A Stanford team trained humanoid pick-and-place policies on a Unitree G1 with zero teleoperation demos.
Sim data only.
And it matched or beat human teleop on every task.
The headline reads "teleop-free."
But the part worth stealing is how they made sim images real enough to actually transfer.
Most sim pipelines render the whole scene as mesh.
The images come out flat and fake, so the policy never crosses the sim-to-real gap.
LEGS renders only the robot and objects as mesh.
Then it drops them onto a 3D Gaussian Splat of the real room, a reconstruction trained against actual photos.
So the background is photorealistic, and the visual gap mostly closes.
On the hardest long-horizon task, teleop collapsed to 0/10.
Same across all three VLA backbones they tested.
The sim-trained policy still landed 2 to 6 out of 10.
The neat part: motion is recorded as a command stream, separate from how the scene looks.
So one recorded dataset re-renders under new rooms and new objects at GPU cost alone.
Covering a new scene drops from ~1.5 operator-hours to ~0.1 GPU-hours.
Worth being honest about the setup though.
SONIC runs the same whole-body controller binary in sim and on the robot.
So locomotion sim-to-real is already handled before LEGS even starts.
And the teleop baseline was a non-expert at 50 demos.
Read it as: photorealism plus procedural motion can replace teleop at the data step, in this setup.
The real lesson is the decoupling.
Once appearance and motion are separate, adapting to a new scene becomes a render job, not a collection job.
Nice work from the Stanford team
1
14
Fati Sule retweeted
Article très intéressant sur la société chinoise Unitree, qui fabrique des robots. Certains passages sont parfois assez techniques, mais ça accessible.
Enfin bref, à lire.
1
4
11
466
Fati Sule retweeted
China's Unitree Will Dominate Global Robotics
The Fastest Iteration Cycle In Next-Gen Robotics Should See Unprecedented Acceleration
newsletter.semianalysis.com/…
1
4
303
McMurphy retweeted
whoa unitree humanoid robot becomes extremely good at everything he does.
Full fire from start to finish! Top-tier motion control
The OmniXtreme framework keeps pushing the Unitree G1 body past its physical limits.
From : @ 铁卫士
Wow, it's so realistic, it looks like a human gesture and not a humanoid robot.
4
6
276
Vivek Raj retweeted
your robot is about to become your most profitable asset.
demo in 2 weeks with a Unitree U2
1
2
18
Jorge Aponte Careaga 🇲🇽 🇵🇸 retweeted
¡El kung fu se encuentra con la tecnología del futuro! Un robot🤖 desarrollado por la empresa Unitree Robotics, con sede en Hangzhou, sorprendió al público demostrando artes marciales tradicionales chinas con movimientos fluidos y precisos, así como un impresionante nivel de equilibrio y coordinación. ¡Parece sacado de una película de ciencia ficción! ✨ #UnitreeRobotics #robot #kungfu #martialarts
3
37
109
3,774
Geopolitics: the Pentagon expands its blacklist to 188 Chinese companies (Alibaba, Baidu, BYD, Unitree, WuXi).
The US closes access. China accelerates in parallel.
1
8
Geopolítica: el Pentágono amplía su lista negra a 188 empresas chinas (Alibaba, Baidu, BYD, Unitree, WuXi).
EE.UU. cierra el acceso. China acelera en paralelo.
1
16
"يحاول جذب تعاطف المارة بعبارة ساعدني من فضلك تكاليف الصيانة والكهرباء غالية جداً".. روبوت "شحات" يتسول المال في الصين باستخدام رمز استجابة سريعة "كيو آر" (QR) موجود أمامه
🔴 الروبوتات غزت الشوارع الصينية وأصبح تواجدها أمرا معتادا للمارة
🔴 فوجئ الناس في الشوارع الصينية بظهور روبوت بشري جالس يتسول على قدميه عبر مجموعة من الإشارات الواضحة،
🔴 الربوت المتسول يطالب المشاة بالتبرع له باستخدام رمز استجابة سريعة "كيو آر" (QR) موجود أمامه وفق مجموعة من مقاطع الفيديو تم تداولها في منصات التواصل الاجتماعي
🔴 الروبوت يبرر طلبه للأموال بارتفاع تكاليف الصيانة والكهرباء مما يجعل المعيشة أمرا صعبا، كما تفاعل الروبوت مع كل المارة
🔴 ينتمي الروبوت الذي ظهر على مواقع التواصل يتسول إلى شركة "يونيتري" (Unitree) الصينية ويصل سعره إلى نحو 16 ألف دولار.
4
18
9,417
Lady Tipster retweeted
May 25
Conference Room Mess Cleanup Test: Unitree WVLA 2.0 Model🎉
This video was recorded in a single take.
Multi-task, fully autonomous operation. Strong external interference present.
86
152
1,482
4,799,040
Paul Myles retweeted
20 Mar 2024
Unitree H1 The World's First Full-size Motor Drive Humanoid Robot Flips on Ground.
Unitree H1 Deep Reinforcement Learning In-place Flipping !
#Unitree #UnitreeRobotics #AI #Robotics #Humanoidrobots #Worldrecord #Flips #EmbodiedAI #ArtificialIntelligence #Technology #Innovation
210
407
1,538
230,086
Benjamin Jost retweeted
Jun 5
By @spaceandtech_
“Unitree Robotics has unveiled the
GD01, a giant mecha that can carry
a human operator inside its cockpit”
| This 500-kg robot can either walk
on 2 legs or switch to 4-limb mode
#Tech #Unitree #robotics #China
Space and Technology
6
24
880
Şu görsel, yapay zeka süper döngüsünü tek bir çerçeveye oturtuyor ve çoğu yatırımcının kafasındaki karışıklığı güzel çözüyor. Ürettikler tez bunun 15 yıllık bir döngü olduğu ve bizim daha 4. yılında olduğumuz.
Evreleri kısaca geçeyim. Birinci evre için çoktan maliyetlenildi, yani çip, bellek ve optik kazananları, Nvidia ve benzerleri. Buradaki yükselişin büyük kısmı yakalandı bile.
İkinci evre tam şu an olduğumuz yer, yani veri merkezi ve güç altyapısının devasa inşası. GPU kümeleri, soğutma, elektrik şebekesi ve ağ tarafı. Haftalardır konuştuğumuz enerji ve şebeke hikayesi tam olarak burası.
Üçüncü evre fiziksel dünyanın yapay zekayla buluşması, yani robotik, dronlar, savunma, uzay ve nadir toprak elementleri. Unitree ve NEURA üzerine yazdıklarım bu evreye birlikte hazırlanabilmemiz için.
Dördüncü evre ise 2030 sonrası yazılım ve kurumsal tarafın yapay zekaya teslim olması, AGI altyapısı ve kuantum.
Evreler tabii ki bu kadar keskin ayrılmıyor ve listelerdeki isimleri de kesin ve değişmez görmemek gerek alternatifleri var. Yine de büyük resmi göstermek için fena bir harita değil.
May 8
The AI supercycle will last 15 years. We're in year 3.
Most investors are still buying Phase 1 names while the real money is already rotating into Phase 3.
I mapped the entire cycle into 4 phases with the tickers that matter at each stage:
The AI supercycle is the biggest investment theme of our generation. Bigger than mobile. Bigger than cloud. A 15 year structural shift that will reshape every sector of the global economy. Hyperscalers just committed $725 billion in capex for 2026, nearly doubling last year. Microsoft, Google, Amazon, and Meta each spending over $100 billion individually.
This is not speculation. I've mapped the entire supercycle into four phases so you know exactly where we are and where the asymmetric opportunities sit.
🔴 Phase 1: Already Ran (2023 to 2025)
The foundation layer is complete. $AMD ran 78% in 2025, $NVDA 39%, and $INTC just posted a blowout Q1 that sent the Philadelphia Semiconductor Index above 10,000 for the first time. Chips still power every phase but the generational entries are gone and risk/reward has compressed.
- $NVDA, $AMD, $ARM, $INTC, $AVGO, $MU, $GLW
- Semiconductors, Memory & Storage,Photonics/Optics
- Foundation complete. Still growing but priced for it.
🟠 Phase 2: Peak Buildout (2025 to 2027)
The phase most investors just woke up to. $CEG acquired Calpine to become the largest U.S. private power producer at 55 GW. $GEV up over 200% in a year. $VRT co engineering cooling for NVIDIA's Rubin architecture. $GLW up 74% YTD on optical fiber demand. Nuclear SMRs are the breakout with $OKLO, $SMR, and $BWXT positioning to power data centers directly. Still upside but the obvious names have moved.
- $CEG, $GEV, $VRT, $VST, $TLN, $ANET, $GLW, $MOD, $EQIX $OKLO, $SMR, $BWXT, $NNE
- Power/Grid, Cooling, Networking, Nuclear/SMR Peak buildout.
- Nuclear SMRs are the sleeper.
🟡 Phase 3: The Positioning Window (2026 to 2028)
Where AI escapes the data center and enters the physical world. Most will be late. Tesla converting Fremont to Optimus production, $25B capex, mass production targeted H2 2026. Rocket Lab posted record $602M revenue with $1.85B backlog. $LUNR up 47% YTD with $943M in contracts. $KTOS Valkyrie drone selected for the Marine Corps. The window to position is open right now.
- $TSLA, $RKLB, $LUNR, $KTOS, $AVAV, $PATH, $ISRG $MP, $FCX, $ALB, $ASTS
- Robotics/Autonomy, Space/Defense/Drones, Rare Earths
- This is where the asymmetric risk/reward lives.
🟢 Phase 4: Final Frontier (2028 )
The endgame. Microsoft capex $190B. Alphabet $190B. Amazon $200B. Meta $145B. Google Cloud backlog past $460B. They're building the rails for AI software dominance and AGI. Quantum still early but $IONQ and D Wave are laying groundwork. The platforms that control the software layer win the entire supercycle.
- $MSFT, $GOOGL, $AMZN, $META, $ORCL, $IONQ
- AI Software Dominance, AGI Infrastructure Decade long thesis.
- Accumulate on weakness.
💊 Key Takeaway
- Phase 2 is confirmed ($725B hyperscaler capex)
- Phase 3 is where the smart money positions nowRobotics, space, defense, nuclear
- SMR are the 2026 to 2028 trades
- Most will rotate into these names 12 months too late
15 year supercycle. Not a trade. Phase 1 ran. Phase 2 is priced. Phase 3 is where you want to be.
1
32
878