Devashish Paul: From Semiconductors and Supercomputing to AI, EVs and the Energy Transition, Part 1
Updated: Aug 28
I spoke with Devashish Paul, CEO and Founder of BluWave-ai, about the path that led him from mission-critical aerospace systems and the global semiconductor industry into artificial intelligence and clean energy. In Part 1 of our conversation, we discussed the early expectations around transportation electrification, the realities of scaling electric fleets, BluWave-ai’s evolution toward connected consumer EVs and grid applications, the growing role of energy storage, and the experiences that ultimately led Dev to build the company. We will continue the conversation in Part 2 with a deeper focus on AI in transportation, policy, regulation and BluWave-ai’s current projects.
How did the early market for transportation electrification compare with what actually happened?
A lot of municipal transportation electrification, and even the UPS, FedEx and DHLs of the world, has been very dependent on the regulatory environment, incentives and other external factors.
The uptake of municipal transit has been slower than people expected. The same thing happened with various OEMs that were emerging to build last-mile electric transportation. You had companies such as Arrival, Canoo and Proterra, along with established manufacturers such as New Flyer. A lot of companies were raising money at very high valuations in 2021 and 2022 based on the assumption that the next valuation would be higher and that volumes would arrive quickly.
Then the capital markets changed and a number of those expected vehicle deployments did not happen. We had a contract with UPS to work on electrification in Utrecht, Düsseldorf, Cologne and Dortmund. UPS had plans for a very large deployment of electric vans in Europe, but those programs changed as the vehicle market changed.
BrightDrop also entered the electric delivery-van market, and you saw significant investment in manufacturing capacity around those types of vehicles.
There was a compounding set of circumstances. Vehicle manufacturers were trying to scale, fleets were trying to transition, infrastructure had to be built, and the expected deployment volumes did not always arrive on the timelines everybody had originally assumed.
What do you think fleets underestimated about the transition from internal-combustion vehicles to electric vehicles?
From a distance, my observation is that a lot of organizations were trying to shoehorn electrified transportation into an ICE business process.
When I bought my wife a Hyundai Kona, she had to change the way she thought about using the vehicle. It is not like going to the gas station once every two weeks, getting 500 kilometres of range and forgetting about it.
I intentionally did not install a Level 2 charger at home initially because I wanted to understand what it was actually like to live with Level 1 charging. You have to be more proactive. If you are going to the cottage for the weekend, you cannot arrive on Friday afternoon with a 25% state of charge and assume you can quickly stop at the gas station. You have to think ahead.
Individual people who buy EVs change their family business process. They accept the performance characteristics of the EV and organize themselves around it.
I had endless discussions with managers at UPS depots, and a lot of the conversation was about fitting electrification into an existing process. That is viable in some situations and very difficult in others.
Forget for a moment about ordering the vehicles, installing chargers, pouring concrete and upgrading transformers. People are accustomed to more than a century of internal-combustion infrastructure simply being there. With electrified transportation, you have to organize and deploy an entirely new infrastructure around the vehicle. All of that takes time.
How did those market realities affect BluWave-ai’s strategy in transportation?
On the electrified transportation side, we reached the point where we said we can connect to Tesla, Kia, Porsche, Hyundai, Volkswagen, GM, Ford, Mini, Jaguar and others. Once we have completed an integration with a model and vehicle ecosystem, we can use that across vehicles in that geography.
That led us to invest more heavily in our EV Everywhere product line, which is about EVs connected to the grid at your home, my home, an office or elsewhere.
We can work with local distribution companies and system operators to aggregate EVs or provide services to the utility so that charging can be managed in constrained areas. Then you can figure out how to provide incentives so consumers participate. That is more aligned with our broader grid business.
Fleet depots can be relatively customized. A depot in Toronto can be different from one in Ottawa or Oakville. You can have different scheduling systems, different utilities and different grid operators depending on the jurisdiction.
With personal vehicles, the volume already exists on the street. We felt the unit economics could ultimately become much more scalable.
Why do you see consumer EVs as an important part of the energy system?
The volume of personal EVs is inherently going to be large.
There is also now a significant second-hand EV market developing. I recently bought a three-year-old Polestar 2 with about 55,000 kilometres for roughly $27,000. New, that was around a $60,000 vehicle.
You can also find older Nissan Leafs and other EVs at price points that are becoming accessible to a much wider group of consumers. Early adopters are upgrading their vehicles, and those cars are entering the used market.
That changes the economics. More people can buy an EV without buying a new premium vehicle.
There are also new manufacturers coming into the market. I recently met with the CEO of Geely North America, who is based in Mexico City, at an event focused on electrified transportation. There are a lot of things changing on the personal-vehicle side.
From BluWave-ai’s perspective, that means there will increasingly be a very large installed base of batteries connected to the electricity system.
How does that connect to the broader grid and energy-storage opportunity?
We also work in energy storage, solar and wind forecasting and optimization.
When I started BluWave-ai in 2017, I thought there would be battery storage everywhere on electricity grids within two or three years. Almost ten years later, the deployment has taken much longer than I expected.
But now construction is happening at scale. Ontario is on a path toward several gigawatts of storage being added to the IESO market. Texas and California are already much further ahead because they have had to integrate very large amounts of solar and other variable generation.
In Ontario, energy storage is becoming increasingly important as the electricity system goes through major changes, including nuclear refurbishment and growing demand.
Storage gives grid operators another tool for balancing supply and demand. You can charge batteries when energy is available and use that stored energy during peak periods.
For BluWave-ai, that sits naturally alongside what we are doing with EVs. Whether it is a stationary battery or an electric vehicle, you increasingly have distributed energy assets that can potentially be coordinated with the electricity system.
You have said that the EV itself may become the battery people put in their homes. What do you mean by that?
Years ago, I visited Hydro Ottawa and we were talking about batteries in people's homes.
I left that meeting thinking: I am motivated by the energy transition, and even I am not particularly motivated to install a dedicated battery in my house.
The battery that I am going to bring into my home is an EV.
Eventually that vehicle may have vehicle-to-grid capability. Even if it does not, it is still a significant storage asset connected to the electricity system.
That was one of the steps that ultimately led us toward thinking about EVs not only as transportation, but as part of the grid.
Before BluWave-ai, you worked in aerospace, semiconductors, telecommunications and high-performance computing. How did that career develop?
I started as an engineer working on the CF-18 mission computer in the early days, including during the period around the first Gulf War. I was not deployed. I was part of the engineering team working on the aircraft software.
That started me in mission-critical systems.
In the mid-1990s, the Canadian Armed Forces was closing bases and changing significantly, and I decided to move into the private sector.
I entered the semiconductor industry and eventually joined a publicly traded Ottawa company called Tundra Semiconductor. We built switching chips that connected CPUs, field-programmable gate arrays and digital signal processors inside wireless base stations.
We ended up winning a large part of the 3G market in China. Then we started building switching chips for 4G and competing globally with Integrated Device Technology in San Jose.
We won designs with companies including Ericsson, Nokia, Alcatel, Motorola, Panasonic, Fujitsu, NEC, Samsung, Huawei, ZTE and Datang.
Eventually Integrated Device Technology acquired Tundra. By around 2012, we had won a significant portion of the global 4G market.
At that point, we could already see 5G coming, but there was a long period before the next major deployment cycle. That was when I started working on chips for supercomputers, hyperscale cloud data centres and AI workloads.
What made you decide to take that technology background into clean energy?
I started thinking more seriously about how computing and data were being used.
A lot of the technology we were building was going into social media, hyperscale computing and eventually AI. There are obviously many positive applications of those technologies, but I started asking myself where else I could apply the same capabilities in a way that was more directly aligned with what I wanted to work on.
There was also an Ironman connection.
In 2006, I did the Ironman World Championship in Hawaii. They had recently installed a wind farm near Hawi at the far end of the bike course, so you were literally riding through the wind farm.
Later in the day, about 20 miles into the run, you go through the lava fields near the Natural Energy Laboratory of Hawaii Authority, where they were testing different forms of renewable and ocean energy.
I remember looking at the wind turbines, solar panels and all of this energy technology and thinking: these people are doing something useful for society. How do I take all of this computing technology I know and apply it to something like this?
Wind and solar fluctuate. They are variable and difficult for operators to predict. I started thinking: what if we could use data and AI to make those resources more predictable for electricity-system operators?
Then you add battery storage. Then you add EVs.
All of those little steps eventually led to where we are today.
Was there also a personal motivation behind founding BluWave-ai?
Yes. I wanted to focus the technology on an area that was aligned with my values.
Technology is rarely inherently good or bad. The same computing technologies can be used in communications infrastructure, defence systems, social media, artificial intelligence or many other applications.
I had reached a point where I wanted to work on something whose purpose was very clear to me.
Energy is fundamental. Everybody needs it. Making the electricity system cleaner, more efficient and better able to integrate renewable energy and electrified transportation was something I could commit myself to.
That was ultimately the motivation behind BluWave-ai: taking the computing, semiconductor and AI experience I had accumulated and applying it to the energy transition.
Part 2 of my conversation with Dev will focus on AI’s practical applications in transportation, the policy and regulatory environment, and the next phase of BluWave-ai’s work..
Comments