high-performance computing

And then do your deep neural network to train the LLM.

So the main architectural issue is how do you do that without it being a science experiment? And people have been leveraging traditional HPC storage systems that were used in government labs.

Lustre comes to mind.

It's 25 years old.

And so getting separate things over separation of metadata from data and being able to scale those independently, having parallel access to data using standard protocols.

Lots of people are inventing custom protocols and custom clients.

And while that's great for HPC in the enterprise and in AI factories, people need standard and The open systems have flourished with standards.

This is not any different.

And so that's first and foremost is if you look at hash rate, hash rate is back to almost all time highs.

which is fascinating because the Bitcoin miners, many of them started to transition to become HPC or AI data centers.

So you took a bunch of the large players and basically took them offline, but hash rate is coming back and almost back to all time highs.

My friend, Will Clemente pointed that out.

Well, I think that anytime you have change in a market, there's always volatility.

Volatility can work in both directions.

Now, one of the questions that people have had historically is, is AI sucking the air out of the room and capital is rotating towards AI and it's moving away from the Bitcoin market? Now, I definitely think that there's been a lot of capital rotation, specifically mentioned, you know, the miners moving to HPC, things like that.

in my own investing.

And so there's been a lot of volatility and, you know, obviously in the last month, but even before that.

And it's not unique to either the miners or the HPC companies or, you know, AI in general right now.

I think you have to be prepared for the fact that there's just like constant bouts of, you know, fear, greed, self-doubt as these things move along.

I mean, it's not that long ago that we all experienced the deep seek kind of, you know, craze, right? And then it gripped us and gripped the markets, huge sell-offs.

Exactly, yes.

Because today, if you look at the entire narrative of this year, people were in gold in Q4 last year, and then they moved to oil with the conflict, and now they're moving back to AI and HPC in Q3.

How do you keep up? And so it keeps moving all the time.

And if you look at the speed of rotation happening right now, it has never been that brutal.

Exactly, yes.

Because today, if you look at the entire narrative of this year, people were in gold in Q4 last year, and then they moved to oil with the conflict, and now they're moving back to AI and HPC in Q3.

How do you keep up? And so it keeps moving all the time.

And if you look at the speed of rotation happening right now, it has never been that brutal.

I came here in 2016 after an announcement for a replacement of the professorship in my field.

At that point I'd been working at Linköping University for a number of years and my engagement was partly in HPC, that we're going to talk about today to a large extent, but also in software development.

Within this,

through your days in Linköping and here, can you describe the Swedish HPC ecosystem in a nutshell?

Yeah, I've been involved in that ever since I started as a PhD student, more or less.

So that's now some 30 years.

And from my perspective in Linköping, I felt that KTH had more of an international profile and then they had an ambition to go more to large scale, massive parallel computing, whereas Linköping had more of a close user interaction based activities.

Then I moved to KTH and I found that very stimulating because then I could also take steps in that direction and make my research more going in the direction towards massive parallel computing.

So they were successful in that research project and then made the decision to invest in building a commercial infrastructure with this network.

So they built that out, continued to work with a quantum startup called Cubatech, and built the first commercially available quantum network that was launched roughly four years ago so all of a sudden chattanooga the mid-sized southern city is in the quantum space the piece of this i think that is really critical and needs to be noted is that this network is not a research tool they made it commercially available meaning you don't have to go to a national lab or a university in order to access this if you're a business that's trying to work on whether that's cyber security or secure communications or testing on how qubits are traveling across the network you can come to chattanooga and that's exactly what's been happening for the past four years last year in addition to the network epb partnered with ionq which is a trapped ion quantum computing company and they purchased the enterprise forte quantum computer So now we have in Chattanooga a commercially available quantum network co-located with a commercially available quantum computer and also co-located with HPC.

So a supercomputer really on loan from Oak Ridge National Lab.

So this environment exists in Chattanooga that again, the emphasis is on how do we turn this research into commercially viable infrastructure that supports the growth of that for companies and enterprises.