Current Evidence on Programme Maturity
On physical manufacturing evidence, XPeng has advanced beyond Tesla by commissioning a dedicated assembly facility for the IRON biped. XPeng opened an assembly line in Guangzhou on 7 September 2026 with over 80 percent of core assembly processes automated, as reported by Electrek. Tesla has stated an intention to begin limited production of its Optimus robot inside its Fremont vehicle facility, but Tesla has not publicly commissioned a dedicated humanoid manufacturing line.
The verification gap between the two projects stems from visible hardware versus forward-looking statements. As of 10 August 2026, Tesla had not publicly shown its third-generation Optimus model, previously expected in 2025 and delayed across multiple projected windows in early 2026. XPeng demonstrated its production-intent IRON unit walking in public, followed by factory documentation of robots moving along an active line. Buyers assessing commercial timelines find tangible tooling in Guangzhou, whereas Tesla continues internal iteration without an official public spec sheet.
Neither manufacturer provides an immediate commercial solution for warehouse operators or factory managers today. Neither robot is buyable. There is no preorder for either. Tracking both programmes through our wider humanoid robot comparison index indicates that while manufacturing footprints differ, both developers remain focused on captive internal deployments before external deliveries begin.
XPeng IRON vs Tesla Optimus Specification Matrix
Direct specification comparisons between XPeng IRON and Tesla Optimus reveal a distinct divide between verified public disclosures and unofficial circulating figures. XPeng claims 76 degrees of freedom (DOF) across the chassis, 21 per hand. Three in-house Turing Artificial Intelligence (AI) chips supply the compute, which XPeng puts at up to 2,250 Tera Operations Per Second (TOPS). Tesla has not published an official specification sheet for its third-generation Optimus, leaving circulating metrics regarding its AI5 processor and tendon-driven hand unconfirmed by Tesla.
Examining the official XPENG IRON official page reveals that XPeng omits physical dimensions, mass, battery capacity, runtime, and pricing, stating the robot is provided strictly for technical demonstration purposes. Unofficial online databases list arbitrary figures such as a 173-centimeter height and a 70-kilogram weight for IRON, yet those aggregator listings lack citations and contradict verified announcements. Tesla similarly maintains no public engineering sheet on its Tesla AI and robotics portal, making direct physical comparisons dependent on company announcements rather than customer data sheets.
Neither machine is an off-the-shelf industrial asset. Buyers investigating robotic hands or chassis mobility must judge XPeng and Tesla on verifiable milestone deliveries rather than unconfirmed specification leaks.
| Factor | XPeng IRON | Tesla Optimus |
|---|---|---|
| Public hardware status | Next-generation hardware demonstrated publicly | Third-generation hardware not publicly shown |
| Dedicated production line | Commissioned in Guangzhou (>80% automated) | Stated intention at Fremont; uncommissioned |
| Targeted production date | Mass production targeted end of 2026 | Limited production timeline unconfirmed |
| Total body degrees of freedom | 76 DOF stated by XPeng | Unofficial reports suggest iterative redesign |
| Hand degrees of freedom | 21 DOF per hand stated by XPeng | 22 DOF per hand reported, unconfirmed |
| Onboard compute platform | Three in-house Turing AI chips (2,250 TOPS) | Reported AI5 chip, unconfirmed by Tesla |
| Physical height and weight | Not published officially by XPeng | ~173 cm, ~57 kg reported, unconfirmed |
| Public price and purchasing | Not purchasable; no public price | Not purchasable; no public price |
Public Demonstration Verification History
Public hardware verification tests distinguish physical electromechanical machinery from digital video presentations. XPeng unveiled the next-generation IRON at its 2025 AI Day in Guangzhou, directing the robot down a presentation runway with an organic gait. Because the movement appeared remarkably fluid, online observers asserted that XPeng had placed an actor inside a suit. Roughly 24 hours later, XPeng Chief Executive Officer (CEO) He Xiaopeng returned to the stage with scissors and cut open the flexible lattice leg skin. Underneath were actuators, metal joints and wiring. The robot kept walking, as documented by New Atlas.
Chief Executive Officer Elon Musk noted that third-generation hardware delays stem from iterative engineering, emphasizing that the engineering group is getting close to a production-ready design while navigating hand complexity. Tesla has not exhibited its third-generation platform walking untethered in an interactive public setting without environmental controls.
Cutting the skin open on stage settled whether the machine was real. The scissor demonstration confirmed that XPeng engineered a multi-jointed physical skeleton rather than a human illusion. For readers following broader industry progress in our Chinese humanoid robots analysis, physical demonstrations provide necessary baseline evidence, though walking on flat stage flooring remains distinct from real manufacturing labor.
Unproven Industrial Performance Metrics
Neither XPeng nor Tesla has proven that its bipedal robot can perform sustained autonomous industrial labor. XPeng asserts that IRON operates autonomously without human teleoperation, using spatial perception models adapted from its electric vehicle division. However, XPeng has not published multi-hour reliability logs, mean time between failures (MTBF), or task success rates in industrial settings. Showing a biped walking down a runway or traversing an assembly line does not prove the machine can assemble automotive components for an entire shift.
Tesla faces an identical verification threshold regarding autonomous operation. Until Tesla publishes unedited, continuous logs of Optimus completing manipulation tasks inside the Fremont vehicle facility across hundreds of cycles, claims of functional utility remain unproven.
Operating expense and component durability represent additional unknowns across both platforms. Humanoid actuators experience sustained thermal stress, mechanical fatigue, and gear wear during cyclic industrial operations. Neither manufacturer publishes expected actuator lifespans, joint replacement intervals, or battery pack runtime during continuous locomotion. Buyers reviewing humanoid robots in warehouses must recognize that until MTBF metrics are verified in multi-month pilots, claims regarding labor displacement remain speculative.
Dedicated Assembly Lines Versus Stated Intentions
XPeng has commissioned tooling. Tesla has stated an intention. On 7 September 2026, XPeng commissioned an automated assembly line in Guangzhou designed specifically for IRON production. He Xiaopeng stated that the assembly lines were created from scratch with no precedent to follow, achieving an automation rate exceeding 80 percent across core manufacturing processes, as detailed by XPENG official news.
XPeng outlined explicit factory scaling milestones backed by corporate funding. XPeng Robotics raised over $900 million in August 2026 at a $6.3 billion valuation, led by IDG Capital with Alibaba and Tencent participating, as reported by Electrek. Mass production is targeted for the end of 2026. The initial build run targets internal operations in retail stores and corporate campuses, progressing to commercial customer deliveries in 2027. XPeng established a monthly target exceeding 1,000 units, with a projected volume approaching one million robots annually by 2030.
Tesla has taken an alternate industrial approach by integrating early development within vehicle facilities. Rather than erecting a purpose-built robotics facility with automated tooling, Tesla has discussed piloting initial assembly at its vehicle plant in Fremont, California. While Tesla possesses extensive automotive manufacturing automation, it has not presented completed industrial tooling designed specifically to produce bipedal chassis at scale. Building custom robotic arms and hand assemblies in volume requires bespoke fixtures that Tesla has not yet revealed to the public.
Evaluation Milestones for Commercial Robotics Buyers
Commercial robotics buyers must evaluate verified factory integration milestones rather than promotional stage presentations. Because neither machine is available for commercial purchase, operators must track verifiable milestone indicators. For XPeng, the primary validation signal is whether the Guangzhou plant meets its target of initial mass assembly by the end of 2026. If XPeng places units into customer-facing retail stores and publishes cycle times, buyers gain verifiable operational data regarding uptime and customer interactions.
For Tesla, the primary milestone is an official unveiling of the third-generation hardware accompanied by an audited specification sheet. Enterprise operators tracking Tesla must watch for the formal release of actuator specifications, hand degrees of freedom, and unassisted runtimes inside the Fremont facility. Until Tesla exhibits the robot handling real automotive subassemblies without remote teleoperation, procurement teams cannot evaluate its commercial viability relative to alternatives covered in our Chinese humanoid robots guide.
Capital commitment provides another metric for assessing long-term supplier stability. XPeng Robotics operates with dedicated capital reserves following its $900 million private financing round, shielding the humanoid effort from immediate passenger vehicle balance sheet pressures. Commercial planners must evaluate whether robotics divisions possess the runway to support long-term hardware iterations, spare parts distribution, and field maintenance before considering pilot commitments.
Conditions That Would Shift the Assessment
Specific physical demonstrations and factory milestones would immediately reverse the current comparative assessment between XPeng and Tesla. The current assessment favors XPeng on physical manufacturing readiness due to its operational Guangzhou production facility. However, that advantage would evaporate if XPeng encounters assembly bottlenecks, actuator supply deficits, or software failures that delay its targeted end-of-2026 manufacturing milestone or disrupt planned 2027 commercial deliveries. Commissioning an 80 percent automated line does not guarantee high yields or reliable finished units.
Tesla would immediately alter the assessment by publicly demonstrating its third-generation Optimus executing continuous, untethered manipulation tasks at the Fremont facility. Tesla would take the lead by pairing a working public demonstration with an audited engineering sheet confirming mass production tooling at Fremont. Fully autonomous material transport without teleoperation would count for more than a runway walk.
Pricing disclosures and export accessibility also represent decisive variables for global purchasers. If Tesla announces an enterprise pilot pricing model with domestic service guarantees, domestic commercial buyers will prioritize Tesla over foreign hardware subject to international trade barriers. Readers investigating equipment investments can explore our humanoid robot cost analysis to examine how platform pricing structures dictate return on investment across commercial robotics programs.
Bottom Line
Procurement directors should not issue purchase orders or attempt contract negotiations for either the XPeng IRON or Tesla Optimus in 2026, as neither robot has an active sales pipeline, certified pricing, or proven autonomous work history. XPeng holds the current advantage in verified manufacturing readiness due to its operational, automated production line in Guangzhou and live physical verification tests. Tesla has not shown third-generation production hardware or published a spec sheet for it. This assessment will shift the moment Tesla demonstrates verified, autonomous factory labor, or if XPeng fails to achieve its end-of-2026 mass production timeline.
Compare technical specifications and deployment timelines across commercial competitors in our complete humanoid robot comparison.
FAQs
Is the XPeng IRON better than Tesla Optimus?
Neither machine can be labeled better overall because neither is available for public purchase or deployed in verified commercial labor. XPeng has shown more physical evidence: it commissioned an automated assembly line in Guangzhou, and its robot walked on stage. Tesla has not publicly unveiled its third-generation hardware or published an official specification sheet. Match evaluation to verified physical milestones.
Which is further along, XPeng IRON or Tesla Optimus?
XPeng is further along in physical production infrastructure. On 7 September 2026, XPeng commissioned an assembly line in Guangzhou with over 80 percent automated core processes and confirmed plans for end-of-2026 mass production. Tesla has discussed limited production at Fremont but has not publicly commissioned a dedicated humanoid manufacturing line, and as of 10 August 2026, Tesla had not shown its third-generation Optimus hardware.
Can you buy either the XPeng IRON or Tesla Optimus?
No, neither the XPeng IRON nor Tesla Optimus can be purchased today. Neither company offers a public checkout, preorder list, or commercial delivery contract. XPeng lists its robot for technical demonstration purposes only, targeting internal deployments before commercial sales begin in 2027. Tesla has announced no release date, pricing, or preorder system for its third-generation Optimus.
How many degrees of freedom does each robot have?
XPeng claims 76 degrees of freedom across the IRON body plus 21 degrees of freedom in each hand. Tesla has not published official degrees of freedom for its third-generation Optimus. Widely circulated reports suggest 22 degrees of freedom per hand across 50 total hand actuators for Tesla, but those figures remain unconfirmed until Tesla publishes an official engineering specification sheet.
Which one will be cheaper?
Neither manufacturer has announced an official price, making direct cost comparisons speculative. XPeng's official portal publishes no pricing data, while third-party databases list unverified estimates of $150,000 without citations. Tesla has published no pricing for its third-generation Optimus. Total acquisition costs will ultimately depend on volume manufacturing scale, actuator longevity, warranty terms, and regional import duties rather than early speculative estimates.
When will Tesla Optimus actually be shown?
Tesla has not announced a confirmed public release date for its third-generation Optimus hardware. The unveiling was anticipated in 2025, then pushed to early 2026 and late July or August 2026. Chief Executive Officer Elon Musk stated that the engineering team is getting very close to a production-ready design, attributing delays to the complexity of the biomimetic hand mechanism.