KAIST research team’s independently developed humanoid robot boasts world-class driving performance, reaching speeds of 12km/h, along with excellent stability, maintaining balance even with its eyes closed or on rough terrain. Furthermore, it can perform complex human-specific movements such as the duckwalk and moonwalk, drawing attention as a next-generation robot platform that can be utilized in actual industrial settings.

Professor Park Hae-won’s research team at the Humanoid Robot Research Center (HuboLab) of KAIST’s Department of Mechanical Engineering developed the lower body platform for a next-generation humanoid robot. The developed humanoid is characterized by its design tailored for human-centric environments, targeting a height (165cm) and weight (75kg) similar to that of a human.

The significance of the newly developed lower body platform is immense as the research team directly designed and manufactured all core components, including motors, reducers, and motor drivers. By securing key components that determine the performance of humanoid robots with their own technology, they have achieved technological independence in terms of hardware.

In addition, the research team trained an AI controller through a self-developed reinforcement learning algorithm in a virtual environment, successfully applied it to real-world environments by overcoming the Sim-to-Real Gap, thereby securing technological independence in terms of algorithms as well.

Currently, the developed humanoid can run at a maximum speed of 3.25m/s (approximately 12km/h) on flat ground and has a step-climbing capability of over 30cm (a performance indicator showing how high a curb, stairs, or obstacle can be overcome). The team plans to further enhance its performance, aiming for a driving speed of 4.0m/s (approximately 14km/h), ladder climbing, and over 40cm step-climbing capability.

Professor Hae-Won Park’s team is collaborating with Professor Jae-min Hwangbo’s team (arms) from KAIST’s Department of Mechanical Engineering, Professor Sangbae Kim’s team (hands) from MIT, Professor Hyun Myung’s team (localization and navigation) from KAIST’s Department of Electrical Engineering, and Professor Jae-hwan Lim’s team (vision-based manipulation intelligence) from KAIST’s Kim Jaechul AI Graduate School to implement a complete humanoid hardware with an upper body and AI.

Through this, they are developing technology to enable the robot to perform complex tasks such as carrying heavy objects, operating valves, cranks, and door handles, and simultaneously walking and manipulating when pushing carts or climbing ladders. The ultimate goal is to secure versatile physical abilities to respond to the complex demands of actual industrial sites.

During this process, the research team also developed a single-leg “hopping” robot. This robot demonstrated high-level movements, maintaining balance on one leg and repeatedly hopping, and even exhibited extreme athletic abilities such as a 360-degree somersault.

Especially in a situation where imitation learning was impossible due to the absence of a biological reference model, the research team achieved significant results by implementing an AI controller through reinforcement learning that optimizes the center of mass velocity while reducing landing impact.

Professor Park Hae-won stated, “This achievement is an important milestone that has achieved independence in both hardware and software aspects of humanoid research by securing core components and AI controllers with our own technology.

“We will further develop it into a complete humanoid, including an upper body to solve the complex demands of actual industrial sites and furthermore, foster it as a next-generation robot that can work alongside humans.”

The results of this research will be presented by JongHun Choe, a Ph.D. candidate in Mechanical Engineering, as the first author, on hardware development at Humanoids 2025, an international humanoid robot specialized conference held on October 1st.

Additionally, Ph.D. candidates Dongyun Kang, Gijeong Kim, and JongHun Choe from Mechanical Engineering will present the AI algorithm achievements as co-first authors at CoRL 2025, the top conference in robot intelligence, held on September 29th.

The presentation papers are available on the arXiv preprint server.

The Stargate Project has brought the global artificial intelligence race to the West Texas desert.

And on Sept. 23, it also brought a flock of industry leaders, U.S. congressmen, other policymakers and a gaggle of regional and news outlets.

All eyes are on the collaboration between OpenAI, Oracle and SoftBank to construct data centers and other infrastructure to support the artificial intelligence boom. On Sept. 23, Stargate also announced it would also be building five additional data center sites across the country.

There are plans to build more capacity near the flagship Abilene site, as well as sites in two other Texas counties, Shackelford County and Milam County. Other locations include Doña Ana County, New Mexico, Lordstown, Ohio, and another soon-to-be-disclosed location in the Midwest.

The joint venture was first announced at the White House in January with President Donald Trump, as part of a broader push for investment in American AI infrastructure.

As the high-stakes international competition to develop and deploy the technology escalates, the companies are betting big on the $500 billion program, with AI kingpin NVIDIA recently joining the fray by investing $100 billion in OpenAI, it announced.

Locals chat about the data centers over their morning coffee downtown―and anyone who’s paying attention to anything relatively tied to the AI industry at least knows about Stargate, even if they can’t point Abilene out on a map.

“Texas is ground zero for AI,” U.S. Center Ted Cruz told a crowd. He praised the state’s availability of low-cost energy, open-for-business environment with low taxes and low regulations, and the way the state lionizes entrepreneurs.

“So in my view, Texas and tech and AI are a perfect match,” Cruz said.

Over 1,000 acres of high-tech

The campus, about 180 miles from Dallas, is on track to provide OpenAI with the world’s largest supercluster when fully built, according to Oracle.

The 1,100-acre campus will have eight near-identical buildings, totaling up to 4 million square feet and is expected to be fully completed around this time next year.

The buildings house servers filled with graphics processing units (GPUs). There are numerous metal boxes with blinking lights and wires of various sizes and color.

Fiber is being installed both below ground and above, tubes are designed to pump a cooling liquid using a closed-loop system. A lot of pieces work together to support the highly technical compute needs.

A portion of the campus is already operating on Oracle Cloud Infrastructure after Oracle began delivering the first NVIDIA GB200 racks in June. NVIDIA’s deal with OpenAI will build and deploy at least 10 gigawatts of AI data centers.

It’s surrounded by rugged scenery: red dirt kicked up by gusts of wind, rocky terrain and short trees. The Abilene skyline is visible through a few miles of hazy air. There are roadways throughout the campus, including a makeshift six-lane “highway,” to ease the traffic from roughly 6,400 workers traveling on and off the site alongside semi-trucks.

OpenAI CEO Sam Altman said the Abilene campus is a fraction of what the partnership is building.

Even then, more infrastructure will still be needed to serve the demand of ChatGPT.

“We’ve got to make this investment,” Altman said. With global competition heating up between the U.S. and other major powers, ” … we cannot fall behind in the need to put the infrastructure together to make this revolution happen.”

Commitment to Abilene

The data centers being built at the new locations, which were selected among 300 proposals from more than 30 states, drive Stargate ahead of schedule to secure a full $500 billion, 10 gigawatt commitment by the end of the year.

“We’re really focused on enabling AI to have all the compute capacity needs,” new co-CEO of Oracle, Clay Magouyrk said.

Abilene Mayor Clay Weldon Hurt said his city is steeped in tradition, and acknowledged there is a mix of feelings among local residents. However, the town is open to progress, he added.

“I have a commitment to our citizens of Abilene to make Abilene a better place, and we have that commitment to grow,” Hurt said.

“So, even though we’re very proud of our heritage, and we’re always going to be proud of that heritage, we’re always going to be open [for business], and we’re so excited that this opportunity has come to Abilene, and we welcome it.”

Sen. Cruz called Stargate an impressive start, but encouraged more building and hiring.

“This is the beginning of a long-term effort to invest in American jobs, supply the additional power needed for AI, and deliver products and services that will benefit all Americans,” he said.

2025 The Dallas Morning News. Distributed by Tribune Content Agency, LLC.

Comparing Our Favorite Sheets on Amazon

Honorable Mentions

Boll & Branch Signature Sheets for $259: I love this brand’s flannel, and this signature sheet set is a solid choice as well (though pricey compared to other cotton sheet sets I prefer).

Buffy Eucalyptus Lyocell Sheet Set for $200: This set uses lyocell made from eucalyptus trees for a similar feel to bamboo sheets—silky and chilly to the touch. It’s another unique cooling option, and great for folks who love soft bedding.

Cariloha Classic Bamboo Sheet Set for $239: For being Cariloha’s cheapest set, it still feels fantastic to sleep on and delivers on the softness and comfort you want in a good bamboo sheet. Upgrade to Cariloha’s Retreat Sheets ($289) if you want a securing strap, which is handy for silky bamboo.

Coyuchi Organic Crinkled Percale Sheet Set for $189: I love a few different sets from Coyuchi, but none of the ones we’ve tested are available on Amazon. Still, it’s a good organic brand, so swinging for the crinkled percale and classic sateen ($178) will likely leave you satisfied.

Cozy Earth Bamboo Sheet Set for $330: This is a big favorite in the bamboo world for the super-cozy, silky soft feel. It’s unlike any other bamboo set I’ve tried. It’s a splurge, but truly worth it if you love a cozy-soft feel and want sheets that feel cool to the touch at the same time (but warm up nicely as you sleep).

Ettitude Signature Sateen Bamboo Sheets for $333: These are an impressively decorated set of bamboo sheets, with certifications from Oeko-Tex, the Forest Stewardship Council, Eco-Cert, and B-Corp. Bamboo doesn’t have the same organic certifications yet that you can find on linen and cotton, but this is the closest combo you’ll get. These sheets are also super soft yet still very breathable, and my favorite bamboo for hot sleepers and organic shoppers.

Sheets to Skip

Amazon Basics Microfiber Sheet Set for $20: We wanted to like these cheap sheets made by Amazon, but this set is a major skip. This set both felt and sounded like the paper robes you wear at the doctor’s office, and our testers hardly managed sleeping on these for a single night.

Coop Comphy Cool+ Sheet Set for $219: This set promises cooling properties, but I found myself waking up sweating after sleeping on these, which isn’t a huge surprise after seeing they’re made of nylon, polyester, and spandex, none of which are as breathable as simple linen or cotton.

Rest Evercool+ Cooling Starter Sheet Set for $299: These are supposed to be cooling sheets, but I found myself waking up feeling moisture trapped around me. Whatever’s in Rest’s proprietary cooling fabric didn’t work for me at all.

Power up with unlimited access to WIRED. Get best-in-class reporting and exclusive subscriber content that’s too important to ignore. Subscribe Today.