Mira Murati’s Thinking Machines Inks Massive NVIDIA Compute Deal

Read original article

Mira

Murati’s Thinking Machines Inks Massive NVIDIA Compute Deal Mira Murati’s two-year-old AI research startup, Thinking Machines Lab, announced a massive multi-year strategic partnership with NVIDIA. The deal, which includes a strategic investment from Nvidia, secures at least one gigawatt of the newly released Vera Rubin compute systems for Murati's team starting in 2027. The 1-gigawatt power play: Vera Rubin Deployment: The lab will deploy at least one gigawatt of Nvidia’s latest Vera Rubin systems starting in 2027. Strategic Investment: Nvidia is joining a $2 billion funding pool alongside Andreessen Horowitz and rival AMD’s venture arm.

Architecture Synergy: Both companies committed to co-developing specialized training and serving systems specifically for Nvidia’s hardware.
Founding Team Turmoil: The deal lands even as three co-founders recently left Murati to return to OpenAI, following co-founder Andrew Tulloch’s exit to Meta.

Jensen Huang recently explained that NVIDIA is finished investing in OpenAI and Anthropic because the "IPO window" for such consequential companies is closing. He told investors that if Nvidia simply provides the compute, "revenues will follow" regardless of whether they own a piece of the company.

Special Highlight from our network:

AI agents can think. Now they can use the web AI agents can reason, generate and plan. But most still struggle to access the live web. Blocks, CAPTCHAs and anti-bot systems stop them before they start.

Bright Data's Web MCP Server changes that. One tool that connects any AI agent to any website, reliably and at scale. It gives your agent four capabilities: Search the web in real time Crawl and extract websites into LLM-ready formats Access any public web content, with JavaScript rendering Navigate and interact with live pages It plugs straight into Claude, LangGraph, Google ADK, Cursor, CrewAI, Amazon Bedrock and more. 5,000 requests a month are free. Try it here This is the infrastructure layer that is taking AI agents from conversation to action on the open web.

Broad Topic Coverage: The feature launched with 70+ modules covering advanced subjects like Coulomb’s law, Charles’ law, and compound interest.
Educational Integration: Dynamic visuals join existing tools like "Study Mode" and "QuizGPT" to help the 140 million people using ChatGPT weekly for STEM help.
Competitive Landscape: This follows Google’s November 2025 launch of interactive diagrams for Gemini, intensifying the AI "EdTech" arms race.

KPMG: Why 55% of Top CEOs Are Hiring Because of AI A new KPMG survey of 100 top CEOs reveals that only 9% plan to cut jobs due to AI in 2026, while 55% are actually increasing hiring as a direct result of their AI investments.

Investment Reality: Nearly 80% of CEOs now dedicate at least 5% of their total capital budgets to AI. The Process Bottleneck: The primary barrier to ROI is the time-consuming work of rewiring legacy business processes.
Long-term Optimism: While 77% believe GenAI was overhyped last year, they admit its 10-year disruptive power is being underestimated.

Core Functions: Organizational Hierarchy: Establishes clear roles where an AI CEO delegates tasks to a CTO and engineers, ensuring structured goal alignment.

Clipmart Marketplace: An upcoming feature to download pre-built company templates—such as trading desks or content agencies—for one-click deployment.

Special Highlight from our network:

Chronicle approaches presentations differently You start with the story. Drop rough notes or an outline onto a flexible canvas. Chronicle turns that structure into a clear visual presentation. Less time aligning boxes and fixing slides.

More time shaping the narrative. It’s a small shift, but an important one. AI won’t replace creative thinking. But it can remove the formatting work between ideas and communication.

Explore Chronicle here ChatGPT Can Now Create Interactive Visuals: The New AI Teacher? OpenAI introduced dynamic visual explanations , allowing ChatGPT users to interact with math and science concepts in real time. Instead of static diagrams, users can now manipulate formulas and variables within interactive modules to see immediate changes. This learning tool currently supports over 70 topics including Ohm's law, kinetic energy, and the Pythagorean theorem.

I nteractive learning stack: Real-Time Manipulation: Users can adjust lengths or variables, such as triangle sides or lens equations, and watch the results update instantly.

OpenAI is effectively turning ChatGPT into a real-time simulator. While Google Gemini launched similar interactive diagrams in November 2025, OpenAI is leveraging its massive user base to normalize AI as a "sandbox" for learning.

Corporate America is shifting from experimentation to structural transformation, treating AI as a long-term play rather than a quick way to slash payroll. State of the AI workforce: The Hiring Surge: 55% of leaders at $500M+ companies are adding headcount to support their AI roadmap.

We are currently in a "messy but exciting" transition phase. While 77% of CEOs admit generative AI was overhyped in the short term, they simultaneously believe its long-term disruptive power over the next decade is being dangerously underestimated. Tool of the Day: Paperclip (Code) Paperclip (Code) is the latest open-source sensation for managing AI agent swarms by organizing them into a functional company structure. It moves beyond messy shell scripts, allowing tools like Claude Code, OpenClaw, and Codex to operate within a coordinated hierarchy rather than in isolated terminals.

Automated "Heartbeats": Agents wake on a set schedule to check and execute work autonomously, removing the need for constant "babysitting". Cost & Audit Controls: Includes per-agent monthly budget limits to prevent surprise bills and maintains a full audit trail of every decision and tool call.

Try this yourself: If you have Node.js (v20+) installed, run npx paperclipai onboard --yes in your terminal. The app will launch in your browser to help you "hire" your first AI CEO and coder to get your autonomous company running.

Mira Murati’s Thinking Machines Inks Massive NVIDIA Compute Deal

Read original article

Mira

Architecture Synergy: Both companies committed to co-developing specialized training and serving systems specifically for Nvidia’s hardware.
Founding Team Turmoil: The deal lands even as three co-founders recently left Murati to return to OpenAI, following co-founder Andrew Tulloch’s exit to Meta.

Special Highlight from our network:

Broad Topic Coverage: The feature launched with 70+ modules covering advanced subjects like Coulomb’s law, Charles’ law, and compound interest.
Educational Integration: Dynamic visuals join existing tools like "Study Mode" and "QuizGPT" to help the 140 million people using ChatGPT weekly for STEM help.
Competitive Landscape: This follows Google’s November 2025 launch of interactive diagrams for Gemini, intensifying the AI "EdTech" arms race.

KPMG: Why 55% of Top CEOs Are Hiring Because of AI A new KPMG survey of 100 top CEOs reveals that only 9% plan to cut jobs due to AI in 2026, while 55% are actually increasing hiring as a direct result of their AI investments.

Investment Reality: Nearly 80% of CEOs now dedicate at least 5% of their total capital budgets to AI. The Process Bottleneck: The primary barrier to ROI is the time-consuming work of rewiring legacy business processes.
Long-term Optimism: While 77% believe GenAI was overhyped last year, they admit its 10-year disruptive power is being underestimated.

Core Functions: Organizational Hierarchy: Establishes clear roles where an AI CEO delegates tasks to a CTO and engineers, ensuring structured goal alignment.

Clipmart Marketplace: An upcoming feature to download pre-built company templates—such as trading desks or content agencies—for one-click deployment.

Special Highlight from our network:

More time shaping the narrative. It’s a small shift, but an important one. AI won’t replace creative thinking. But it can remove the formatting work between ideas and communication.

I nteractive learning stack: Real-Time Manipulation: Users can adjust lengths or variables, such as triangle sides or lens equations, and watch the results update instantly.