Jun Li

So hopefully good things will come out of it and maybe they'll realize that there needs to be a major shift in their education system.

Okay, and then moving to the last news story of the day is Google unveils Project Suncatcher, gotta love these project names, to explore AI compute in space.

Oh, yes.

Google announced Project Suncatcher, a research moonshot aimed at scaling machine learning compute off planet using solar powered satellites equipped with TPU chips.

A moonshot is a high risk, long horizon R&D.

The company published a preprint and said it will partner with Planet.com or Planet on a learning mission that plans to launch two prototype satellites by early 2027.

So if you didn't know, Planet is...

A satellite company.

They essentially design satellites and deployment for geospatial work.

So a lot of the information that you're getting, you're probably getting out of one of their satellites or their new satellites, I should say.

The concept ships from beaming space power down to Earth towards running AI workloads directly in orbit, where sunlight is consistent and land and water use on Earth are zero.

The blog says early work covers constellation design, satellite control, communications, and radiation testings of TPUs.

If it works, some training or preprocessing could move to low earth orbit, LEO, defined as altitudes roughly 160 to 2000 kilometers, while latency sensitive tasks stay on the ground.

The differentiator here is putting proven AI accelerators into a tight satellite network rather than a single giant platform.

So some of the questions here, economics versus cheaper terrestrial power, orbital debris and traffic management, data downlink bottlenecks, and security rules for off-planet compute.

Independent coverage frames this as an ambitious bet that leans on falling launch costs and maturing commercial space supply chains.

So super exciting, right?

Yes.

Yeah, exactly.

Yep.