The T800 is available in four editions ranging from approximately $25,000 (Basic Edition, RMB 180,000) to $50,000 (Max Edition, RMB 360,000). The Basic Edition includes the full 29-DOF body and solid-state battery but lacks dexterous hands and is not open-source. The Eco Edition at $33,000 adds dexterous 5-finger hands and open-source support. The Pro Edition at $38,500 upgrades to an NVIDIA Thor chip. The Max Edition at $50,000 features industrial-grade long-life components designed for continuous deployment. A $700 (RMB 5,000) deposit secures a pre-order on JD.com.

As of March 2026, the T800 is available for pre-order exclusively through JD.com in China. There is no official North American distributor, authorized reseller, or confirmed shipping date for the US or Canada. Purchasing from China involves import logistics, potential tariffs, and no local warranty or service support. BotInfo.ai is tracking EngineAI's international expansion plans and will update this page as new distribution channels are announced.

EngineAI's JD.com listing states that first shipments will begin no later than June 2, 2026. This aligns with the company's stated roadmap to begin "scenario-based verification and large-scale deployment" in 2026. Mass manufacturing is still ramping up, so early units may ship in limited quantities before broader availability later in the year.

As of early 2026, Tesla's Optimus remains in an R&D and internal testing phase — CEO Elon Musk confirmed on a recent earnings call that Optimus humanoids are not yet commercially available. Tesla has committed $20 billion in 2026 capex partly for Optimus production, but no public pricing, shipping dates, or commercial deployment details have been confirmed. The T800, by contrast, has published pricing ($25K–$50K), a JD.com listing, and a June 2026 shipping target. On raw specs, the T800's 450 Nm torque and solid-state battery are competitive, though direct comparison is difficult given Tesla's limited public disclosures. Tesla's advantage lies in its manufacturing scale and AI ecosystem (FSD data pipeline, Dojo); EngineAI's advantage is being commercially available now.

The T800's base compute combines an Intel N97 CPU with an NVIDIA AGX Orin module (64GB), delivering 275 TOPS of AI computing power. The Pro and Max Editions upgrade to the NVIDIA Jetson Thor, which delivers approximately 2,000 TOPS — a significant leap that enables more sophisticated autonomous behaviors, faster perception processing, and support for advanced AI models. This tiered compute approach lets buyers match their investment to their actual deployment complexity.

Only the Eco Edition ($33,000) is open-source and supports secondary development. The Basic Edition is closed-source. The Pro and Max editions include more advanced hardware but EngineAI has not confirmed full open-source access for those tiers. For the UKRL combat league, EngineAI has indicated they will open-source the robot's code for customization and training purposes, which could benefit the broader developer ecosystem.

The T800 uses what EngineAI describes as the industry's first solid-state power battery designed specifically for humanoid robots. Solid-state batteries replace the liquid electrolyte in traditional lithium-ion packs with a solid material, offering theoretical advantages in energy density, safety (lower fire risk), and cycle life. This gives the T800 a claimed 4–5 hour runtime — roughly double what the Unitree H2 achieves. The battery is also modular and hot-swappable, enabling extended operation shifts. However, solid-state battery technology is still emerging at commercial scale, and long-term reliability under high-torque humanoid workloads remains unproven.

The Ultimate Robot Knockout Legend (UKRL) is the world's first professional combat league for full-size humanoid robots. Launched in February 2026 in Shenzhen by EngineAI, it features 16 teams competing with identical T800 robots (provided free of charge) through a year-long tiered season. Matches follow a Best-of-Three format with 5-minute rounds. The winning team receives a gold championship belt valued at approximately $1.44 million (10 million yuan). The league is open to universities, enterprises, and research institutions worldwide. It serves as both an entertainment spectacle and an extreme R&D testing ground for humanoid durability, balance, and autonomous decision-making.

Yes — the Eco Edition ($33,000) is specifically designed for research and educational use. It includes dexterous hands and open-source support for secondary development, making it suitable for universities and R&D labs. The UKRL league also provides T800s free of charge to qualifying teams, which could be an alternative path for academic institutions looking to access the platform. For smaller-budget educational needs, EngineAI's PM01 compact humanoid (initially under $15,000) may be a more appropriate entry point.

No. When EngineAI released the original T800 demo in December 2025, the footage showed such fluid martial arts movements that many viewers online accused it of being computer-generated. EngineAI responded by releasing raw behind-the-scenes footage from the shoot — same movements, no color grading or dramatic editing — and CEO Zhao Tongyang posted a video in which the T800 delivered a padded kick to his chest, knocking him to the ground. The company states that the demonstrations used "no CGI, no AI, no video speed-up." The T800's 450 Nm torque and dynamic stability algorithms are the engineering basis for the movements seen in the videos.

The T800 and Unitree H2 are the two leading full-size Chinese humanoids competing head-to-head in 2026. The T800 is heavier (75 kg vs ~47 kg), more powerful (450 Nm vs ~360 Nm torque), and has significantly longer battery life (4–5 hours vs ~2 hours) thanks to its solid-state battery. The T800 is also cheaper at $25,000 vs the H2's ~$29,900. However, the Unitree H2 benefits from Unitree's broader ecosystem (G1, R1), its approaching IPO momentum, and a more established track record in commercial deployments. Unitree has also partnered with UFC and hosted its own humanoid kickboxing tournaments, making the rivalry between the two companies intense and highly visible.

EngineAI positions the T800 for industrial production lines, logistics, commercial services (retail "Cyber Staff"), inspection, material handling, and entertainment. Some T800 units have been placed in retail stores in Shenzhen. However, like all humanoid robots in 2026, the gap between demo capabilities and daily operational reliability is significant. The T800's martial arts demonstrations showcase its hardware limits, but sustained autonomous task performance in unstructured environments remains an industry-wide challenge. Buyers should evaluate the T800 as a capable but still-maturing platform.

EngineAI was founded by Zhao Tongyang in October 2023. Zhao has a long background in robotics and IoT — he started in IoT in 2012, invested in bipedal robots by 2016, and in 2018 co-founded Pengxing Intelligent Robotics alongside XPeng Motors, one of China's largest EV companies. After taking patents and IP with him, he launched EngineAI and built it to over 110 employees with $140+ million in funding, anchored by JD.com as lead investor.

EngineAI highlights the T800's 5% force control accuracy — meaning the robot can precisely regulate how much force it applies, distinguishing between a padded kick and a handshake. The company also publishes safety disclaimers noting that "due to the complex structure and extremely powerful actuators, users must maintain a safe distance and exercise extreme caution during operation." The humanoid robotics field as a whole is in an early-stage exploration phase, and EngineAI encourages buyers to fully understand the technology's current limitations before purchasing. The T800 is designated for civilian use; any hazardous modifications or misuse is strictly prohibited.

That depends on your use case, risk tolerance, and support requirements. The T800 offers the best raw specs-per-dollar of any full-size humanoid in 2026 — $25K for 450 Nm torque and a solid-state battery is remarkable. However, US-based alternatives like Apptronik Apollo and Figure 02 come with domestic support infrastructure, partnerships with Fortune 500 companies (Mercedes, BMW, GXO), and access to Western AI ecosystems (Google DeepMind, NVIDIA GR00T). Chinese humanoids like the T800 carry import risks, tariff exposure, and service limitations. For R&D and evaluation, the T800's price-to-capability ratio is hard to beat. For production deployment in North America, US platforms may offer lower total cost of ownership when factoring in support and reliability.