The convergence of technology can be very powerful. When advancements in one area affect others, we see exponential changes. Therefore, while progress is happening in parallel, it often converges to multiply.
Probably the best example was the iPhone. This was the convergence of compute, the Internet, and mobile technology. There is little wonder why it become popular in such a short period of time.
With communications, we see similar impacts. When progress is made in routers, this affects everything. Of course, bandwidth expansion provides like results. Battery efficiency leads to longer up time. Basically, each advancement assists in the expansion of the totality.
AI is getting a lot of attention. Coupled with this is the energy question. Could we be looking at Space AI?
Datacenters are the major conversation piece. We see tens of billions being poured into this segment. If some forecasts are correct, this could eventually enter the trillions.
The quest for energy is on. Major corporations such as Meta and Google are cutting deals to power their data centers. This includes building on (or near) a nuke plant. Everyone knows the thirst for energy is only increasing, making alternatives necessary.
Here is where the idea of space comes in. Many feel that solar arrays eventually reaching deep space could be a solution.
The challenge is that fact that sending the energy back to Earth is dangerous and would have to be limited. Only a small portion of the sun's energy hits the planet. If too much is sent back, it could fry everything.
As a solution, instead of sending the energy back, why not build Space AI. This is a concept that Elon Musk is already thinking about with Tesla's chips. Why not power the data centers in space and have the data fed back to Earth?
If AI related services are going to consume massive amounts of energy, why not have it done in Space? This removes the power consumption for the planet.
Here is a breakdown of the costs taken from the nextbigfuture website:
As we can see, there are cost benefits to looking in this direction.
So why not do it?
Industries running in parallel comes into play here.
At the moment, it is not feasible to set up Space AI. The first barrier is rockets. We simply cannot send up arrays large enough to power these datacenters. It is a situation that will change with SpaceX Starship.
Of course, we could add another industry into the discussion. Advancements in solar conversion and cell materials could provide efficiency gains. This could mean harnessing more energy with less surface space on the solar panels.
Then we have the AI chips themselves. Right now, they are designed for Earth. They operate in controlled environments while being in close proximity. When dealing with space, a host of other factors are introduced.
Tesla is already thinking about this with its AI 7 and 8 designed. They are presenting working on AI5, a chip that will be introduced towards the end of next year, with production coming in 2027.
A lot of this makes sense if we are talking about interplanetary operations. There are proposals out for quests to both the Moon and Mars. Many floated the idea of space hotels along with manufacturing. Over the next 20 years, entire industries will crop up.
Having AI operating off-planet only makes sense. There is no need to focus as much on Earth if people are starting to spend time in space.
Entire new industries are about to open up. Convergence is going to be greater in the future. The further down the path we head, the more powerful the uniting of the technologies is.
