🛎️ Lunar AI Factories

Good Morning, AI Enthusiasts!

The real competition is shifting from models alone to who controls the infrastructure that makes them run.

ANNOUNCEMENT

We’re excited to announce that, effective today, MyClaw.ai has officially ended its waitlist and is now fully open for public beta access.

MyClaw.ai is a fully managed, plug-and-play OpenClaw deployment that runs on a secure, isolated Linux VPS — built to give anyone a persistent, always-on 24/7 AI agent without any technical setup.

All users can now sign up and start using the platform immediately.

Thank you to everyone who supported us during the waitlist phase and welcome to the next stage of MyClaw.ai.

MOON

Lunar AI Factories Move From Vision to Plan

👀 What’s happening: At an internal meeting, Elon Musk proposed something most executives would not say out loud. Build AI satellites on the Moon and use a lunar factory and an electromagnetic mass driver to launch hardware into orbit. He framed the Moon as the near term construction site, not Mars. This is no longer distant colonization talk. It is infrastructure planning for AI beyond Earth.

🌍 How this hits reality: For almost everyone, lunar AI sounds like science fiction. Power, life support, mining, manufacturing, launch systems. Each piece is a national program. Musk already controls rockets, satellites, launch cadence, and a fast scaling AI company. Starship aims for high frequency reuse. Starlink proves orbital deployment at scale. As AI data centers start consuming as much power as small cities and regulators push back, shifting compute off Earth stops sounding crazy and starts sounding like vertical integration in space.

🛎️ Key takeaway: This is still fantasy for most governments and startups. For Musk, it is vertical integration extended upward. If Earth becomes the constraint, the Moon becomes the workaround. The foundation is already in place.

TOGETHER WITH QODER

GLM-5 Just Launched. Now on Qoder

GLM-5 dropped today—Zhipu's biggest upgrade yet. The base model scales from 355B to 744B parameters, trained on 28.5T tokens. Benchmarks show it surpassing Claude Opus 4.5 on complex coding and agentic tasks.

The shift: from "vibe coding" to agentic engineering. GLM-5 is built for production-grade work—multi-step debugging, system architecture, and long-running agents that don't break mid-task.

Qoder added support on day one.

INFRA

Grid Wars Heat up

👀 What’s happening: Anthropic says it will pay 100 percent of the grid upgrade costs needed to connect its data centers, including expenses normally passed to residents. This follows its $50 billion expansion plan across New York and Texas. The pledge mirrors moves by Microsoft and OpenAI, but Anthropic’s language is more absolute and direct.

🌍 How this hits reality: Electricity prices are now politically sensitive, and large data centers can add hundreds of megawatts of demand in a single region. Utilities often spread transmission and substation upgrades across ratepayers. By absorbing those costs, Anthropic signals it is willing to internalize infrastructure risk. That raises the bar for competitors. Power access is no longer a background constraint. It is a competitive weapon.

🛎️ Key takeaway: The fight is shifting from benchmark charts to substations. Capital is being deployed not just into models, but into grid leverage. AI scale will increasingly be determined by who can secure and finance electricity first.

SKILLS

OpenAI Formalizes Agent Skills as Workflow Layer

👀 What’s happening: OpenAI officially introduced Agent Skills in its API, allowing developers to package reusable workflows as versioned bundles with a SKILL.md manifest. These skills mount into hosted or local shell environments and execute only when invoked. Instead of embedding complex procedures inside system prompts, workflows now live as attachable, executable units.

🌍 How this hits reality: Until now, multi step logic lived in bloated prompts or fragile tool schemas. Skills separate global behavior from procedural execution. They support version pinning, file bundles up to 50MB, and controlled execution inside container environments. Unlike Anthropic’s more open, cross ecosystem skill direction, OpenAI’s approach ties workflow packaging directly to its runtime layer, tightening integration between execution, governance, and model reasoning.

🛎️ Key takeaway: This is structural, not cosmetic. Prompts defined behavior. Skills define process. As agents move from answering to executing, the prompt era gives way to workflow engineering skill as the core discipline.

DAILY TL;DR

OpenClaw Trending

• Peter Steinberger joked that even without software knowledge, his OpenClaw now brews beer and runs half of his design company.
• Jason said on X he’s offering $25,000 to back openclaw-based startups, with investment support.
• 0xMarioNawfal said that OpenClaw has launched for TradingView, enabling AI agents to analyze charts and execute trades in real time.
• Eli Mernit argues that modeling a company as a filesystem, orchestrated by Claude, lets agents like OpenClaw run the business by reading and writing files.
• Matthew Berman said OpenClaw paired with GPT5.3 Codex and Opus 4.6 forms a powerful stack, which he demonstrated across automation use cases in a video.

AI Trending

• Mozart AI raised $6 million to expand its generative audio workstation and bring AI-powered music creation to a broader audience.
• OpenAI researcher Zoë Hitzig resigned over ChatGPT ads, warning they could erode user trust and incentivize data-driven manipulation.
• ByteDance is reportedly developing its own AI chip and in talks with Samsung to produce over 100,000 units this year.
• Threads launched “Dear Algo,” an AI feature that lets users publicly request short-term feed adjustments to personalize and enhance real-time content.
• The New York Times built in-house AI tools to transcribe and summarize podcasts to track the “manosphere” and support investigative coverage.

TRENDING

Daily AI Launches

• Tinkerer Club has launched a lifetime-access community for self-hosters and local AI enthusiasts, featuring tools like Clawdbot.
• Agent Builder by Thesys has launched a no-code platform to build AI agents that output charts, forms, and slides.
• Normain has launched a document AI that prioritizes structured, traceable data extraction over chat-based summaries.
• Spawned has launched an AI app builder with a built-in launchpad and Solana token integration for community ownership.

Featured AI Tools

• ✨ MuseMail.ai makes creating stunning, on-brand emails from one prompt with AI precision.
• 🪶 CopyOwl is the First AI Research Agent, deep research on any topic in one click.
• 🦾 Flot AI writes, reads, and remembers across any apps and webs.

READ MORE

Let the Future Come to Your Inbox

Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.

• AI Secret uncovers what really matters in AI
• Bay Area Letters decodes tech and business shifts from Silicon Valley
• Robotics Herald tracks how robots move from labs into daily life
• Marketing Secret breaks down real growth and go-to-market playbooks
• The Hardwire explores hardware, consumer tech, and what’s coming next
• Oncely helps you discover the next wave of AI products and launches

TOGETHER WITH US

AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.

We've helped promote over 500 Tech Brands. Will yours be the next?

Email our co-founder Mark directly at mark@aisecret.us if the button fails.