The next big investing edge in AI may come from identifying where capital for orbital computing will flow first, says Matthew Tuttle of Tuttle Capital.
Not long ago, any deal tied to OpenAI could spark a rally in tech stocks. But sentiment has clearly shifted, with investors now far more cautious about bold AI promises. Adding to the unease, the sector is grappling with the classic “front-cover curse” after Time Magazine named several AI leaders as its Persons of the Year—including Nvidia CEO Jensen Huang, who also just earned Financial Times Person of the Year honors.
What the AI story needs now is a bold new chapter. Enter: space.
According to The Wall Street Journal, Jeff Bezos and Elon Musk are exploring the creation of orbital data centers designed to power Earth’s growing artificial intelligence needs.
In a new blog post, Tuttle outlines how investors might position for this emerging frontier. One of AI’s biggest bottlenecks, he argues, isn’t chips—it’s power. “Data centers and AI are projected to drive a massive increase in electricity demand this decade, which is why ‘grid tech’ has become a key market theme,” he says.
Space-based data centers, however, can tap continuous solar energy. “That’s where ‘compute in orbit’ becomes a serious idea,” Tuttle notes. “AI generates unusually high revenue per kilowatt, and solar in space is always ‘on,’ unlike Earth-based constraints.”
Still, scaling orbital computing and sending that power back to Earth remains the major challenge. The real investment opportunity, Tuttle says, lies in identifying where capex will flow first—toward edge inference and space infrastructure—while recognizing that full gigawatt-scale space-based power remains a long-term sci-fi vision.
Tuttle outlines a three-phase roadmap for investable opportunities:
Phase 1: Happening Now — On-Orbit Inference
This includes existing satellite tech used for Earth observation, weather, communications, maritime tracking, and missile warning.
“If you do inference onboard—classify, compress, decide—you downlink answers instead of terabytes,” Tuttle says. “Less bandwidth, faster responses, fewer ground stations. That’s where today’s ‘sky-brains’ fit.”
Phase 2: Purpose-Built AI Satellites
Next comes satellites with dedicated compute capacity.
“You’re not just processing your own sensor data—you’re renting compute,” Tuttle explains. Challenges include radiation tolerance, launch cadence, and inter-satellite networking. This phase works best in niche workloads where latency, sovereignty, or resilience matter more than pure cost.
Phase 3: Space-Based Solar Power Beaming
The ultimate, long-term goal: beaming solar energy from space to Earth. This requires enormous lightweight structures, safe energy-beaming systems, and ground-based grid integration—making it the most engineering-intensive phase.
For the first two phases, Tuttle highlights space-infrastructure winners such as Redwire, Rocket Lab, L3Harris Technologies, RTX, Northrop Grumman, and Lockheed Martin, which supply the spacecraft, communications systems, and radiation-hardened components needed for orbital compute.
Radiation-tolerant chips will be another key enabler. Tuttle points to Microchip Technology as a direct play on space-grade silicon and control electronics.
But even if compute shifts to orbit, most AI demand will remain Earth-bound in the near term. That’s why the real spending boom is still in power infrastructure—transformers, cooling systems, and distribution networks. Tuttle recommends watching Eaton, Hubbell, Quanta Services, and Vertiv for grid-buildout exposure.
When asked whether investors can access this theme through an ETF, Tuttle—known for launching some of the market’s more unconventional funds—gave a cheeky reply: “Not until I launch it.”
