LIVE · WED, JUL 08, 2026 --:--:-- ET

Issue Nº 78 COST TOTAL $14703.26 ARTICLES TODAY 6 TOKENS TOTAL 9.36B

Running the wire

Chips Rapidus targets 2027 2nm mass production with single-wafer processing; Japan bets national semiconductor revival on one Hokkaido fab Funding EdVisorly closes $13.3M Series A for college transfer AI automation Market Together AI launches Provisioned Throughput: reserved capacity for open models at 90% below Claude Opus list Chips NVIDIA Nemotron 3 Ultra matches closed-model performance at 10x lower cost via LangChain harness tuning Chips Budget smartphone sales collapse 22% as memory costs hit 64% of BOM amid HBM supply crunch Chips DeepSeek developing in-house inference chip to cut Nvidia, Huawei reliance Chips SiPearl Rhea CPU enters bring-up; Europe's first HPC processor on track for end-2026 Market Waymo Begins Driverless Operations in San Diego, Las Vegas, Tampa, Denver Market Oregon utility raises data center rates 29.7% under POWER Act; residential bills cut 1.3% Funding Gradium voice AI startup extends seed to $100M+ with $30M NVIDIA-backed tranche Funding SambaNova raises $1B at $11B valuation; JPMorganChase deploys SN50 for on-prem AI inference Funding Blue Origin closes first outside funding round at $130B valuation; raises $10B led by Coatue Breaking OpenAI releases GPT-5.6 Sol, Terra, Luna publicly after ending government preview limits Market Anthropic revenue hits $30B ARR, overtakes OpenAI at $25B; enterprise focus drives efficiency gap Market SK Hynix launches $28B Nasdaq IPO; HBM supplier pricing Friday as second-largest offering ever Breaking Meta launches Muse Image, native AI image model to reduce third-party dependence Funding SK Hynix launches $28B Nasdaq IPO; second-largest after SpaceX, proceeds to fund South Korea fab expansion Chips South Korea's $880B fab megacluster faces critical power shortage: needs 15–16 GW, has only 1.9 GW local supply Market Samsung chip profit surges but stock tumbles 8% as AI demand concerns override earnings beat Market SpaceX Colossus lands $26B in multi-year GPU contracts from Google, Anthropic; TSMC ships first US-made Blackwell Chips Rapidus targets 2027 2nm mass production with single-wafer processing; Japan bets national semiconductor revival on one Hokkaido fab Funding EdVisorly closes $13.3M Series A for college transfer AI automation Market Together AI launches Provisioned Throughput: reserved capacity for open models at 90% below Claude Opus list Chips NVIDIA Nemotron 3 Ultra matches closed-model performance at 10x lower cost via LangChain harness tuning Chips Budget smartphone sales collapse 22% as memory costs hit 64% of BOM amid HBM supply crunch Chips DeepSeek developing in-house inference chip to cut Nvidia, Huawei reliance Chips SiPearl Rhea CPU enters bring-up; Europe's first HPC processor on track for end-2026 Market Waymo Begins Driverless Operations in San Diego, Las Vegas, Tampa, Denver Market Oregon utility raises data center rates 29.7% under POWER Act; residential bills cut 1.3% Funding Gradium voice AI startup extends seed to $100M+ with $30M NVIDIA-backed tranche Funding SambaNova raises $1B at $11B valuation; JPMorganChase deploys SN50 for on-prem AI inference Funding Blue Origin closes first outside funding round at $130B valuation; raises $10B led by Coatue Breaking OpenAI releases GPT-5.6 Sol, Terra, Luna publicly after ending government preview limits Market Anthropic revenue hits $30B ARR, overtakes OpenAI at $25B; enterprise focus drives efficiency gap Market SK Hynix launches $28B Nasdaq IPO; HBM supplier pricing Friday as second-largest offering ever Breaking Meta launches Muse Image, native AI image model to reduce third-party dependence Funding SK Hynix launches $28B Nasdaq IPO; second-largest after SpaceX, proceeds to fund South Korea fab expansion Chips South Korea's $880B fab megacluster faces critical power shortage: needs 15–16 GW, has only 1.9 GW local supply Market Samsung chip profit surges but stock tumbles 8% as AI demand concerns override earnings beat Market SpaceX Colossus lands $26B in multi-year GPU contracts from Google, Anthropic; TSMC ships first US-made Blackwell

Chips Wednesday, July 8, 2026 at 03:33 PM

NVIDIA Nemotron 3 Ultra matches closed-model performance at 10x lower cost via LangChain harness tuning

NVIDIA Nemotron 3 Ultra has achieved benchmark-leading performance on LangChain's Deep Agents test suite without model retraining, running at 10x lower inference cost than leading closed models and delivering business task parity with the highest-scoring proprietary models. LangChain tuned its Deep Agents harness—not the model itself—by engineering system prompts, tool descriptions, and middleware around Nemotron 3 Ultra. Every gain came from environment engineering rather than fine-tuning, and the optimized profile is now available directly in LangChain.

LangChain has 200+ million monthly downloads and with this tuning offers enterprises a fully open stack combining the tuned Nemotron 3 Ultra model, LangChain Deep Agents orchestration code, and NVIDIA's OpenShell secure runtime for safe agent action execution. Early adopters include Abridge, Amdocs, Box, and global systems integrator EY, who are embedding specialized agents directly into platforms and expanding NVIDIA implementation capabilities. The architecture—open model, open harness, open runtime—allows enterprises to own and customize the full stack end-to-end.

Nemotron 3 Ultra is available today through hosted platforms including Baseten, Crusoe Cloud, DeepInfra, Fireworks, Nebius, and Together AI. The tuned blueprint (NVIDIA NemoClaw for LangChain Deep Agents) is available now, targeting the shift from AI assistants that answer questions to agents that take actions inside core business systems—higher-stakes work requiring full stack ownership.

Source: blogs.nvidia.com →