Google Just Put AI Chips in Orbit. These 5 Stocks Could Build What Comes Next.VIEW IN BROWSER | FOLLOW LUKE ON 𝕏 
Imagine waking up in 2090 to a thin blue sunrise beyond your bedroom window. You’re several hundred miles above Earth, starting a six-month assignment at a commercial research station. Down the hall, scientists are developing medicines in microgravity. A cargo ship outside your window is preparing for the next leg of its journey to a lunar outpost. Before breakfast, you ask your AI assistant to review the overnight experiments and adjust the day’s schedule. You barely think about where the computing happens. But somewhere beyond the station, a cluster of satellites is processing those requests, drawing electricity from sunlight and exchanging data through laser links. An orbital data center. As ordinary to you as a server farm in Virginia was to your grandparents. That’s one possible future. But getting there would require cheaper launches, reliable hardware and businesses that can earn a return on the enormous expense of operating in space. For investors today, though, there’s a more immediate question: Which companies could get paid to build the first pieces of that future? We’re about to find out. On Oct. 1, Alphabet (GOOGL) put its first Project Suncatcher prototype into orbit. Developed with Planet Labs (PL), the satellite will test how Google’s Tensor Processing Units (the custom chips it uses for AI) perform in orbit. The mission has a modest but essential assignment: Find out how that hardware handles the journey and the conditions waiting above Earth. Its results could help determine what engineers build next, and which suppliers they need to build it. That’s why I’m watching orbital data centers. Long before computing in space becomes routine, the effort to make it work could create demand for solar power systems, specialized electronics and communications equipment. And several publicly traded companies already make the kinds of technology those systems will need. That supply chain is also a key part of my XPANSE thesis: how the convergence of AI, energy, robotics and space could create opportunities for the businesses supplying the buildout. Later, I’ll show you where to find my full XPANSE briefing, including a free investment recommendation. But first, let’s look at what Google’s experiment could mean for the companies helping take AI into orbit. Why Google Thinks AI Compute Could Move Into SpaceBack in April, we explored the case for moving AI infrastructure into orbit. The underlying idea remains compelling: If power availability limits where we can build computing capacity, finding a new source of power could expand the opportunity. In the right orbit, Google estimates that solar panels can generate up to eight times as much energy as comparable panels on Earth, with nearly continuous sunlight reducing the need for batteries. Project Suncatcher would use that electricity where it is generated. Google envisions satellites carrying AI processors and communicating through optical links, eventually functioning as a distributed computing system. Its original research announcement describes the engineering work needed to make that possible. The appeal of this is easy enough to understand, but a computing facility that generates its own electricity in orbit would have a different set of constraints from one waiting for a connection to the terrestrial grid. And those different constraints could create a different set of winners. Think about the hardware required: efficient solar cells, deployable arrays, reliable electronics, communications equipment and systems that can keep working without a technician walking down the aisle. Every one of those requirements represents a problem somebody needs to solve, and potentially a product somebody needs to buy. What Project Suncatcher Has to Prove in OrbitGoogle’s latest update identifies the immediate priorities: surviving launch, operating through radiation exposure and managing heat. Ground testing has been encouraging, but an orbital mission will expose the hardware to conditions a laboratory cannot fully reproduce. Cooling deserves particular attention. There is no airflow in a vacuum to carry heat away; Google is testing heat pipes and radiators to move heat out of its chips and reject it into space. Then comes the networking test. Google plans to put two satellites in orbit in 2027 to evaluate the laser connections needed for distributed AI workloads. Those are separate milestones. First, establish how the hardware operates in space. Then test whether multiple spacecraft can work together effectively. For investors, each stage should sharpen the questions we ask. Which components perform reliably? What needs redesigning? Which suppliers win follow-on work? And how much does the next version cost? That’s how a promising concept starts becoming an investable business. The results need to support the next round of spending.
Google Isn’t the Only Company Exploring Orbital AIGoogle’s effort also sits alongside another significant development since our early-August issue. On Aug. 24, Nvidia (NVDA) confirmed that SpaceXAI plans to extend its computing platform into orbit. The planned first-generation Starmind AI satellite would use an optimized version of the Vera Rubin NVL72 system. That supports an important part of our original thesis: established AI computing platforms could provide the foundation for orbital systems. Nvidia and SpaceXAI are working to preserve a common architecture and software ecosystem while adapting the hardware for space. Nvidia explicitly identifies power, thermal management, bandwidth, reliability and physical integration as constraints that differ dramatically from terrestrial data centers. An Earth-based rack cannot simply be loaded onto a rocket and declared ready for business. What interests me is the parallel investment in solving those problems. Google is exploring an approach built around its TPUs. SpaceXAI is pursuing one built around Nvidia’s platform. That gives the broader orbital-computing thesis more than one path forward. It also suggests that investors should study suppliers whose products could be useful across different designs. 5 Space Stocks With Exposure to Orbital Data CentersThe latest developments sharpen our orbital data center watchlist: Planet Labs (PL) has the clearest connection to the Suncatcher milestone because it is Google’s named mission partner. That gives investors a specific program to keep tabs on. I would watch for evidence that the relationship expands into additional development work or commercial contracts; participation in a prototype alone does not establish a material earnings stream. Rocket Lab (RKLB) deserves attention for more than launch services. In February, it introduced silicon solar arrays aimed at applications including space-based data centers. And on Sept. 8, it announced production availability of its IMM Apex solar cell, touting lower cell mass and a design that eliminates conventional germanium substrates. Those developments address two practical considerations for space infrastructure: weight and supply-chain availability. Redwire (RDW) offers another way to follow the power-system opportunity. Its Roll-Out Solar Array technology has applications including the International Space Station. I see that experience as relevant to a potential orbital-computing buildout, though relevance still needs to translate into customer orders. Microchip Technology (MCHP) supplies radiation-tolerant programmable chips for space applications. These supporting electronics deserve attention alongside the processors doing the AI calculations. The investment question is whether new programs produce enough incremental demand to matter to the broader business. Nvidia (NVDA) remains a direct participant through the announced SpaceXAI development effort. Orbital computing could eventually expand its addressable market, but I would treat that as a long-term possibility when evaluating the stock, rather than assume a large near-term revenue contribution. These companies have different exposures, valuations and execution risks. A good orbital-computing story is a starting point for research; orders, margins and cash flow will determine its value to shareholders. The Bottom Line: Orbital Data Centers Are Starting to Create a Supply ChainWhat excites me about Suncatcher is how clearly it illustrates the next investment question in AI: What has to be built to make much more computing possible? Answering that question takes us beyond the model itself and into energy, advanced chips, manufacturing and space infrastructure. That convergence is central to the broader investment thesis I call XPANSE. In my briefing, I explore how Elon Musk’s businesses could become more closely integrated around AI, robotics, energy and space. I also explain why some of the most interesting opportunities may be among the smaller businesses supplying the technology and materials those ambitions require. More companies exploring orbital computing could create opportunities beyond any single founder or corporate deal. The challenge for investors is figuring out which businesses can turn that interest into durable growth. That’s where my XPANSE briefing comes in. I walk through the three investment moves I’m recommending and reveal the name and ticker of my top investment for this theme… for free. Watch the XPANSE presentation and get my free recommendation here. The grid is going up whether you’re positioned or not. Which side of the trajectory will you be on? Sincerely, |
