One might think that the Toyota Research Institute (TRI), founded with a billion-dollar investment and a mission to develop artificial intelligence and advanced robotics, wouldn’t need to look elsewhere for fresh ideas. But even with the resources of one of the world’s biggest automakers behind it, executives with company’s research arm acknowledge that they can only do so much.
“The [number] of startups is incredible,” said Gill Pratt, chief executive officer at TRI, speaking specifically of the transportation technology sector. “No matter how successful I or other OEMs or even tech companies are at hiring, we’re still missing out on innovation at these startups.”
To harness some of that talent, TRI announced at the Automated Vehicles Symposium in San Francisco that it’s starting a venture-capital fund that will provide early-stage funding for startup technology companies. The company has set aside an initial $100 million and has already made investments in three companies.
“The auto industry is changing from data being an exhaust of the product to, now, data forging the product.”
– Jim Adler, Toyota AI Ventures
The initial class includes: Nauto, a Silicon Valley startup creating computer-vision software that monitors driving habits and distills data-based insights for fleets; SlamCore, a British company developing mapping for autonomous cars with a focus on power efficiency; and Intuition Robotics, an Israeli startup focused on robotic companions for lonely adults. Terms of these initial investments were not disclosed.
On its face, that last one is a curious addition for any automotive-related venture. But since its inception less than two years ago, TRI has maintained a broad focus on robotic applications that branch beyond cars and into houses. Jim Adler, who now takes the role as managing director of the fund, said the artificial intelligence developed by Intuition Robotics could have voice-recognition and response applications for both cars and homes.
Toyota’s foray into venture capital, formally called Toyota AI Ventures, comes at a time when the automotive-technology sector is flush with cash. Startup financing in advanced transportation exceeded $1 billion in 2016, according to CB Insights, a market analytics firm, and it has already surpassed $1.6 billion in the first seven months of 2017.
Competitors have made similar efforts, and in some cases they’re not new. In 2009, Ford executive chairman Bill Ford was a co-founder of Detroit-based Fontinalis Partners, which started funding new mobility startups long before that was an industry trend. More recently, in late 2016 German supplier ZF Group formed its own fund, called Zukunft Ventures, to focus on efficiency, safety, and autonomous tech.
With so many investors seeking stakes in promising startups, Toyota and other traditional automotive companies believe they have something to offer beyond money.
“The world is awash in venture cash, but it’s not awash in companies that make 10 million cars a year that could make for extremely excited customers for the technology,” Pratt said. “That’s far more rare. I hope our money can be a smarter choice for some of the startups, even though there is a lot of cash incentive.”
“The world is awash in venture cash, but it’s not awash
in companies that make 10 million cars a year.”
– Gill Pratt, Toyota Research Institute
In disbursing it, he hopes to borrow a strategy from his days as a program manager with the Defense Advanced Research Projects Agency (DARPA), which funded projects both by reviewing traditional pitches for funding and by identifying gaps in technology and calling for people to form companies dedicated to solving them.
Some current gaps include a need for greater power efficiency in autonomous-vehicle subsystems and a need to better understand why deep-learning systems make the decisions they do. While those lean toward future systems, there are also needs for automotive-grade sensors for autonomous vehicles and better tools for distilling insights from data.
“The auto industry is changing from data being an exhaust of the product to, now, data forging the product,” Adler said. “So we know data is really important.”
While TRI has already funded a data company in Nauto, Adler said the venture fund won’t hesitate to fund more data-insight companies. Part of the plan may include funding multiple companies working in the same realm and simply waiting to see which develops the stronger product.
“The nice thing about startups is that you can run many experiments in parallel, and winners will emerge,” Pratt said. “We invest, and we buy the winners. I think that’s a great model.”