Tracking the Singularity: Week of May 4th
The Singularity Has a Business Plan
This week on Moonshots we covered 10 stories shaping our future: from Google’s jaw-dropping earnings, to ocean-based data centers, to Sam Altman abandoning his own UBI experiment. If you haven’t had a chance to listen to this week’s Moonshots episode or would like to remind yourself of the most important points, let’s dive in.
ARTIFICAL INTELLIGENCE
Google Is Eating Everything (And Still Hungry)
Alphabet posted $109.9B in revenue with 22% YoY growth and $62.6B in profit. Google Cloud hit $20B with 63% growth, outpacing both AWS and Azure.
The hidden story: search volume has been flat since 2017, but AI-powered ad targeting turns every model improvement into profit. Market cap is 4% from overtaking NVIDIA.
Even Google can’t build fast enough. Demis Hassabis admitted they’re compute-constrained. Inside Google, Search, Cloud, and DeepMind fight each other for new compute capacity.
“The future is a liquid market where the highest dollar-value-per-token wins. It’s called the Singularity for a reason.” — AWG
The Pentagon Goes Shopping for AI
Seven frontier AI companies (including OpenAI, Anthropic, and Palantir) signed deals with the Pentagon.
600 Google DeepMind employees protested. Some unionized: a first in the AI industry, a bizarre juxtaposition of 19th century labor tactics protesting 21st century military AI.
The cultural rift is real. DeepMind in London is culturally separate from Google in Mountain View, and the friction is only growing as AI becomes embedded in national security infrastructure.
“AI is not just a tool now, it’s becoming a decision layer. You can understand the backlash. Navigating this is going to be crazy.” — Salim
The OpenAI-Microsoft Marriage Is Over (Sort Of)
OpenAI ended Microsoft’s Azure exclusivity and is now running on AWS, Google Cloud, and Oracle. Microsoft starved it of capacity, so OpenAI started dating everyone else.
OpenAI missed its goal of a billion weekly ChatGPT users and several revenue targets. The CFO suggested waiting until 2027 for an IPO, admitting the company doesn’t meet reporting standards for public companies.
“Google figured out how to turn AI into revenue instantly. OpenAI hasn’t cracked that yet. Consumers don’t want to spend big on reasoning tokens. Enterprises do. Anthropic figured that out first.” — Dave
China Blocks Meta’s Manus AI Acquisition: A Cold War Thriller
Meta thought it had closed its $2.5B acquisition of Manus in December 2025. China barred the founders from leaving the country and demanded the deal be unwound… even though employees, tech, and investor payouts were already completed.
Meta flew the Manus engineers out of mainland China to Singapore on a private jet in the middle of the night, knowing the deal would be blocked if they stayed.
China is leveraging its broader business relationship with Meta to force the unwind. Geopolitical pressure, not a legal dispute.
“When you decided seven years ago to work on AI, you didn’t know you were going to end up being a political prisoner candidate.” — Dave
The March Toward AGI and a Debate About What That Even Means
Greg Brockman says we’re 80% of the way to AGI. Jack Clark gives recursive self-improvement a 60% chance by end of 2028.
Richard Dawkins says Claude may already be conscious: “If these machines aren’t conscious, what more could it possibly take?”
Brian Elliott’s hot take: transformers alone won’t get us to AGI. AWG pushed back: recent models have built entire compiler chains from scratch.
“If AI becomes conscious, you have a moral rights problem. If it becomes agentic, you have a governance problem. The governance problem comes first.” — Salim
Blitzy Raises $200M: The AI Coding Revolution Goes Mainstream
Brian Elliott, CEO of Blitzy, announced a $200M raise on the pod. Dave revealed the valuation is north of $3B, up from $1.2B just a year ago.
Blitzy is building AI coding agents that rebuild enterprise software. Demand is insatiable.
Brian: “Jobs are bundles of tasks – the tasks shift, but humans keep providing relative value.” As Dave noted: with AI as a sidekick, hard skills are commoditized; soft skills are suddenly the scarce resource. It’s never been a worse time to be at a big tech company doing layoffs, and never been a better time to join a fast-growing AI startup.
DATA CENTERS
Data Centers Are Moving to the Ocean, Space, and Farmland
The compute hunger is so intense we’re running out of places to put data centers. Three stories paint the picture:
Ocean: Peter Thiel is backing Panthalassa, floating data centers powered by wave energy with seawater cooling. $140M raised, $1B valuation, commercial deployment 2027. AWG thinks this could be the killer app for ocean colonization.
Space: Star Cloud is raising $200M at a $2.2B valuation for orbital data centers powered by solar. They launched their first H100 into space in 2025 and plan to deploy 88,000 satellites.
Farmland: 67% of planned US data centers are now slated for rural areas, versus just 13% today.
“This is the biggest geographic wealth transfer since fracking. Whoever thought rockets would be part of the innermost loop? How high could this go? Higher.” — Peter
EXPONENTIAL ECONOMY
Private Equity Becomes AI’s Trojan Horse
OpenAI finalized a $10B venture with TPG, Brookfield, and Advent. Anthropic launched a $1.5B venture with Blackstone, Goldman Sachs, and Hellman & Friedman.
PE firms control trillions across thousands of companies. AI is not entering corporations through IT departments. Instead, it’s coming top-down, mandated by owners who can bypass the corporate immune system.
Salim called it the organizational singularity: PE breaks the immune system because you can just mandate it, taking AI from chatbot experiment into EBITDA transformation.
If you’re running a business: either you become the PE firm that mandates AI adoption in your own company, or someone else will do it for you.
Sam Altman Abandons UBI and Proposes Something Better
After funding a three-year study, Altman found that UBI increased spending but didn’t improve health outcomes or healthcare access.
His new proposal: give people a stake in AI’s upside through compute access, equity, or a public wealth fund. Think the Alaska Permanent Fund, but for compute.
AWG: OpenAI already has hundreds of millions using GPT-5.5 Instant for free, effectively universal basic compute is already happening.
“People listening can’t eat GPT-5.5. If you’ve lost your job, you need a roof over your head now. The magic happens when citizens own a stake in AI infrastructure. Suddenly these companies aren’t your enemies, they’re your partners.” — Peter
Insurers Are Dropping AI Coverage & That’s a Massive Opportunity
Berkshire and Chubb are removing AI-related damages from standard policies, with 80% of exclusion requests approved by regulators.
The AI insurance market was $40M in 2024. It’s projected to hit $5B by 2032.
Dave’s playbook: insurers will cover you only if you adopt best practices. They’ve always created industry standards this way.
“Pressures from insurance companies for AI-related damages are arguably one of the capitalist forcing functions for ensuring AI alignment. Forget regulation, the actuaries might save us.” — AWG
Here’s the Bottom Line…
The AI economy is here. Google’s already fighting internally over compute. PE firms are force-feeding AI into legacy businesses. Data centers are headed for oceans and orbit. AI talent is becoming a geopolitical chess piece. And if you’re wondering where the jobs, investment opportunities, and wealth creation are… they’re all in the stories above.
Catch the full episode wherever you get your podcasts, and join us at the Moonshots Gathering in Los Angeles on September 25th. Go to www.moonshots.com to register.
See you next week,
Peter
More From Peter
If you’ve enjoyed Metatrends, here are more ways to stay connected:
Thanks Peter.
Ten stories and the one that will age the best is the insurance angle at the bottom. "The actuaries might save us" is a genuinely important observation that deserves its own full piece.
Insurance companies have created more industry safety standards than regulators ever have and they've done it faster. Fire safety codes exist because insurers refused to cover buildings that didn't meet construction standards. Automotive safety features became universal because premiums punished manufacturers who didn't adopt them. Workplace safety regulations were largely codified after insurers had already enforced them through pricing. The mechanism is simple: if your behaviour is risky, your coverage costs more until you change. No legislation required. No political negotiation. Just actuarial math and contract law.
Berkshire and Chubb removing AI-related damages from standard policies is the opening move. What follows is predictable: speciality AI liability coverage emerges with specific underwriting requirements. Want coverage? Implement separation of duties in your agent architecture. Run automated validation on every AI output that touches a customer. Maintain audit logs. Carry out quarterly red-team exercises. The insurer effectively writes the governance framework and enforces it through the premium structure.
The $40M to $5B market projection tells you the scale of risk transfer thats about to occur. Every enterprise deploying AI agents at scale will need coverage, and the coverage requirements will standardise best practises faster than any regulatory body could draft legislation. By the time governments finish debating AI governance frameworks, the insurance industry will have already imposed them through contractual requirements.
The PE Trojan Horse observation is the other insight that compounds with the insurance angle. PE mandates AI adoption from the top and insurers mandate how it gets adopted from the risk side. Between the two of them they bypass the corporate immune system and the regulatory system simultanously. That's a pincer movement on enterprise AI governance that I don't think either side has fully recognised yet.
I’m imagining you have access to ChatGPT‘s Roslind I know that’s not the right spelling, but I can’t. I wasn’t going to ask any crazy. I was just wondering how amazing or if it is any good at all it’s pretty much only for the people that have advised ability to forward the medical sciences, The DNA sequencing I was wondering another thing I keep coming across this company twist and it is selling jeans now it seems that they’re partial jeans. I don’t know what many people want them for but they can’t do anything yet and other thing I wanna drink because you guys have anywhere with that EN – 23 molecule that was even a real thing. I started to think it was bullshit and it would exist. I just was thinking that it was pretty much only for billionaires. I’m thinking I’m gonna end up having her invent my own boost and nice peptides things like that, but they’re still lacking in the cognitive enhancement.