🙀 SpaceX raised $75B for AI in space

Welcome, humans.

If you know ChatGPT and Gemini but every time you hear the word “Anthropic” you scratch your head, this video is for you. Anthropic makes Claude, and they’ve been the main character of the AI industry for pretty much all of 2026.

Emily Chang of Bloomberg dives deep with Claude Code’s creator Boris Cherny, and co-leads and co-founders Daniela and Dario Amodei, and it’s kinda like an abridged history of the entire company, including why Dario, Daniela, and Anthropic’s fellow cofounders left OpenAI to start OpenAI… they just didn’t trust Sam. We break down the key insights here.

Here’s what happened in AI today:

• 🙀 SpaceX priced a record IPO around AI infrastructure.

• 📰 Anthropic walked back invisible Claude Fable safeguards after backlash.

• 📰 OpenAI acquired Ona to expand Codex workspaces for agents.

• 🍪 Perplexity put Deep Research inside Computer for agents.

• 💡 Dario Amodei warned AI policy is moving too slowly.

…and a whole lot more that you can read about here.

Hey: Want to reach 700,000+ AI-hungry readers? Advertise with us! 

P.S: Love robots? We just launched our new robot-specific newsletter! Check it.

🙀 SpaceX's IPO Turned Data Centers Into the Next Space Race

Every few years, Wall Street gets one number that tells you what future it wants to buy.

This week, that number was $75B.

SpaceX priced its IPO (the first public sale of stock) at $135 a share, raising $75B and valuing the company around $1.77T. Demand reportedly hit roughly $250B, with BlackRock placing a $5B order.

Here's what happened:

• SpaceX raised more than any company in IPO history.

• Retail investors got an unusually large slice of the sale.

• Investors are pricing rockets, Starlink, and AI infrastructure into one story.

• The data-center angle is the part to watch.

The obvious story is rockets. The actual story is data. SpaceX is one of the few companies where launch capacity, internet distribution, and compute ambition live under the same roof.

Why this matters: AI companies are running into very physical limits: chips, electricity, cooling, land, and the pipes that move data around. The industry is already acting less like software and more like heavy industry, with data-center fights over power, water, labor, and permits.

SpaceX sits in a strange place inside that bottleneck. It already owns a global satellite network through Starlink. It also has the launch capacity to put more hardware in orbit than anyone else. If AI demand keeps climbing, that combination starts to look like infrastructure, not only transportation.

That gives investors a clean thesis: if AI needs more compute, and Earth keeps making data centers harder to build, SpaceX might become part of the next AI infrastructure layer.

The bull case looks like this:

• Starlink becomes the data pipe.

• SpaceX launches the hardware.

• AI customers buy capacity, connectivity, or both.

The bear case is a lot more basic: what the heck are you smoking bro, you’re going to put computers in space and try to sell it to AI companies? What could go wrong?? 

Hey, at least it would be “actual” cloud computing! 

Our take: The market did not buy a space company, or at least, not a space company worth whatever the trillion it ends up at later today when retail investors can buy in. 

No, the market bought a possible answer to the AI infrastructure crunch, with a giant Elon premium stapled to the top (and a seemingly lucrative lower-earth satellite business). The rocket business got the headline, but the IPO price only makes sense if investors believe SpaceX becomes a new layer of AI infrastructure. 

Now, SpaceX has to prove orbit can be more than the most expensive investor mood board ever assembled. Here’s to hoping, but y’know… not financial advice either way!

FROM OUR PARTNERS

Your AI lab just got a desk address

Most AI experiments start in a browser tab. The serious ones need more room.

Dell Pro Max with GB10, powered by the NVIDIA GB10 Grace Blackwell Superchip, gives builders a compact desktop for AI development with 128GB memory, NVIDIA DGX OS 7, and 4TB storage.

Use it to test agents, prototype workflows, and run model experiments without turning every idea into another cloud line item.

For founders, technical teams, and AI-obsessed operators, this is the “let’s actually build it” machine.

Click here to see the Dell Pro Max with GB10.

🎓 AI Skill of the Day: Cut Coding-Agent Token Waste by Routing Work

Claude Code is amazing until your weekly limit disappears because the model spent premium tokens typing boilerplate.

CJ Zafir shared a simple routing workflow that he says cut his Claude Code limit burn by 50%: use Claude Fable 5 for planning and final review, then hand the actual implementation to Codex GPT-5.5. In plain English: let Claude do the thinking and quality control, while Codex does the typing.

The setup is simple:

1. Install the OpenAI Codex plugin inside Claude Code.

2. Use Claude Fable 5 High for the plan.

3. Use Codex GPT-5.5 xhigh for execution, using your Codex plan and no API.

4. Bring the result back to Claude Fable 5 Max for review.

Use this workflow when you have a big coding or research job where planning quality matters, but the execution would waste your best Claude tokens.

Use this routing workflow for the task below:

Task: [paste task]

Step 1: Claude Fable 5 High should create the plan.
- Define the goal.
- Break the job into clear implementation steps.
- Identify files, tools, tests, or sources needed.
- Write the execution instructions for Codex.

Step 2: Codex GPT-5.5 xhigh should execute the plan.
- Follow the plan exactly.
- Make the needed changes.
- Run checks or tests where possible.
- Return a concise report of what changed.

Step 3: Claude Fable 5 Max should review the result.
- Check whether the work matches the original goal.
- Identify bugs, missing context, or weak assumptions.
- Suggest final fixes.
- Give me a plain-English verdict: ship, revise, or rerun.

Favorite insight: expensive models should handle judgment. Cheaper execution models should handle the grind.

Want more tips like this? Check out our AI Skill of the Day Digest for June.

Total AI beginner? Start here (goes with this video).

Have a specific skill you want to learn? Request it here. 

🍪 Treats to Try

1. Perplexity Deep Research turns hard questions into reports, dashboards, and decks inside Computer, using connectors and long-running research workspaces, no pricing details.

2. Use Computer gives your agents real Mac, iPhone, iPad, Apple Watch, and Ubuntu sandboxes to click, type, inspect screens, and record workflows, no pricing details.

3. Pool organizes your screenshots into searchable collections, recovers original links, and suggests what to do next, free to try.

4. ElevenLabs Avatars creates talking videos from a script, voice, and reusable avatar, so your course or marketing character stays consistent, no pricing details.

5. OpenCreator turns one-sentence video ideas into scripts, casting, storyboards, model generation, and final edits in one workspace, free to start.

Learn Agents in 2026 With This Total Beginner’s Guide to AI Agents & Automation

AI agents sound intimidating until you break them down into one basic idea: using AI to handle repetitive work so humans can focus on judgment calls. In this beginner-friendly Neuron Live, Grant and Corey explain what agents are, how automation workflows actually work, and where tools like ChatGPT, Claude, Make, and ClickUp fit in.

Watch/read: YouTube | Blog version

📰 Around the Horn

• Anthropic walked back invisible Claude Fable 5 safeguards after researchers said legitimate AI, cyber, and bio work was blocked or silently switched to Opus 4.8.

• OpenAI agreed to acquire Ona, giving Codex secure, persistent cloud environments for long-running enterprise agents.

• Google DeepMind and partners launched up to $10M in multi-agent safety research funding.

• Anthropic launched Claude Corps, a $150M fellowship that will place 1,000 fellows inside nonprofits for year-long AI coaching.

• Google committed $50M to train 300,000+ skilled-trade workers for data-center and infrastructure jobs.

• CISA shortened the fix window for serious federal cyber vulnerabilities to three days as AI-enabled threats rise.

Want absolutely EVERYTHING that happened in AI this week? Read the full ATH digest.

FROM OUR PARTNERS

Quick question about newsletter ads

We’re running a super short survey to see if our newsletter ads are being noticed. It takes about 20 seconds and there's just a few easy questions.

Your feedback helps us make smarter, better ads.

Take the survey now.

💡 Intelligent Insights:

• Dario Amodei argues AI is moving faster than government and wants mandatory third-party testing for high-risk models.

• Carnegie argues AI infrastructure is now geopolitics, with time-to-power becoming the key bottleneck.

• John O’Farrell warns that pro-AI political spending could preempt democratic debate before voters understand the tradeoffs.

• Mitchell Hashimoto says Fable 5 shines in targeted optimization loops, not broad everyday coding, which is a useful model-selection reminder.

• In Interconnects, Nathan Lambert weighed in on Anthropic walking back the silent model manipulation it shipped with Claude Fable, arguing the secret nerfing of AI researchers cost more user trust than an openly disclosed restriction would have.

• In AI as Normal Technology, Arvind Narayanan and Sayash Kapoor argued the "AI is replacing software engineers" narrative is based on AI-washing of layoffs, and the best data shows U.S. software engineer employment is still growing, and they expect software engineering skills and judgment to remain in demand (Dario said similar in the Bloomberg interview; the pie of engineering work will grow overall, it's the slow to adapt players that could dwindle or collapse)

• A group of AI researchers and policy experts launched Europe 2031, a fictional five-year scenario tracing how Europe slides into AI-driven irrelevance through compute dependence and institutional inertia, ending with five recommendations to change course.

A Cat’s Commentary

P.S: Before you go… have you subscribed to our YouTube Channel? If not, can you?