5Q PropTech Interview – Paul Benoit Founder at Qarnot Computing

5Q PropTech Interview – Paul Benoit Founder at Qarnot Computing

Why did you start Qarnot Computing and what is the company’s mission?

I started Qarnot Computing for 2 main reasons:

Firstly, in 2010, after a long experience with large computing farms in the banking industry, it looked obvious to me that computers should be embedded and distributed in buildings to retrieve the waste heat instead of being exclusively concentrated on inexpensive and energy-hungry data centers.

The experience led me to think about designing a heater embedding computers as a heat source.  

Secondly, I met Miroslav, our COO, who accepted to follow me! 

The mission has never changed: democratize HPC by making them greener and cheaper. 

Would you explain what is High-Performance Computing (HPC), as well as who is driving the demand for it?

Aerodynamics optimization, protein simulation, oil & gas exploration, financial risk analysis, weather forecast, climate change studies, or 3D animation… all these tasks have one thing in common: High-Performance Computing (or HPC).

It can be seen as all computational tasks that need important computing resources to be done in a reasonable amount of time.

HPC is strategic for competitiveness and innovation.

As an example, HPC is increasingly used for new drug testing to focus on costly field tests on the best candidates. It’s faster and cheaper.

How do you differentiate yourself from the dominating cloud players such as AWS, MS Azure, and GCP?

Since its inception, Qarnot is focused on HPC services while most cloud players provide generic IaaS.

HPC nodes are marginal in terms of occupied space in data centers, compared to storage servers, but their energy footprint is far more important.

With AI and simulation development, energy consumed by computation servers will grow exponentially.

Will 5G deployment assist in accelerating the climate-conscious edge computing approach that Qarnot Computing has adopted?

5G constitutes an opportunity for what we call edge HPC, which means HPC performed closer to the end-users producing data.

Edge HPC has many advantages in terms of latency, energy efficiency, and privacy.

With 5G, operators will need to provide more services with local computing infrastructures, e.g. for gaming or AI.

On the other hand, latency and bandwidth offered by 5G will be an opportunity to deploy Qarnot infrastructure where fiber optics is not available.

What are the main challenges of building and operating both hardware products and software platforms?

There’s a famous quote among tech-savvies saying that people who are really serious about software should make their own hardware.

It’s a bit radical and we used to say that Qarnot is mostly a software company; however, designing our own hardware is a great opportunity for our teams to develop a strong and concrete experience in product development.

Today, we have reached sufficient maturity to develop partnerships with appliance manufacturers for potential co-designs.

Press enter or esc to cancel