
6.22.22-Showcase-Xinterra

-
Video details
AI-Powered Sustainable Materials and Formulations Development
-
Interactive transcript
PATRICK TEYSSONNEYRE: Hi, everyone. My name is Patrick Teyssonneyre. I'm the CEO and co-founder of Xinterra and also, as well, fellow class of 2019. Two of my co-founders are also MIT alumni. And Professor Tonio Buonassisi from the mechanical engineering department is also a co-founder of Xinterra. At Xinterra, we are on a mission to fundamentally change the development of sustainable materials to meet the urgent challenge of the climate crisis that were extensively explored here today.
So the current materials R&D process is very slow, expensive, and limited. And also, all of the materials that we use in our day to day, they work, they do their job. However, they are suboptimal in terms of performance costs and, unfortunately, suboptimal in terms of sustainability, as well.
To solve this problem, we created our platform, named Xinterra Design Factory, or XDF. We use a unique combination of artificial intelligence systems, high throughput experimentation tools, and deep material science expertise.
In this case here, this customer, they spent seven months just trying to develop a thermal insulating coating. They tried 69 formulations. None of them worked. So they hired us to help them to overcome this challenge. And the additional challenge here, on average, they use around 15 ingredients in these coating formulations. They wanted to reduce that. Besides the improvement of performance, they wanted to eliminate certain ingredients that were not environmentally sustainable.
So we trained our machine learning with their data. Actually, we used only 49 out of the 69. Then you see here, once we trained our machine learning, it recommended them to explore a totally different space from the original one that they were exploring.
As you can see here, this was the initial space. They were trying, trying trying, here. That's why they were not being able to find a solution. This happens all the time. It's a pattern on the materials R&D process. And once the machine understood what was happening there, so guided them to explore this other space.
And they executed the experiments. They prepared this formulation in their lab. They make the measurements. And we were able to find four formulations that by far exceeded their expectations in terms of this thermal insulating platform and also the thermal insulating properties and the other properties that we need to keep or improve.
One of the biggest differential of our company, of Xinterra-- so we are in the material informatics space. Probably you heard about other companies in this space-- the biggest, differential is this component, the hardware component of our platform.
So we have a strong expertise in the team to design, develop, and fabricate high throughput experimentation tools and to run proxy measurements. We do we do this to generate data fast, cheap, to train our machine learning, and come up with a much more accurate prediction and reduce even further the materials R&D development times.
Here are examples, so tools that we adapted from the biotech industry, like an open [INAUDIBLE] mixer. We have a rig here to measure the Hansen Solubility Parameters for paints and coatings and beauty and personal care industry, accelerated thermal conductivity measurement.
And we are using our platform to-- we are starting to develop our own materials IP portfolio for that direct air capture, so basically materials that can absorb CO2 from the air. And this is a rig that we are developing, we are building, to be able to measure and to screen and to generate thousands and thousands of results of different materials, training our machine learning, and start developing our own materials.
So we are targeting companies in each one of these five fields here. Because we want to advance with at least this for our United Nations Sustainable Development Goals. And on doing that, one of the contributions that we want to leave is a more environmentally sustainable planet.
In terms of business model, we named our model materials as a service. We provide service to company. We co-create materials with companies through joint development agreements. And as I mentioned, we are starting to develop our own materials IP portfolio to license in the future very focused on direct air capture.
We are looking for commercial pilots globally. We are already serving companies in the pulp and paper space, polymers, additives for chemical industry, paints and coatings. And we'll partner with you to accelerate innovation and especially to improve your sustainability footprint. Thank you.
[APPLAUSE]
-
Video details
AI-Powered Sustainable Materials and Formulations Development
-
Interactive transcript
PATRICK TEYSSONNEYRE: Hi, everyone. My name is Patrick Teyssonneyre. I'm the CEO and co-founder of Xinterra and also, as well, fellow class of 2019. Two of my co-founders are also MIT alumni. And Professor Tonio Buonassisi from the mechanical engineering department is also a co-founder of Xinterra. At Xinterra, we are on a mission to fundamentally change the development of sustainable materials to meet the urgent challenge of the climate crisis that were extensively explored here today.
So the current materials R&D process is very slow, expensive, and limited. And also, all of the materials that we use in our day to day, they work, they do their job. However, they are suboptimal in terms of performance costs and, unfortunately, suboptimal in terms of sustainability, as well.
To solve this problem, we created our platform, named Xinterra Design Factory, or XDF. We use a unique combination of artificial intelligence systems, high throughput experimentation tools, and deep material science expertise.
In this case here, this customer, they spent seven months just trying to develop a thermal insulating coating. They tried 69 formulations. None of them worked. So they hired us to help them to overcome this challenge. And the additional challenge here, on average, they use around 15 ingredients in these coating formulations. They wanted to reduce that. Besides the improvement of performance, they wanted to eliminate certain ingredients that were not environmentally sustainable.
So we trained our machine learning with their data. Actually, we used only 49 out of the 69. Then you see here, once we trained our machine learning, it recommended them to explore a totally different space from the original one that they were exploring.
As you can see here, this was the initial space. They were trying, trying trying, here. That's why they were not being able to find a solution. This happens all the time. It's a pattern on the materials R&D process. And once the machine understood what was happening there, so guided them to explore this other space.
And they executed the experiments. They prepared this formulation in their lab. They make the measurements. And we were able to find four formulations that by far exceeded their expectations in terms of this thermal insulating platform and also the thermal insulating properties and the other properties that we need to keep or improve.
One of the biggest differential of our company, of Xinterra-- so we are in the material informatics space. Probably you heard about other companies in this space-- the biggest, differential is this component, the hardware component of our platform.
So we have a strong expertise in the team to design, develop, and fabricate high throughput experimentation tools and to run proxy measurements. We do we do this to generate data fast, cheap, to train our machine learning, and come up with a much more accurate prediction and reduce even further the materials R&D development times.
Here are examples, so tools that we adapted from the biotech industry, like an open [INAUDIBLE] mixer. We have a rig here to measure the Hansen Solubility Parameters for paints and coatings and beauty and personal care industry, accelerated thermal conductivity measurement.
And we are using our platform to-- we are starting to develop our own materials IP portfolio for that direct air capture, so basically materials that can absorb CO2 from the air. And this is a rig that we are developing, we are building, to be able to measure and to screen and to generate thousands and thousands of results of different materials, training our machine learning, and start developing our own materials.
So we are targeting companies in each one of these five fields here. Because we want to advance with at least this for our United Nations Sustainable Development Goals. And on doing that, one of the contributions that we want to leave is a more environmentally sustainable planet.
In terms of business model, we named our model materials as a service. We provide service to company. We co-create materials with companies through joint development agreements. And as I mentioned, we are starting to develop our own materials IP portfolio to license in the future very focused on direct air capture.
We are looking for commercial pilots globally. We are already serving companies in the pulp and paper space, polymers, additives for chemical industry, paints and coatings. And we'll partner with you to accelerate innovation and especially to improve your sustainability footprint. Thank you.
[APPLAUSE]