OpenAI Lost $21 Billion in 3 Months — Here’s What the Headlines Missed
June 28, 2026
Every Claude Tool Explained: Master Full Ecosystem in 20 Minutes
June 29, 2026
By C. Rich
There is a distinct flavor of irony in watching the tech elite stumble blindly through the AI infrastructure bottleneck. For the past two years, the evangelical hyperscalers have operated under a singular, blunt-force thesis: buy more land, secure more terrestrial gigawatts, and print more GPUs. But the math on the ground has already broken. We are staring down a severe regulatory and power-grid wall. Terrestrial data centers are choking on their own thermal output and structural delays. If you want to find alpha in this market, you don’t look at the metrics retail investors are screaming about on quarterly calls. You look for the pattern deviation. You look at what the biggest players are doing quietly, out of view, while the rest of the herd is looking down.
My investment philosophy has always been built on high-conviction positions in the foundational architecture of the sovereign supply chain, the unglamorous infrastructure that makes the rest of the world function. And right now, the dots are spelling out a narrative that almost the entire market is mispricing. It’s a story about the ultimate convergence of irony sitting directly at the crossroads: Planet Labs (PL). To understand where this is going, you have to look back to 2017. Google sold its SkySat constellation (Terra Bella) to Planet Labs. At the time, Wall Street applauded. The consensus was that running hardware in orbit was a capital-intensive distraction. Google wanted to be an archivist, not an aerospace engineer. They traded the physical assets for an equity stake and a data contract, viewing satellites merely as cameras meant to take static snapshots for Google Maps. But history doesn’t repeat; it rhymes. When the compute demands of large-scale AI models exploded, the limiting friction shifted from silicon availability to sheer power capacity.
Enter Project Suncatcher.
When Google quietly initiated this project, the market treated it like a standard corporate moonshot. But look closer at the mechanics: they are targeting prototype orbital data center launches by early 2027. Why? Because the physics of space offer an elegant escape hatch from Earth’s environmental and energy constraints.
- Unfiltered, 24/7 solar efficiency without atmospheric degradation offers an eight-fold advantage over terrestrial solar arrays.
- The infinite cryogenic sink of space solves the massive liquid cooling crisis plaguing earthbound clusters.
The mainstream media spent months talking about Google negotiating launch deals with SpaceX. But the crowd completely missed the structural component: Planet Labs is the partner building the actual satellite architecture. Planet Labs isn’t just an imagery vendor anymore; they are the group ruggedizing Google’s custom edge TPU chips to survive and compute in orbit. This isn’t a speculative thesis; it’s a structural necessity. Planet Labs possesses a data asset that cannot be replicated: a massive constellation of small “Dove” satellites capturing a high-resolution visual scan of the entire Earth’s landmass every single 24 hours. If you view Planet Labs through the lens of traditional SaaS or defense imagery contracts, you are missing the forest for the trees. The true strategic value isn’t the photos, it’s the temporal cadence.
[Daily Orbital Scan] ---> [Continuous Data Stream] ---> [Gemini Multi-Modal Architecture] ---> [Live Predictive Simulation Engine]
Current geospatial engines treat the world like a static museum. Integrating Planet’s real-time daily feed transforms a digital archive into a living, breathing <b”>World Model. When you feed a continuous, 24-hour global update directly into a multimodal AI architecture, the system stops merely labeling static pixels. It begins to map, track, and predict global supply chains, crop yields, port congestion, and maritime traffic in real time. It turns the digital map into an active simulation engine of human and natural activity. For a hyperscaler trying to maintain an absolute edge in AI utility, that data stream isn’t just a feature; it’s a sovereign moat. When you synthesize these pieces, the severe constraints of the terrestrial power grid, Google’s deep institutional relationship with Planet Labs, and the absolute requirement for proprietary, high-cadence datasets to feed the next generation of AI models, the conclusion becomes a matter of industrial logic.
Planet Labs represents a natural, vertical extension of the tech infrastructure layer. For a trillion-dollar entity like Alphabet, absorbing a mid-cap defense and aerospace powerhouse at these levels is a minor financial exercise. But the strategic payoff is massive: it instantly deprives Microsoft and Amazon of a foundational data asset while locking down the off-planet hardware architecture before the orbital launch supply chain gets completely constrained. The market is currently treating PL like a volatile, speculative space play, missing the reality that it is fast becoming core utility infrastructure for the AI era. When the broader market finally recognizes that the infrastructure of the next decade isn’t just buried in concrete on Earth, but orbiting overhead, the re-pricing will be swift. I prefer to see the pattern before the rest of the room wakes up to it.
