Bonjour tout le monde,
This is a special article featuring a fascinating Canadian technology company, Brainbox AI. The company is at the forefront of the climate tech revolution and trying to solve one of the biggest contributors to greenhouse gas emissions, the built environment. I had the chance to speak with their CEO, Sam Ramadori and CFO, Francis Trudeau who were gracious enough to provide a behind the scenes look at what they’re building. I also interviewed Sam on my podcast.
I’ve been a fan of the company for a while and excited with what they are working on. Selfishly, I wanted to have the chance to learn more about them and shine a light on their work.
In this post, I want to highlight the global problem they are trying to solve, how they are applying deep learning in an innovative way and the tantalizing possibilities of making a big impact.
Let’s dive in
The world is burning
This past summer alone, many places in the world experienced unprecedented droughts and natural disasters. The Poe river in Italy, the Rhine river in Germany, Yangtze river in China and Lake Mead in the US all endured historic droughts. It’s not just droughts, let’s not forget the apocalyptic floods in Pakistan this summer as well.
In isolation, each country deals with natural disasters from time to time. When looked at together, the picture is starker. Natural disasters are happening at a faster rate than before and human caused emissions are undoubtedly a contributing factor.
When we zone in on the sources of these emissions, there are five inescapable areas we need to tackle in the fight against climate change: energy, large industrials, transportation, food & agriculture and the built environment. The built environment represents all the buildings and homes we live & work in and is alone responsible for 38% of carbon emissions. Of that, around 12% is the heating & cooling in buildings. This is the segment that BrainBox is going after.
This market has several systemic challenges:
- The market itself is extremely fragmented. There are countless stakeholders in the ecosystem.
- Emissions from the built environment are also a huge externality that no one is paying attention to. Building owners or tenants don’t care and the cost isn’t passed on to anyone.
- The built environment is slow in nature. All the buildings around us will be here long after we’re all gone.
- There isn’t any real government regulation enforcing change.
Who in their right minds would enter such a difficult market?
From BrainBox’s inception, the goal was to tackle this intractable problem. Jean-Simon Venne is the CTO and co-founded the company in 2017. The vision was to build software that would help buildings autonomously adapt and learn, just like the potential of self-driving cars.
How it works
Buildings have HVAC systems (heating, ventilation and air conditioning) which connect to hundreds of sensors. These control systems monitor and adjust the temperature of the building. The sensors capture a bunch of internal data but nothing from the outside world since they are not cloud connected. This internal data is collected each day, used and then thrown away. The whole system simply resets every morning.
BrainBox developed a technology that layers on top of HVAC systems with their software and helps make these systems smarter. The software brings in external data to improve the learning cycle and make adjustments in real-time. The software will predict that the room will get cold in a few hours and starts adjusting the temperature of the room in advance.
It may sound like simple stuff until you understand how these systems work.
Today, all HVAC systems only react when the room temperature is already too cold. For example, when a room goes from 22 to 21 degrees, it will react and start increasing the temperature. No one will wait around for the heat to gradually increase so the system needs to react quickly. It’s like driving a car at 100kph, then stopping, then accelerating quickly again. It’s terribly inefficient and uses a lot of energy.
BrainBox’s software is able to predict in advance when it will be too cold and starts adjusting in advance at a much more gradual pace. This allows a step change in how energy is deployed in a building. BrainBox uses data from the outside world in order to achieve this. Weather data is one obvious data point. Another one is the type of energy available on the grid – renewable or not.
BrainBox will install a simple edge device that connects the HVAC to the cloud and provides an interface to the person sitting in the control room. The interface is not just a dashboard but also provides the controller the important capability to turn the autonomous system on or off – just like the idea of the self-driving car. All the data is moved to the cloud, properly categorized which then generates meaningful insights and most importantly, actions.
It’s a straightforward SaaS business model with low upfront cost and promises to be cash positive in less than a year for building owners and property managers. As with any startup there is technology risk and market risk. They’ve proven their platform works with deployments in over 400 buildings. The key issue to scale is market risk in a giant slow moving industry.
Going to market
HVAC control systems come in all sorts of different ages and quality of commissioning. This adds a degree of complexity in their go to market strategy since the BrainBox team needs to do an assessment for each deployment. It’s been a challenge right from the launch of their product. Pitching skeptical property managers with the promise of saving money & doing good for the planet at the same time.
Oh but how the times are a’changing. They are benefiting from exceptional tailwinds. The markets are moving away from a solely return on investment mindset. As noted in the dire opening paragraph, the world is burning and people are starting to pay attention. Their sales conversations used to be about payback periods and dollars saved – this isn’t the case anymore.
Governments, banks and boards are helping with this shift and greenhouse gas emissions are no longer an unfortunate externality. People want to improve the carbon footprint of their buildings and need to start showing tangible results. Their go to market has been radically accelerated and sales conversations are moving across all types of property; from shopping centers to small retail to large offices. The need is clear and present.
Building an impact culture
The company’s motto is to save the planet with AI and one of their values is solving big ass problems. They are definitely not taking an easy route by tackling this thorny problem. Sam and I spoke at length about their team and how he thinks about their culture. One key principle he tries to instill in the company is allowing people the freedom to make mistakes. There isn’t a blueprint for what they are building here and errors are unavoidable. This gives the team agency to push the envelope and execute at a high velocity. For instance, they went global far faster in their life cycle than one would expect.
Despite being a scrappy startup, the talent density is remarkably high. The people working there can all go work at some big tech company for more stability & higher wages. Yet, they are choosing to work on a project that has true global scale and impact.
Making energy use smart
BrainBox’s technology is currently deployed in over 400 buildings, in over 18 countries, saving thousands of metric tons C02 every year.
The biggest shift this kind of technology represents is flexibility. It allows building owners to understand their assets at a much more granular level – even down to the individual fan. Energy usage in buildings is moving from a blunt instrument to a scalpel. Not only will equipment be used less, energy will be reduced during peak periods in turn reducing the strain on utility providers.
The shift to renewable energy is also a key component. One important source of external data is knowing how carbon intensive the electricity received in the building is. BrainBox’s algorithm monitors energy use throughout the day and shifts to green energy over fossil fuels when possible.
The entire economy is inexorably moving from the industrial era to the connected era. It’s just a question of when, not if, our buildings and energy grid get smart.
This is such an obvious use case of what software can do. It starts off by measuring something that we couldn’t before and by doing so, new paradigms emerge. The next phase of what BrainBox is working on is one of those mind blowing possibilities.
The future of distributed energy
Demand for EVs is growing exponentially and this is going to impose a massive strain on buildings which are expected to provide charging stations. Buildings will need all the flexibility possible over and above the transition to renewables. Do we need to install batteries & powerwalls in all buildings? Not necessarily.
Buildings absorb energy to turn into heat through a charge & discharge process. By forecasting the HVAC load in advance, BrainBox is able to pre-emptively store thermal energy when a relevant event is predicted on the horizon. As noted in the previous example, in summer the building can be pre-cooled to the lower comfort limit before a high-demand event, thus using the equipment less.
BrainBox uses this characteristic to provide more flexibility to the grid. It can store thermal energy and then avoid using electricity later therefore freeing up electrons that it would have normally demanded from the grid. Two-way communication with the grid becomes super important. A utility provider can stop sending energy during peak periods and let the building float on its own energy. It’s a super green form of energy and the exact same concept of a thermal battery. No lithium batteries required.
This concept of distributed energy resources (DERs) is an emergent phenomenon with huge potential. You might have heard of the US infrastructure bill which focuses specifically on DERs. This means that commercial buildings will become participants in the energy grid and a core part of scaling renewable energy. Energy consumption and storage will be optimized, efficient and green. This unlocks a new addressable market for BrainBox which is poised to take advantage of.
The road ahead remains volatile amid a slow moving industry. Will BrainBox succeed in overcoming the massive challenges? I sure hope so. Our planet is counting on it.