AI’s New Orbital Silicon Valley

Less than a week from today, one of our major space picks will launch its next satellite.

This won’t just be the company’s first next-generation satellite; it’ll be the company’s biggest leap yet toward building the world’s first space-based communications network.

With an array 10 times the data capacity currently seen in orbit, it’s the dawn of a new age.

Proving that it can deliver the kind of high-bandwidth, high-power orbital infrastructure that AI companies, carriers and government agencies are suddenly racing to deploy.

As we’ve seen this year, space is quickly becoming the next battleground in the AI buildout — for connectivity, for power, and for national security.

The Space-based Data Center Opportunity

While everyone’s focused on satellite communications, something new is forming.

Earlier this week, Elon Musk dropped hints about “Galaxy Mind” — a venture that would combine SpaceX’s launch capability, Tesla’s solar technology, and xAI’s artificial intelligence to deploy solar-powered AI datacenters in space.

He’s not alone.

Google CEO Sundar Pichai announced Project Suncatcher will launch pilot satellites in 2027 to test AI hardware in orbit.

Jeff Bezos is predicting gigawatt-scale datacenters in space.

Former Google CEO Eric Schmidt bought an entire rocket company — Relativity Space — specifically to pursue orbital datacenters.

And last month, Starcloud (backed by Y Combinator and NVIDIA) launched the first H100 GPU into orbit.

This isn’t speculation anymore. This is happening.

The Maths Makes Too Much Sense

Here’s the problem AI companies are facing right now: The largest models are already hitting the limits of terrestrial power grids.

By 2027, training the next generation of AI will require multi-gigawatt clusters that simply cannot be built on Earth’s existing infrastructure.

Space solved that immediately.

According to a Starcloud whitepaper, a 40 megawatt datacenter cluster costs $8.2 million to operate for 10 years in space versus $167 million on Earth.

Energy costs are about 20 times cheaper, or $0.002/kWh versus $0.04/kWh.

Solar in space has a 95% capacity factor versus 24% on Earth, generates 40% more peak power with no atmosphere to block sunlight, and provides free cooling using deep space at -270°C.

Plus, there are no permitting delays or water usage.

Space isn’t optional. It’s inevitable.

And that creates a massive need for large, resilient, power-dense, secure orbital infrastructure — the very capabilities ASTS is building right now.

The Golden Dome Question

Which brings us to something the market hasn’t connected yet:

The U.S. Space Force just awarded multiple contracts for the Golden Dome missile defense program. The contract winners? Classified. Awards under $9 million don’t require public disclosure.

We know some names — Northrop Grumman, Lockheed Martin, Anduril, True Anomaly. But we don’t know all of them.

Could others be on that list?

Here’s why the possibility matters:

The same technologies needed for orbital AI datacenters are the same technologies needed for orbital defense systems.

Golden Dome is the government’s near-term push into the exact orbital infrastructure that AI companies will need long-term.

And our pick is one of the few companies already building the hardware at that scale.

Consider this: they already secured a multi-million Space Development Agency contract and prime contractor status. Management has explicitly said they’re “well positioned” for Golden Dome.

Production contracts could be worth $1.8 billion to $3.4 billion annually.

I’m not saying they have the Golden Dome contract. I’m saying if they do, the market hasn’t priced it in yet. And at current levels, that’s a hell of a setup.

The stock doubled from our earlier recommendation, then back 40% from the October highs, and now sits at an enviable position.

The next launch could be the catalyst that reminds everyone why this company matters.

The Next Space-Sized Opportunity

Space datacenters represent the next wave. But here’s the challenge for retail investors:

Most of the pure plays are private. Starcloud is backed by Y Combinator and Nvidia but hasn’t gone public. Galaxy Mind doesn’t exist as a separate entity yet. Project Suncatcher is a Google division – a tiny part of a huge company.

We’re watching this space closely. When investment opportunities emerge — whether through IPOs, SPACs, or publicly traded component suppliers — we’ll flag them immediately.

For now, our pick remains the best way to play the broader space infrastructure buildout. The company is executing, the technology is validated, and the December launch puts the next catalyst on the calendar.

The stock might be volatile. The sector might be speculative. But the direction is clear.

Space infrastructure isn’t coming. It’s here.

And the companies building it are just getting started.