Why the Next Generation of Robots Might Be Soft and Squishy
13:40 Share

Marketers, you know that feeling when your content just works? When you crush a viral trend before 10 a.m.? That's Contentful. Dynamic content made blissfully simple. Contentful helps you create and launch personalized experiences instantly across any digital channel. No limits, no stress, only possibilities. Come get the feels at Contentful.com.

Hey, TNB listeners. Before we get started, heads up. We're going to be asking you a question at the top of each show for a couple more weeks. Our goal here at Tech News Briefing is to keep you updated with the latest headlines and trends on all things tech.

Now, we want to know more about you, what you like about the show, and what more you'd like to hear from us. Our question today is, how often do you want new episodes, and how long do you want them to be? Do you want more shorter shows or longer shows less frequently? If you're listening on Spotify, look for our poll under the episode description, or you can send an email to tnb at wsj.com. Now, onto the show.

Welcome to Tech News Briefing. It's Tuesday, May 27th. I'm Julie Chang for The Wall Street Journal. The chip industry has entered the era of so-called sovereign AI. What that means and what it'll tell us about NVIDIA's future as the chipmaker prepares to report earnings tomorrow.

And the next generation of robots may not be humanoid, but rather inspired by animals like turtles and worms. We got an inside look at the lab rethinking robotics. But first, NVIDIA CEO Jensen Huang has talked about this idea called Sovereign AI.

Essentially, it suggests countries will invest directly in artificial intelligence instead of relying on companies for AI services. And we saw this idea play out recently with major deals NVIDIA struck with the likes of Saudi Arabia, India, and the UAE to build up AI infrastructure in their countries. We realize now we are an AI infrastructure company, an infrastructure company that's essential all around the world.

Every region, every industry, every company will build these infrastructures. That was Huang speaking at the Computex 2025 conference in Taiwan this month.

And while Huang sees opportunity, our Heard on the Street columnist Dan Gallagher recently wrote about the risks that sovereign AI also brings to chip makers. Dan, what are some of those challenges? I don't think there's a business right now that's more politicized than chips, especially AI chips. These AI chips are front and center in the trade war with China between the U.S. and China. The Trump administration and the Biden administration before that really wanted to make sure China did not get its hands on the latest chips.

AI chips to get more competitive with the U.S. And so that governs everything that NVIDIA can do in terms of how it can sell its chips into other countries. And it's been a back and forth. The Trump administration did recently kill a rule that the Biden administration had set up where it was going to be a lot harder for NVIDIA to sell chips into other countries, even like friendly ones to the U.S. There's going to be all these different rules.

Those were scrapped, but the signs now are that NVIDIA's ability to sell into other countries might hinge on bigger picture stuff. Are these countries striking trade deals with us? Is there a concern that chips sold to one country could eventually find their way to China? So it's just going to be a much more complicated business than NVIDIA just getting the country to sign on the dotted line to make some sales. There's going to be a lot of politics involved. How?

How important is this idea of sovereign AI to the future of NVIDIA? Sovereign AI is important because it's another potential growth leg for them. Right now, they've been growing so much, doubling revenue every year for the last couple of years. Investors expect to see strong growth continue for the company. And so much of that is relied on a few very big companies spending a really large amount of money.

doing that. Capital spending by the companies like Microsoft, Google, and so forth, that's continuing. But eventually, that's going to get to a point where it levels off and gets to like a normalization. So the fact of having whole countries

also become very big customers for NVIDIA chips is another potential avenue for growth for them. So this takes us to earnings. NVIDIA earnings are coming out tomorrow. What are you going to be looking out for? Right now, there's a lot of worry about NVIDIA just because they've already made clear that these chips that they designed specifically for the Chinese market that they thought would be compliant with export controls, they can't sell those anymore. We learned that last month.

And so how that's going to impact the rest of their year is kind of a question mark because NVIDIA tends not to give forecasts for the whole year. They just tend to forecast for the quarter. So there's a potential that their forecast might be a little disappointing just because they're still dealing with the shutoff of this market. The deals that we've been talking about, sovereignty ideals and so forth, those are more like mid to longer term things. We don't know how much money is going to be flowing into the near term. So it's going to be an interesting report to see how they try to frame that up.

That was Heard on the Street columnist Dan Gallagher. Coming up, soft and squishy robots. How a lab at MIT is rethinking robot design. That's after the break.

Marketers, you know that feeling when your content just works? When you crush a viral trend before 10 a.m.? That's Contentful. Dynamic content made blissfully simple. Contentful helps you create and launch personalized experiences instantly across any digital channel. No limits, no stress, only possibilities. Come get the feels at Contentful.com.

Everyone these days seems to be thinking of robots as hard, metallic, and humanoid. But researchers at one MIT lab are taking a different approach. Lots of different approaches, actually.

They're working on robots that are squishy. If you have a hard hand and you don't grip something exactly right, well, it's going to be stuck with that kind of bad grasp. But if you have something that's like really, really soft, if you're a bit in error with what your grasp is, the silicone actually conforms to the shape. Robots that swim. My toes are roughly inspired by actual sea turtles. And even robots that are edible.

Our enterprise tech reporter, Isabel Busquets, visited MIT's Computer Science and Artificial Intelligence Laboratory. She spoke about what she saw and what researchers there are working on with our colleague, Victoria Craig. VICTORIA CRAIG: So when we think about robots,

tend to lean toward Elon Musk's view of these very hard humanoid style robots. But the ones you've seen are soft and squishy and decidedly not human. What do they look like and what are they made of? Historically, robots have either looked like humanoids, metal humans like C-3PO, or they've looked like like a

a box on wheels almost. And those are the kind of robots you see in factories, more like the R2D2s of the world. But the director of MIT's computer science and AI lab, a woman named Daniela Roos, she's really driving this idea that we can think differently about robots. They don't have to be either humanoids or boxes on wheels. She's made robots out of paper, out of food, out

Robots we can eat. She's a proponent of soft and squishy robots, things that can exist in the human world and won't hurt you if you accidentally touch it. One of the really fun things her lab is working on is a sea turtle robot named Crush.

after the Finding Nemo character. He's kind of a combination of hard and soft components. So he has silicone flippers, essentially, and he's designed to monitor coral reefs and sea life and swim around. And his flippers are soft so that, you know, he doesn't damage anything.

So are these soft-shelled robots mostly used in water, or what other use cases do they have? Yeah, I mean, they can really be anywhere. The possibilities are really limitless. They're also working on, like, a soft-shelled

robotic hand, something that can lift fruit in a grocery store. For example, you may not want a hard metal robot in a grocery store helping you pick out pears or oranges. Daniela Roos' lab, it's focused on a really broad array of things. They've really pioneered a lot of the underwater robotics, but they're also working on drones,

aka flying robots. They're also working on autonomous vehicles. They're also working on the humanoids. They're working on robots that will, you know, cook and potentially stir and pour things in the kitchen.

Overall, the focus is just on robots that will integrate better with the natural human world. And Danielle told me one of the things that got her into this in the first place was that when she was doing her PhD at Cornell, she was assigned to design a robot that could pick something up.

And it seems like a really simple task. But if you think about just having like your hand being a totally hard shell, it's really hard to pick something up. That's what made her realize that having these soft components on something like a robotic hand is really crucial.

And she's been working on this research at MIT since the early 90s. So this isn't necessarily a new area of study for her or the people there. Isabel, you visited this lab and the testing facility there. What's it like being there and what other kinds of innovation is going on? It's totally different from anything I ever pictured. You think about a lab and you think everyone is sitting at a desk inside a room building something. But

we were all over. We were outdoors with the drones in the parking lot with the self-driving car at the pool where there were actual MIT swimmers and lifeguards. So it's expansive. And

It's a huge facility. It has about 1,800 students and researchers at this point. They're also working on the algorithms that will power these robots, power their brains and their movements. She's developed this device

sort of new type of AI algorithm that she thinks will be better at powering robots than the AI algorithms we currently use. Currently, a lot of the algorithms we use are inspired by the human brain, which is

really complex and really hard to emulate. But she had this idea to create these algorithms that are inspired by the neural networks of worms, because their brains are a lot simpler, but they can still do a lot of things. So she

She developed these new types of AI algorithms called liquid networks. And she and a couple other researchers actually spun off a company called Liquid AI, where they're trying to bring that innovation in the software side into more robots as well.

That's so fascinating. Replicating a worm's brain I don't think would be on anyone's bingo card for research into AI. Would it be to create a robotic worm or what kind of uses would that kind of thing have? No, it's not necessarily that the physical...

body of the robot would be a worm. It's more that they studied the way that the neurons in a worm's brain talk to each other. And then they replicated that with math. Essentially, it's different math for how the neurons talk to each other. Maybe there could be a robotic worm.

They have a robotic eel. They have a robotic rabbit. There is no limit when it comes to sort of the animal world and what they can build with robots. That was the WSJ's Victoria Craig speaking with our enterprise tech reporter, Isabel Busquets.

If you want to see photos of the robots, she's got a bunch in her story. We'll drop a link to it in our show notes. And that's it for Tech News Briefing. Today's show was produced by me, Julie Chang, with Deputy Editor Chris Zinsley. We'll be back this afternoon with TNB Tech Minute. Thanks for listening.

Marketers, you know that feeling when your content just works? When you crush a viral trend before 10 a.m.? That's Contentful. Dynamic content made blissfully simple. Contentful helps you create and launch personalized experiences instantly across any digital channel. No limits. No stress. Only possibilities. Come get the feels at Contentful.com.