Apptronik

Table of Contents

⚠ Disclaimer: This entry may be incomplete, out of date, or inaccurate. It is AI-maintained on a best-effort basis. Do not rely on it as a sole source — verify claims independently using the sources listed below.

Summary

Apptronik is an Austin-based humanoid robotics company spun out of the Human Centered Robotics Lab at UT Austin, with deep roots in NASA’s Valkyrie humanoid program. Its Apollo robot targets industrial manufacturing and logistics. The company raised $935M total across a multi-tranche Series A in 2025–2026, co-led by B Capital and Capital Factory with Google participation, at a ~$5B post-money valuation. Investors include Mercedes-Benz, John Deere, Qatar Investment Authority, and AT&T Ventures. On July 1, 2026, Apptronik opened an expanded, nearly 90,000-square-foot “Robot Park” data-collection and training facility in Austin, running fleets of its current Apollo 2 platform (in both bipedal and wheeled-base configurations) to generate training data for Google DeepMind’s Gemini Robotics models. Apollo 2 is explicitly framed as a data-collection and pre-commercial platform; the first mature commercial product, Apollo 3, is targeted for release in 2027. Apollo’s hot-swappable battery (4-hour runtime, swap in under 5 minutes) remains a key differentiator. Commercial partners with active Apollo 2 deployments include Mercedes-Benz, GXO Logistics, and Jabil.

Key Facts

Founded: 2016 (UT Austin Human Centered Robotics Lab spinout)
HQ: Austin, TX
Type: Company — Platform OEM
Status: Active; late pre-commercial / early commercial
Employees: ~300 (as of February 2026)
Total funding: ~$935M
Valuation: ~$5.3B post-money (February 2026 round)
Robot: Apollo — 5'8" (~1.73m), ~160 lbs (~73 kg), 25 degrees of freedom, 55 lb (25 kg) payload
Current platform: Apollo 2 — modular bipedal + wheeled-base configurations; data-collection/pre-commercial platform, in production over a year as of mid-2026
Next platform: Apollo 3 — first mature commercial product, targeted for release in 2027
Training infrastructure: Robot Park — ~90,000 sq ft flagship facility, Austin, TX (expanded facility opened July 1, 2026); additional Robot Parks at Google DeepMind and customer sites (Mercedes-Benz, GXO)
Battery: 4-hour runtime; hot-swappable in under 5 minutes
Actuator technology: UT Series Elastic Actuator (SEA), invented by Nicholas Paine
Value chain position: Platform OEM; actuator design in-house (SEA), components sourced
Commercial partners (active deployments): Mercedes-Benz, GXO Logistics, Jabil
Key investors: B Capital (lead), Capital Factory (co-lead), Google, Mercedes-Benz, John Deere, Qatar Investment Authority, AT&T Ventures, PEAK6
Key executives: Jeff Cardenas (CEO & co-founder), Barry Phillips (Chief Commercial Officer)

Funding & Investors

Round	Date	Amount	Lead / Notable Investors
Pre-Series A (NASA grants)	2016–2024	$5M+	NASA / government contracts
Series A initial	Feb 2025	$350M	B Capital (co-lead), Capital Factory (co-lead), Google (participant), Mercedes-Benz, GXO Logistics
Series A extension	Mar 2025	$53M	Oversubscribed close; brings initial round to $403M
Series A-X extension	Feb 2026	$520M	Existing investors including B Capital, Google, Mercedes-Benz; new: AT&T Ventures, John Deere, Qatar Investment Authority (QIA)

Total raised: ~$935M+ across the full Series A. The Feb 2025 press release confirmed 150 employees at close; ~300 by Feb 2026.

Lead investors: B Capital (Howard Morgan, Chair; co-led initial) and Capital Factory (Gordon Daugherty, Chairman of Apptronik board) — not Google. Google participated in the initial tranche and joined the extension, but was not the lead.

Google DeepMind connection: Google’s participation across multiple tranches is widely interpreted as DeepMind seeking to ground its physical AI research in a hardware platform. Apptronik established a formal strategic partnership with Google DeepMind before the funding round. This makes Apptronik a potential hardware partner in Google’s foundation-model-for-robotics strategy.

Qatar Investment Authority: QIA is a sovereign wealth fund with $475B+ in AUM. Their participation signals interest in diversifying into frontier technology through a US ally’s manufacturing platform — and potentially Middle Eastern deployment of industrial robots.

Mercedes-Benz: Both investor and potential customer. Mercedes has publicly stated interest in testing humanoids for small-batch and precision assembly tasks at automotive plants — the same manufacturing context as GXO/Toyota are for Agility.

John Deere: Agricultural equipment manufacturer. Their investment is the most unusual — suggests interest in humanoids for farm equipment assembly or eventually field operations, although the latter is speculative.

What It Is / How It Works

Apptronik evolved from UT Austin’s Human Centered Robotics Lab, which contributed significantly to NASA’s Valkyrie humanoid (developed 2013–2015 for DARPA’s Robotics Challenge). The Apollo robot inherits Valkyrie’s emphasis on safe human co-location, modular design, and the Series Elastic Actuator technology invented by Nicholas Paine during his PhD work.

UT Series Elastic Actuators: Nicholas Paine’s doctoral dissertation produced a class of SEA designs specifically optimized for full-body humanoid robots. The UT SEA is designed around backdrivability (the robot can be pushed away in an emergency), torque sensing via spring deflection, and shock tolerance. Like Agility’s Digit, Apollo prioritizes safe human interaction over maximum speed or stiffness — the right tradeoff for factory floor co-working. See UT Series Elastic Actuator (UT-SEA) for a full technical deep dive on the design’s origins, ball-screw/RFSEA architecture, and Apptronik’s resulting patent family.

Apollo specifications: 5'8" tall (~1.73m), ~160 lbs (~73 kg), 25 degrees of freedom, 55 lb (25 kg) payload capacity, 4-hour battery runtime with hot-swappable batteries (swap in under 5 minutes). The 4-hour runtime is a significant differentiator over Agility Digit’s 90-minute limit; hot-swap means a single charging station can sustain near-continuous operation across shifts without the robot going offline. The 55 lb payload is meaningfully higher than Digit’s tote-transport capacity, positioning Apollo for heavier manufacturing tasks (automotive parts, equipment assembly).

AI integration with Google: The Google investment is likely coupled with access to Google DeepMind’s Gemini Robotics foundation models. Apptronik would use these models to accelerate task learning across manufacturing environments, reducing the teleoperation data collection burden that limits competitors. This is a significant potential competitive advantage if DeepMind’s physical AI models mature.

NASA Valkyrie lineage: Apptronik’s engineers built, operated, and maintain NASA Valkyrie units at multiple research institutions. This gives the team more full-humanoid integration experience than any other humanoid company except Boston Dynamics — the experience of actually running a full humanoid at scale in research environments is not trivial. See Valkyrie (NASA R5) for a full profile of the platform, its 2013 DARPA Robotics Challenge history, and its ongoing NASA research program.

Robot Park and Apollo 2 (July 2026): Apptronik opened an expanded, nearly 90,000-square-foot “Robot Park” data-collection and training facility in Austin on July 1, 2026, and unveiled Apollo 2 — the current-generation platform, offered in both bipedal and wheeled-base configurations. Apollo 2 has reportedly been running in Apptronik’s facilities and at customer sites for more than a year prior to this public unveiling, and the company says it has already produced more Apollo 2 units than its 2023 predecessor. Apollo 2 is explicitly positioned as a data-collection and training platform rather than a finished commercial product: it generates teleoperation and autonomous-execution data at Robot Park and at partner sites (including a Robot Park at Google DeepMind itself, plus Mercedes-Benz and GXO facilities) to train Google DeepMind’s Gemini Robotics foundation models. The wheeled configuration is designed to meet existing industrial mobile-robot safety standards for faster near-term deployment; the bipedal configuration is used to continue refining walking safety and reliability. CEO Jeff Cardenas has said the company deliberately withheld a public reveal of this generation for roughly a year, wanting to show only capabilities it could reliably demonstrate live rather than in edited demo videos.

Apollo 3 (targeted 2027): Apptronik describes Apollo 3 as its first mature commercial product, intended to launch with substantially more built-in embodied intelligence than Apollo 2 by leveraging the data collected through Robot Park and the Gemini Robotics partnership. As of mid-2026, Apollo 3 has not been publicly demonstrated; Cardenas has indicated multiple incremental Apollo versions will appear “in pretty quick sequence” ahead of it. Per this knowledge base’s convention for undemonstrated products, Apollo 3’s status should be treated as Announced / Undemonstrated until a public hardware demo occurs.

Notable Developments

2026-07-01: Opened expanded ~90,000 sq ft Robot Park facility in Austin, TX; unveiled Apollo 2 (bipedal + wheeled configurations) as the data-collection/training platform feeding Google DeepMind’s Gemini Robotics models.
2026-02-11: Closed $520M Series A-X extension (new investors AT&T Ventures, John Deere, Qatar Investment Authority), bringing total Series A to over $935M at a ~$5B+ valuation.
2025-03: Closed $53M oversubscribed Series A extension, bringing initial round to $403M.
2025-02-13: Announced $350M Series A co-led by B Capital and Capital Factory, with Google, Mercedes-Benz, and GXO Logistics participating.

Founder Backgrounds

Jeff Cardenas — Co-founder & CEO

M.S., Technology Commercialization, UT Austin
Background in management consulting (Deloitte) before joining the robotics lab
Bridges research-to-commercial translation; led Apptronik’s pivot from contract R&D work to product company
LinkedIn: jeffcardenas

Nicholas Paine — Co-founder & CTO

Ph.D., Electrical and Computer Engineering, UT Austin
Invented the UT Series Elastic Actuator during his dissertation — the core actuator technology in Apollo
Previously: research engineer at NASA Johnson Space Center contributing to Valkyrie hardware
One of the small number of people globally who have designed, built, and operated a full-scale humanoid
LinkedIn: nickpaine

Bill Helmsing — Co-founder

Ph.D. or advanced degree in mechanical/robotics, UT Austin affiliation
Mechanical design lead

Bill Welch — Co-founder

UT Austin Human Centered Robotics Lab
Controls and systems engineering

Prof. Luis Sentis — Co-founder

Director, Human Centered Robotics Lab, UT Austin (founding academic)
Ph.D. from Stanford; decades of whole-body humanoid control research
Author of foundational papers on whole-body motion control that underpin modern humanoid designs
LinkedIn: luis-sentis

Pattern: The strongest academic foundation of any US humanoid startup. Five co-founders with PhDs and genuine full-humanoid construction experience. The presence of Prof. Sentis as a co-founder (rather than just an academic advisor) gives Apptronik access to ongoing research output and talent pipeline from UT Austin.

People — Last Reviewed: 2026-07-18

Supply Chain & Dependencies

Actuators (40–60% of BoM): UT SEA design; components (motors, springs, bearings, gearboxes) sourced externally. Motors likely from Maxon (Swiss, premium) or Kollmorgen; precision bearings potentially from Schaeffler or Japanese suppliers. Unlike Figure AI, Apptronik does not claim full actuator vertical integration.

Compute: Not publicly specified. Given Google investor/partner status, Tensor Processing Units (TPUs) for cloud training and NVIDIA Jetson-class for edge inference are the likely stack. Google may provide compute infrastructure as part of the partnership.

Sensors: Camera + depth sensor head; lidar not mentioned. Likely RGB-D sensors from Luxonis or equivalent; proprioceptive sensing via SEA spring deflection.

Battery: 4-hour runtime; hot-swappable packs replaceable in under 5 minutes. This is Apollo’s most important operational specification — it enables near-continuous shift operation with a small bank of spare batteries, unlike competitors requiring full recharging downtime. Cell chemistry and supplier not disclosed.

Manufacturing: Not discussed publicly. No BotQ-equivalent announced. Production scale is small; commercial scale-up would require a manufacturing partner (Mercedes-Benz or another strategic investor could provide this).

Supply chain geography: No explicit China decoupling statement. Given US manufacturing ambitions and the John Deere/Mercedes-Benz investor base, domestic supply chain is likely a priority but not yet documented.

Claim Verification

Claim: Apollo will reach commercial manufacturing deployment at scale

Status: Unverified (forward-looking)

Supporting sources:

$935M in funding with Mercedes-Benz and John Deere as strategic investors signals near-term deployment intent
Google DeepMind partnership suggests access to foundation AI models that could accelerate task generalization

Refuting / qualifying sources:

Active partner deployments at Mercedes-Benz, GXO Logistics, and Jabil confirmed, but described as “early versions working within designated areas” — scope and productivity metrics not disclosed
Agility Robotics has more established multi-customer US deployments with less funding and a simpler robot
Apollo’s 25-DoF complexity is higher than Digit, making reliability validation harder

Summary: Strong institutional backing and credible team. No verified commercial deployment yet. Timeline to commercial scale is uncertain.

Claim: Apollo 3 will be Apptronik’s first mature commercial product, targeted for 2027

Status: Unverified (forward-looking, company statement)