted Q2 2027. Microsoft joins Meta, OpenAI and Oracle as customers; Meta plans 1 gigawatt of Helios racks by year-end against a lo...
A model-routing API covering OpenAI, Anthropic, DeepSeek, Moonshot, Minimax, Nvidia, xAI, and Z.ai, with strategies including preferring flex usage tiers, specifying up to three performance benchmarks for automatic selection, or sending only complex queries to expensive models...
The argument is that AI rack density outgrew AC distribution, and 800-volt DC cuts conversion stages between grid and GPU so more available power reaches compute (NVIDIA). Staged rollout: hybrid AC-compatible power rack in H2 2026, row power center supporting up to 2 MW per ro...
Two data points that tell the same story. First, Value Add Pulse counts four frontier launches in 30 days: Gemini 3.5 Pro, Grok 5, Anthropic's Fable 5 and Mythos 5, plus open-weight GLM-5.2 and Kimi K2.7. The model-layer moat compressed from quarters to weeks. Second, TechCrun...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.