If two years from now you look back wondering what the OpenAI moment was for robotic humanoids – their ChatGPT moment, when things really got real – it very well may be the below video posted last week by privately-held Figure...
Figure 01, as it’s called, debuted in this post from Figure on social media, showing a human robot that while slow acting, sounded and most importantly reasoned like, well, a real person...
My first reaction was that it was a fake video created entirely by AI – kind of an advanced version of Nikola’s purportedly self-propelled hydrogen-powered truck driving down the road. It apparently wasn’t, in which case Figure 01 is not just the latest, but certainly most advanced of any humanoid debut so far.
And that includes Tesla’s ($TSLA) Optimus, which has probably generated the most buzz, certainly among investors. But just yesterday Gizmodo ran a story comparing Optimus to Figure 01 under the headline...
Whether that’s really the case isn’t the point...
The point is that the next big thing investors are likely to chase as the natural evolution of AI – even if it’s two years from now – is likely to be human robots.
‘An Inflection Point’
This is something my futurist friend Paul Kedrosky of SK Ventures
has been telling me for months... but I just couldn’t find the right hook. “Robots are at an inflection point,” he was saying, citing the emergence of large language models, imitation learning and declining costs.
A few days later Figure announced funding from Jeff Bezos, OpenAI and Nvidia, giving it a valuation of $2.6 billion.
That sparked a few stories, including this one on Bloomberg, headlined...
But so far, the coming boom in humanoids hasn’t yet really caught Wall Street’s fancy, which you can clearly see in the below chart, which shows the number of times the word “humanoid” has shown up in searches on AlphaSense…
While mentions in transcripts are clearly up over the past few years, since 2009 it has made its presence in only 765 documents, including just 74 transcripts. Last quarter alone, the word showed up in a mere nine transcripts... and that was the most ever in a single quarter.
By comparison – and to show how fast these ideas can ramp – “large language model” has been mentioned in 3,912 documents over the same span, including 1,353 transcripts. Last quarter alone it showed up in 242 transcripts, which is quite a bit less than its peak a few quarters earlier.
The obvious reason is that there are simply so few public pure plays.
Another reason, Paul believes, is because “it’s caught in what Gartner calls the trough of disillusionment. Robotics has been hyped and failed many times, so people roll their eyes...”
The Nvidia Effect
That could be changing, though, especially after Nvidia’s CEO Jensen Huang mentioned “humanoids” four times during his address yesterday at the company’s annual user conference, while surrounded by a group of human robots. “We’re starting to do some really great work and the next generation of robotics,” he said, adding: “The next generation robotics will likely be humanoid robotics.”
Given Huang’s and Nvidia’s current popularity, that should send investors scampering in search of plays on humanoids...
That’s already been happening out of the limelight by venture investors, with CB Insights citing record funding for humanoids. Many of those, CB says, “are targeting more realistic human-robot interactions using large language models and natural language processing.”
The thing is, there’s more to robots than language, and plenty of money will be made in the less-sexy side of the story: companies that make nuts and bolts, or in this case case, the likes of actuators and sensors.
The challenge, however, is that if they’re not private, most of these manufacturers are buried inside larger companies, like Swiss-based ABB Ltd., Japan’s Keyence and Japan’s SMC, where the “humanoid” aspect is currently too small to have impact.
Enter the ETFs
Which gets to the bigger question: What’s currently the best publicly traded humanoid idea?