We would think that computing is exploding.
On a certain level it is. When we look at the world of AI, specifically GPUs, we see a massive explosion. This is being led by Wall Street's most recent darling, NVIDIA.
This, however, is no reflective of the entire computing spectrum. There is one area where we are seeing a fall off from the norm.
It is the world of supercomputing and this means an exciting future.
How is it an exciting future if supercomputing is falling behind?
To answer this question, perhaps we should look at what is taking place. We can see in graphical terms how things appear. It also presents the opportunity in the future.
Here is the chart from the Top500 website:
We can clearly see the trend lines for all three categories is flattening. This is most noticeable with the #500 computer. It is then followed by the top computer.
The number 500 computer falling off could be explained away. This could be due to the development of cloud computing which is providing alternatives to companies. This means the need for so many supercomputers could be in decline.
That said, it does not cover why the top one is slowing down against the trend or the average. Here is where the immediate concern enters.
Of course, this is only near term.
What we have is the situation where reversion to the norm takes place. With technology, especially something as long established as computing trends, we see these periods. In fact, there was a period of about 10 years where the top computer was ahead of trend.
The same was true for other trend lines.
When the situation is resolved. we see a powerful move.
History shows that, at some point, perhaps towards the end of the decade, we will see a major shift. Perhaps this could be done, in part, by OpenAi and Microsoft's desire to built a $100 billion supercomputer.
If that eventually rolls out, it will be a powerful device.
Supercomputing is important because it is cutting edge technology. Absent quantum, which is presently lacks utility, these can solve problems no other devices can. It is why they were always the go to for Big Pharma and related companies. Simulations could be run that save months (if not year) of traditional research. This could save hundreds of millions of dollars in drug discovery.
With that type of financial impact, it is worth $50K an hour to rent these computers. They traditionally were funded by governments, making them part of the "public" domain. In the United States, they tended to be funded from either the Department of Energy or Defense.
In other words, it takes a lot of money to build them.
This could be changing. Since many of the leading technology companies are also involved in chip development, this means it could create a processor that takes things to the next level. If nothing else, a revolutionary line of Nvidia chips, as an example, could provide the thrust to revert back to the trend line.
As always, the tentacles of technology are not linear.
If we do see a period where new computers, with even greater compute come online, it will affect many areas. Couple this with different software architecture, along with advancements in algorithms and reduced energy consumption, and we can see how problems that cannot be addressed today suddenly become possible.
What happens if there is a breakthrough in the material sciences due to something being run on one of these workhorses?
There is no telling the impact that could have on construction, automotive, aeronautical, or even rocketry.
For now we see how things are slowing down. Hopefully, in a year or two, we will see the trend start to move upward. A lot of money in flowing into different aspects of computing, providing the opportunities for advancement.
