Category: Uncategorized

TL;DR: IEEE Spectrum checked in with the pros, and the consensus is clear: Optimus has made massive strides, but Boston Dynamics and Figure aren’t giving up their leads anytime soon.

It’s easy to get caught up in the hype when a new demo drops, but what do the actual engineers building these things think? IEEE Spectrum recently dove into exactly that, pulling back the curtain on how the robotics community views Tesla’s Optimus right now.

The vibe? Deeply impressed, but not blindly sold. When Optimus was first announced with a guy in a spandex suit, the industry laughed. They aren’t laughing anymore. The actuators and the end-to-end neural net approaches Tesla is showing off are legitimately top-tier. But let’s be real—they’re up against Boston Dynamics, who have been perfecting Atlas for over a decade, and aggressive newcomers like Figure.

The experts point out that while Tesla’s AI integration is incredibly fast, physical compliance and battery efficiency under heavy load are still massively tough nuts to crack. It’s an all-out sprint between Silicon Valley AI speed and legacy robotics precision. Whoever bridges that gap first wins.

Source: IEEE Spectrum

TL;DR: Elon Musk is targeting 2027 to scale the Optimus program, but manufacturing a mass-market humanoid robot brings a whole new set of brutal challenges.

Here’s the thing about building robots: making one incredible prototype is hard. Making ten thousand of them reliably is a nightmare. Tesla is officially eyeing a 2027 timeline to scale Optimus, and the industry is watching closely. Can they actually pull it off?

If you look at the recent analysis out of eWeek, the consensus is cautious optimism mixed with a heavy dose of reality. Tesla has a massive advantage because they already know how to mass-produce complex electric machines—cars. They have the supply chains, the battery tech, and the AI compute infrastructure. But a bipedal robot has entirely different failure modes than a Model Y.

By 2027, Tesla wants Optimus off the factory floor and doing useful things at scale. The trick won’t be teaching it to walk. The trick will be driving the cost down enough so that deploying a fleet of them actually makes financial sense. We’re about to find out if Elon’s aggressive timelines hit reality, or if we’ll be waiting until 2030.

Source: eWeek

TL;DR: Tesla brought Optimus out of the lab and into the crowds at the Boston Marathon, taking photos and proving it can handle real-world public interactions.

Look, seeing a robot in a pristine lab environment is one thing. Watching it interact with exhausted marathon runners is something entirely different. Tesla just pulled off a brilliant PR move by bringing their Optimus humanoid out to the Boston Marathon. It wasn’t running the course, but it was doing something arguably harder: dealing with unpredictable humans in a crowded, noisy environment.

Optimus spent the day greeting runners, waving, and posing for photos. No safety cages. No completely scripted lab routines. Just a humanoid bot standing around like a regular spectator—well, a metal spectator with glowing sensors. This highlights exactly what Tesla is trying to prove right now. They don’t just want a robot that can sort widgets; they want something you’d be comfortable standing next to.

And honestly? It worked. People weren’t terrified; they were taking selfies. Getting the public used to having humanoids in their space is going to be half the battle for companies like Tesla and Figure. If the Boston Marathon is any indicator, we’re adapting pretty fast.

Source: Interesting Engineering

TL;DR: Unitree pushed a massive software update for the $16k G1 humanoid, unlocking dynamic balancing and object manipulation previously reserved for six-figure bots.

Unitree just completely changed the value prop of the G1. They pushed a software update that unlocks dynamic balancing capabilities we usually only see on bots that cost ten times as much. The new demonstration video is wild.

We’re talking about the robot recovering from heavy shoves and navigating uneven terrain without missing a beat. On top of that, the hand dexterity got a massive upgrade. The G1 is now manipulating tools with a level of precision that makes you forget it costs less than a used Honda Civic.

The gap between the budget tier and the premium tier is closing incredibly fast. And Unitree is driving that wedge.

Source: Unitree Robotics

TL;DR: The latest Optimus update ditches the choreographed routines for actual utility, showing fully autonomous navigation and part retrieval inside the Fremont factory.

Tesla’s latest Optimus update shows the bot navigating the Fremont factory floor entirely on its own. Forget the dancing. We’re looking at actual utility. The footage shows Optimus identifying misplaced parts, picking them up, and returning them to the correct bins without any teleoperation.

The walk cycle looks significantly smoother now. The awkward shuffle is mostly gone. It’s using vision to map the aisles and avoid forklifts in real time. This is the kind of boring, repetitive work that actually justifies the R&D spend.

Sure, it’s still early days. But the trajectory is clear. Optimus is shifting from a science project to a piece of factory equipment.

Source: Tesla Updates

TL;DR: Figure AI’s humanoid bots are no longer just running demos. They’re handling complex, multi-step tasks at BMW’s Spartanburg plant across full double shifts.

Figure AI just dropped an update on their humanoid deployment in Spartanburg. Honestly, the speed of progress is staggering. They’ve moved from simple part placement to handling complex, multi-step subassemblies. And they’re doing it consistently.

BMW is reportedly expanding the fleet to three new assembly lines by Q3. It’s not just a PR stunt anymore. These bots are pulling their weight. You can see the fluidity in the new video. The movement isn’t stiff or programmed; it’s reactive. They adapt when a part is slightly misaligned.

Look, it’s easy to get cynical about robotics timelines. But watching Figure 02 keep pace with a live automotive line changes the math entirely.

Source: Figure AI Blog

TL;DR: While Tesla’s bot took photos in Boston, a Chinese humanoid robot actually ran a half-marathon in Beijing—beating the human world record.

Here’s the thing. We were all distracted by Optimus waving at people in Boston. Meanwhile, over in Beijing, a humanoid robot literally destroyed the human half-marathon world record.

It crossed the finish line in a blistering 50 minutes. That’s a full seven minutes faster than any human has ever run that distance.

This completely flips the script on robot mobility. We’re no longer talking about bots that can barely navigate stairs without tumbling over. We’re looking at machines that can sustain high-speed dynamic movement over long distances. The robotics race is heating up, and right now, China is literally sprinting ahead.

Read the full story on The Next Web.

TL;DR: Elon Musk is targeting an insane 10 million Optimus units produced out of a new Texas facility.

Tesla isn’t just making cars anymore. The pivot to robotics is happening faster than anyone predicted, with Elon Musk announcing plans to crank out 10 million Optimus units.

The plan centers on a massive new facility in Texas dedicated entirely to scaling up humanoid robot production. If they hit even a fraction of that 10 million target, it completely changes the labor market overnight.

Sure, Musk has a history of wildly ambitious timelines. But the sheer scale of the investment proves Tesla is betting its entire future on robotics, not just EVs.

Read the full story on The Robot Report.

TL;DR: Tesla deployed its Optimus humanoid to greet runners and pose for photos at the Boston Marathon, scoring a massive, free marketing win.

Tesla just pulled off a masterclass in free marketing. Instead of dropping millions on ads, they sent their Optimus humanoid robot straight to the Boston Marathon finish line.

The bot didn’t run the race, obviously. It spent the day greeting exhausted runners and posing for selfies. But honestly, that’s exactly the kind of public exposure humanoid robotics needs right now to shake off the sci-fi fear factor.

Look, seeing a faceless metal humanoid in a lab is terrifying. Seeing one high-five a sweaty marathoner? That normalizes the tech. And it proves Tesla’s marketing playbook still works beautifully without spending a dime.

Read the full story on Inc.

TL;DR: Chinese robotics companies like Unitree absolutely dominated humanoid sales in 2025, taking a 90% global market share while US rivals struggled to break 150 units.

Silicon Valley loves a good concept video. But while American startups were polishing their marketing, China actually shipped the hardware. In 2025, a massive 90% of all humanoid robots sold globally came from Chinese firms.

Unitree took the crown, selling 5,500 units last year. Agibot was right behind them with over 5,100. Let that sink in. They’re moving thousands of metal bodies while major US players—like Figure AI, Agility Robotics, and even Tesla—sold roughly 150 each.

Why the blowout?

China is running the exact same playbook they used to dominate the EV market. Massive state funding, cheap local components, and ruthless competition. They have government-backed testing facilities specifically designed to accelerate humanoid R&D. It’s an industrial machine dialed in on dominating the next big tech wave.

Even Elon Musk had to admit it. Last month at Davos, he called China the “toughest competition” for Tesla, noting there are barely any serious competitors outside of it. He still thinks Optimus will win in the end, of course.

But right now? The scoreboard isn’t even close.

Catch the full report at Rest of World