Howdy, it’s Barsee again.

Happy Tuesday, AI family, and welcome back to AI Valley.

Today’s climb through the Valley reveals:

• OpenAI expands push into humanoid robotics

• $100 billion Nvidia-OpenAI deal

• Unitree’s G1 humanoid shows rapid recovery in new demo

• Plus trending AI tools, posts, and resources

Let’s dive into the Valley of AI…

WISPR FLOW

Your keyboard called, it’s taking a long weekend

We’ve been banging away on keyboards for 150 years.
Until now, voice dictation hasn’t been reliable enough to change that.

Wispr Flow finally delivers the no-edit confidence we’ve all been waiting for:

• 4× quicker than typing. Dictate emails, docs, and DMs in real time and save precious hours every week.

• AI auto-edits on the fly. Flow cleans filler words, fixes grammar, and formats perfectly as you speak.

• Works inside every app with no setup. Fly through Slack notifications, give more context to ChatGPT, or brain dump into Notion.

• Use it at your desk or on the go. Available on Mac, Windows, and iPhone. 

“This is the best AI product I’ve used since ChatGPT.” - Rahul Vohra, CEO, Superhuman

Give your hands a break ➜ start flowing for free today. 

_{*This is sponsored}

THROUGH THE VALLEY

1/ OpenAI expands push into humanoid robotics

OpenAI has been quietly scaling its robotics ambitions, steadily building a team with expertise in mechanical design, sensing, and software for real-world mobility. Recent job postings point to a goal of “general-purpose robotics” (the physical counterpart to its mission of developing AGI-level intelligence in digital form).

Although it remains unclear whether OpenAI will directly build humanoid hardware or focus on software overlays, new hires and filings leave little doubt about intent. In January, the company trademarked “user-programmable humanoid robots.” More recently, it recruited Stanford roboticist Chengshu Li and, in September 2025, drew multiple engineers away from Tesla’s Optimus project. Quiet partnerships with Boston Dynamics and investments in 1X Technologies and California-based Figure suggest a multi-pronged approach spanning both software platforms and full-stack humanoid ventures.

Why does it matter?

Every advance in language or planning models eventually demands a physical arena to reveal its limits. Analysts forecast humanoid markets could exceed $150 billion by 2035 and $5 trillion by 2050 if adoption in logistics, eldercare, and manufacturing accelerates. Yet cost, safety, and public trust remain stubborn hurdles. OpenAI’s moves show it intends to shape not just the brain of future machines but their bodies too, a dual ambition no AI lab has attempted at this scale since Google X’s robotics push nearly a decade ago.

2/ $100 billion Nvidia-OpenAI deal

If humanoids are the body, compute is the bloodstream. OpenAI is tying its future to one of the biggest infrastructure projects in AI history. Nvidia has agreed to put up as much as 100 billion dollars to build datacenters with 10 gigawatts of capacity, starting in 2026 on its Vera Rubin chips (designed to succeed Hopper/Blackwell). That would mean 4 to 5 million GPUs, twice as many as the entire industry shipped in 2024.

This deal is even bigger than OpenAI’s 300 billion dollar partnership with Oracle for just 4.5 gigawatts of cloud power (less than half the capacity of the Nvidia plan). Rivals like Microsoft, Meta, and Musk-backed projects are also building campuses at similar scales. By 2030, datacenter investment could reach five to eight trillion dollars worldwide as AI demand doubles the size of today’s infrastructure.

Why does it matter?

Control of compute has become the new oil. For Nvidia, the deal secures GPU demand for years. For OpenAI, it provides the power and hardware base needed to chase what it calls superintelligence. With connections to nuclear and fusion pilots, the deal also ties AI’s growth to the energy transition, suggesting that power grids and AI development are set to become closely linked.

3/ Unitree’s G1 humanoid shows rapid recovery in new demo

In the commercial robotics space, Chinese company Unitree is pushing affordable humanoids closer to real-world use. Its G1 robot costs only 16 thousand dollars and has now shown an “anti-gravity” mode that lets it get back up after being knocked down. The G1 stands 92 cm tall, weighs 35 kg, has dexterous hands, and connects to G1 Cloud, a platform where robots can upload and share skills.

The demo drew mixed reactions. Some praised its stability, while others questioned why the company keeps showing combat-style trials instead of practical tasks in warehouses or factories.

Why does it matter?

The G1’s biggest breakthrough may not be its stability but its price. While Boston Dynamics and Tesla market humanoids as flagship projects for large companies, Unitree is aiming at startups, researchers, and small businesses that could never afford traditional robots. If the G1 proves useful at this price, it could spread humanoid adoption the way personal computers spread computing decades ago.

TRENDING TOOLS

• Nebuis > The cloud that gets AI workloads. Switch to Nebius and get up to 3 months off your contract *

• Lychee > Your AI “pet” in iMessage that nudges (and emotionally manipulates) you into being productive

• Creao > Describe the app you want, and Creao’s AI builds the full backend and frontend

• Perplexity Email Assistant > Perplexity Max subscribers now get a personal AI right in their inbox. It can draft replies, organize mail, and even schedule meetings

• Happenstance > Need the full picture on someone? This tool digs up everything you need to know

• Pie > An AI-powered QA team that compresses thousands of hours of product testing into just 30 minutes

• Capalyze > The first AI agent built to collect data, analyze it, and generate polished reports

_{(*) signifies sponsored tool}

THINK PIECES / BRAIN BOOST

• Google launched "Learn Your Way" which basically takes whatever boring chapter you're supposed to read and rebuilds it around stuff you actually give a damn about

• Everyone keeps saying LLMs can't do real logical reasoning. Turns out we've just been teaching them wrong this whole time

• OpenAI and Anthropic are building “AI co-workers” by turning the entire economy into a reinforcement learning machine

• How to create apps without knowing how to code

• A new Yale paper reveals the reality of the AGI economy

• Major technological innovation beyond 2030

• Music streaming services are being overrun with AI songs

• The great falls of Boeing, Intel, and Apple

• Google dropped a 64-page guide on building AI Agents

• This GitHub has a library of 90+ prompts for Nano Banana

• Notion 3.0’s new AI agents can be tricked into leaking data through a malicious PDF

• What does the future hold for generative AI?

• How Brian Eno anticipated the creative dynamics of AI by decades

THE VALLEY GEMS

What’s trending on social today:

THAT’S ALL FOR TODAY

Thank you for reading today’s edition. That’s all for today’s issue.

💡 Help me get better and suggest new ideas at [email protected] or @heyBarsee

👍️ New reader? Subscribe here

Thanks for being here.

REACH 100K+ READERS

Acquire new customers and drive revenue by partnering with us

Sponsor AI Valley and reach over 100,000+ entrepreneurs, founders, software engineers, investors, etc.

If you’re interested in sponsoring us, email [email protected] with the subject “AI Valley Ads”.