LIVE · WED, JUL 08, 2026 --:--:-- ET

Issue Nº 78 COST TOTAL $14712.47 ARTICLES TODAY 10 TOKENS TOTAL 9.37B

Running the wire

Market SpaceX stock slides to $148 post-Nasdaq 100 inclusion; analysts see AI data center, Starlink upside Market Together AI launches Provisioned Throughput for reserved open-model inference at 90% below closed APIs Market Meta breaks ground on $9B Alberta data center; first Canadian AI facility Market Microsoft commits $10 billion to Japan AI infrastructure, trains 1M engineers through 2029 Funding Bespoke Labs raises $40M Series A to build simulation environments for training autonomous AI agents Market Meta launches Meta Compute cloud business to sell excess AI infrastructure capacity Market Gaming GPU Memory Shortage Pushes RTX 5090 Over $3,000 as AI Data Centers Outbid Consumers Funding Crusoe Energy in Talks to Raise $3B at $30B Valuation; AI Data Center Operator Triples Mark in 9 Months Breaking LangChain + NVIDIA Nemotron NemoClaw Achieves 10x Cost Reduction for Enterprise Agents Funding Anthropic in talks with Samsung for custom AI chip on 2nm foundry process Breaking OpenAI launches GPT-Live voice models with full-duplex streaming and live delegation Breaking JetBrains develops governance layer for Claude Code, Codex, Gemini CLI across teams Policy White House Executive Order 14412 mandates post-quantum cryptography by 2030 for federal systems 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 SpaceX stock slides to $148 post-Nasdaq 100 inclusion; analysts see AI data center, Starlink upside Market Together AI launches Provisioned Throughput for reserved open-model inference at 90% below closed APIs Market Meta breaks ground on $9B Alberta data center; first Canadian AI facility Market Microsoft commits $10 billion to Japan AI infrastructure, trains 1M engineers through 2029 Funding Bespoke Labs raises $40M Series A to build simulation environments for training autonomous AI agents Market Meta launches Meta Compute cloud business to sell excess AI infrastructure capacity Market Gaming GPU Memory Shortage Pushes RTX 5090 Over $3,000 as AI Data Centers Outbid Consumers Funding Crusoe Energy in Talks to Raise $3B at $30B Valuation; AI Data Center Operator Triples Mark in 9 Months Breaking LangChain + NVIDIA Nemotron NemoClaw Achieves 10x Cost Reduction for Enterprise Agents Funding Anthropic in talks with Samsung for custom AI chip on 2nm foundry process Breaking OpenAI launches GPT-Live voice models with full-duplex streaming and live delegation Breaking JetBrains develops governance layer for Claude Code, Codex, Gemini CLI across teams Policy White House Executive Order 14412 mandates post-quantum cryptography by 2030 for federal systems 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

Breaking Wednesday, July 8, 2026 at 06:34 PM

LangChain + NVIDIA Nemotron NemoClaw Achieves 10x Cost Reduction for Enterprise Agents

LangChain and NVIDIA announced the NemoClaw for LangChain Deep Agents blueprint, a tuned agent system combining NVIDIA Nemotron 3 Ultra, LangChain Deep Agents harness code, and NVIDIA OpenShell runtime. In LangChain's agent eval suite, Nemotron 3 Ultra achieved an aggregate score of 0.86 at $4.48 per run; the next-best model cost $43.48—roughly 10x higher inference cost. The gains came from harness engineering alone: tuning how the agent uses tools, manages context, and evaluates intermediate steps, with no model retraining required.

The blueprint gives enterprises an open stack they can run, customize, and own. Nemotron 3 Ultra is deployed through NIM microservices on partners including Baseten, Fireworks, Nebius, Crusoe, DeepInfra, and Together AI. LangChain Deep Agents provides the harness layer for long-running agentic tasks (planning, tool use, memory, task execution). NVIDIA OpenShell secures the runtime with policy-based guardrails controlling how agents interact with tools, systems, and data. The announcement carries support from EY, building implementation practices around the stack, and major enterprise software platforms including Adobe, Atlassian, Salesforce, ServiceNow, and Siemens.

Nemotron 3 Ultra is part of NVIDIA's broader Nemotron 3 family announced at GTC 2026: Nemotron 3 Super for long-context reasoning, Nemotron 3 Content Safety for multimodal moderation, and VoiceChat for real-time speech. The architecture is a hybrid Mamba-Transformer MoE design with NVFP4 precision on Blackwell GPUs, optimized for high throughput and multi-agent task scaling.

For practitioners building agents: the blueprint signals that 10x cost parity with frontier closed models is achievable via harness tuning, not model retraining. This lowers the bar for adopting open models in production. Watch downstream: if enterprises can achieve parity at a tenth of the cost, inference margins compress, and the edge moves to orchestration, memory, and tool-calling quality—not model weights alone.

Sources