Featured
TRI’s Accelerated Materials Design & Discovery (AMDD) Program — An Origin Story
By: Brian Storey - Vice President, Energy and Materials
Eight years ago, TRI launched the Accelerated Materials Design and Discovery (AMDD) program, partnering with top universities to develop new tools, models, and methods that push the boundaries of what’s possible at the intersection of AI and materials science.
We’re taking a moment to reflect on the impact of this ambitious journey before launching into what’s next. This article marks the start of an ongoing series that will share the breakthroughs, lessons, and future vision shaped by AMDD.
Follow us on Medium and keep an eye here over the next few months to see new articles as they are published.
At Toyota, materials science research has a rich history that dates back to the company’s transition from making textile looms to making automobiles. Toyota established its Materials Testing Center in the 1930s to develop better steel for mass-produced cars. Materials research grew substantially in the 1970s when Toyota made significant discoveries in materials for catalytic converters that eliminated harmful tailpipe emissions and reduced smog. The importance of catalytic converter materials in improving air quality and human health is hard to overstate.
Today, the need to discover new materials that solve critical problems for society has never been more urgent. Materials are at the heart of all clean energy technologies, such as batteries, hydrogen fuel cells, solar cells, electric motors, and catalysts for creating low-carbon fuels. It’s the materials inside that determine the performance, cost, and durability of all these technologies. At the same time, our material choices are limited by geopolitical, geological, environmental, and economic forces. To fully decarbonize transportation, we must move from a system powered by fossil fuels to one powered by truly sustainable materials.
But the development of new materials is notoriously a multi-decade slog. One example is today’s lithium-ion battery, which took four decades to go from Nobel-prize winning discoveries to ubiquity. Due to the importance of discovering new technologically relevant materials, the aim of accelerating materials development has long been a priority for all industries. In 2011, US President Obama launched the Materials Genome Initiative (MGI) with the goal to “help businesses discover, develop, and deploy new materials twice as fast”. Building on its long traditions in materials science, Toyota began significant investments in internal research aligned with the goals of MGI around this same time.
An Overview of AMDD
In late 2015, we announced the creation of Toyota Research Institute (TRI) with a billion-dollar investment and a mission to “accelerate R&D in a range of fields to help resolve society’s future challenges by using artificial intelligence (AI) and big data.” The founding of TRI came when the growing urgency for new materials discovery intersected with emerging advances in AI. In response to this opportunity, TRI began investing significant resources into accelerating new materials discovery for clean energy. The time was right for Toyota to further accelerate its materials research efforts and to try to advance the art of the possible.
A major component of TRI’s overall investment in materials research has been university-sponsored research. An initial four-year funding program was launched in March of 2017, with a renewed four-year investment announced in 2021. Over the past 8 years, TRI has invested approximately $70M in university research focused on Accelerated Materials Design and Discovery (AMDD). To launch the AMDD program in 2017, we had an open call for proposals with the overarching theme shown in the box below. The strategic goal of the program for Toyota was to enlist some of the top researchers to push the state of the art in AI for materials and see whether practical breakthroughs could be made.
“The long-term vision of the program is to aid in the development of truly emissions-free and carbon-neutral vehicles for the Toyota 2050 Environmental Challenge. Realizing the vision will require the development of new materials for batteries, fuel cells, catalysts, and more. The timescale for new materials development is traditionally measured in decades. Our aim at TRI is to merge advanced computational materials modeling, new sources of experimental data, machine learning, and artificial intelligence to dramatically accelerate the materials design and discovery process.”
From the original AMDD call for proposal
Looking back
Over the course of eight full years, the AMDD university consortium comprised 17 separate projects, involving 14 different institutions, and 36 principal investigators (PIs). Including all the graduate students and postdocs, at any given moment, there were approximately 125 people whose research was funded through AMDD. The largest projects involved multiple institutions and/or PIs with budgets over $1M per year, which supported over a dozen graduate students and postdocs in their research. While some smaller seedling projects had budgets to support one or two students. There was great diversity in the project topics and structure within the consortium. The full scope of the people involved can be seen in the photo below, when we held a conference for everyone supported by the consortium in 2019.
Today, in 2025, TRI’s material discovery research continues with a strong internal research team, additionally supported through Toyota’s increased investment in the Global AI Accelerator (GAIA), and our continued investment in the next generation of scientists through our sponsored research. However, as the original structure of the AMDD university consortium has evolved and many of our original investments have concluded, it is a good time to take stock of what we have done in this first phase of our history.
Looking back on the past eight years, the scientific output from such a large group has been incredible. But we know from many studies on the market failures of research that a financial return on investment is difficult to track, as the fruits of research can take many years to bear. A large fraction of research never makes it to real-world implementation for various reasons (i.e, cost, feasibility, practicality, etc). The value of scientific research often lies in intangible factors that are difficult to quantify, and it will take some time before we can assess the impact of the research on Toyota’s bottom line.
What’s Next
This article kicks off a series in which we discuss some of the highlights, intangibles, challenges, and learnings from our consortium. In the spirit of Kaizen, “continuous improvement,” we will reflect on our experiences and glean lessons that will hopefully be useful for others. We are also excited to celebrate the accomplishments of a world-class group of scientists and engineers with whom we have had the honor and opportunity to work over the years.
Watch for my upcoming series to touch on some of the following issues and themes:
- Scientific advances — The program yielded over 300 publications, many of which have been high-impact. Beyond these singular publications and results, we have numerous cumulative scientific advances that occurred through a series of connected papers and people over multiple years.
- People — AMDD was responsible for training and inspiring the next generation of leaders. TRI was the primary funder of at least 45 PhD theses. About 20 of our alumni are now professors themselves who are working in the same field, around the world. Beyond academia, our alumni are leading efforts at other tech companies working on AI for materials, leading startups in materials science, and carrying the inspiration from the AMDD consortium forward. In some cases, startup founders were connected through our consortium. I am proudest of some of these outcomes.
- Deep collaboration — In the 2021 press release, I was quoted (and meant it!) when I said, “Our focus on collaboration is what makes our research program unique. Rather than acting strictly as a funding source, TRI has formed deep collaborations with researchers which have led to joint publications as well as co-developed open-source data and software.” This direct collaboration led to some of the most unique results in our program.
Recent advances in AI have brought a lot of new excitement and attention to the materials discovery field, and numerous companies have entered the race to find better materials faster. This race includes established tech companies as well as brand-new startups. We at TRI share in that excitement and see a very bright future ahead. But we also know that there remain many challenges we must face head-on. The world of atoms is very different from the world of bits. Materials in isolation are not the same as materials working in a device. Scaling from one sample in the lab to materials at the scale of ~10M cars per year built globally by Toyota is no easy feat. To help solve the grand challenges facing society through materials discovery, our work is not done.
I hope that this series offers insights and lessons as we consider how to realize the true vision of accelerated materials design and discovery.
Stay tuned for more: https://toyotaresearchinstitute.com/amdd/
