Industry News
Amazon Bought a Humanoid Startup Just 59 Days Old
Amazon has acquired Fauna Robotics, a New York City startup whose 107 cm humanoid Sprout launched to developers just 59 days before the deal closed. The acquisition — Amazon's second robotics purchase in March 2026 — brings a team of former Google DeepMind and Meta engineers into Amazon's Personal Robotics Group and signals the company's push into consumer humanoid robotics.
Amazon has acquired Fauna Robotics, a two-year-old New York City startup whose 107 cm, 22.7 kg humanoid robot named Sprout launched to developers just 59 days before the deal closed. The acquisition, confirmed on March 24, 2026, marks Amazon's second robotics purchase in the same month — following its earlier acquisition of stair-climbing delivery startup Rivr — and signals the company's intent to compete directly in consumer and developer humanoid robotics.
Terms were not disclosed, but the roughly 50-person Fauna team, including co-founders Rob Cochran (former Head of Product at CTRL-labs/Meta) and Josh Merel (former Research Scientist at Google DeepMind), will join Amazon's robotics division in New York City. Fauna had raised approximately $30 million from Kleiner Perkins before the deal. Early Sprout customers included Disney, Boston Dynamics, UC San Diego, and NYU.
AI-generated image
Sprout packs a stereo camera, four time-of-flight sensors, and 29 degrees of freedom into a 3.5-foot form factor designed for shared human spaces. Source: AI-generated illustration.
Key Stats
107 cm
Sprout Height
29 DOF
Degrees of Freedom
$30M
Pre-Acq. Funding
2nd
Amazon Robotics Acq. in March
The Sprout Story: Two Years, One Robot, One Acquisition
Fauna Robotics was founded in 2024 by veterans of CTRL-labs, Meta, and Google DeepMind with a specific and contrarian premise: the first mass-market humanoid robot won't look like an adult human. Sprout stands 3 feet 6 inches tall, weighs just 50 pounds, and was designed from the start to be safe around children, pets, and elderly adults in unstructured home environments. The size is a deliberate engineering choice — lower center of gravity, less inertia in collisions, and a visual profile that reads as approachable rather than threatening.
The Sprout Creator Edition, launched January 27, 2026, is explicitly a developer platform. At $50,000 per unit, it targets research labs, technology companies, and entertainment studios rather than household consumers. Early buyers received a robot capable of walking, dancing, kneeling, crawling, and waving — plus a full SDK for building autonomous behaviors. Disney, Boston Dynamics, UC San Diego, and NYU were among the first customers, a list that spans entertainment, robotics research, and academia.
Less than two months after launch, Amazon had acquired the company. The speed of the deal reflects both how competitive the talent market has become and how aggressively Amazon is moving to establish a position in personal robotics before its competitors.
🤝 Why Amazon Moved Fast
Fauna's co-founder Josh Merel spent years at Google DeepMind building AI models for robot locomotion. Amazon acquired those research relationships along with the hardware — and pulled a key AI talent away from a direct rival's orbit.
📦 Two Acquisitions in One Month
Amazon confirmed the Rivr stair-climbing delivery robot acquisition on March 19, then closed Fauna five days later. The back-to-back deals suggest a coordinated push into both delivery robotics and consumer/developer humanoids simultaneously.
Under the Hood: Sprout's Hardware and AI Stack
Sprout's hardware reflects its design philosophy. The 29 degrees of freedom — split across 6-DOF arms, 5-DOF legs, 2-DOF neck, active eyebrows, and single-DOF rubberized grippers — prioritize expressive motion and safe interaction over industrial payload capacity. The soft-to-touch exterior minimizes pinch points throughout the body. The legs are designed for a low center of gravity with compliant walking gaits that yield to external forces rather than pushing through them.
On the compute side, Fauna chose NVIDIA Jetson AGX Orin (64 GB) paired with 1 TB of onboard SSD. The sensing stack centers on a ZED 2i stereo camera for depth perception plus four time-of-flight sensors covering the body's blind spots, a four-microphone array for directional voice interaction, and a torso-mounted IMU for balance. The combination gives Sprout full-room spatial awareness without the LiDAR complexity found in larger industrial platforms.
Battery life runs 3 to 3.5 hours on a single swappable pack — roughly in line with larger humanoids despite the smaller form factor. The SDK includes modular autonomy layers: navigation, mapping, localization, and human awareness are available as building blocks developers can stack rather than starting from scratch.
AI-generated image
Amazon's fulfillment infrastructure — already one of the most automated in the world — becomes the testing ground for a consumer humanoid platform. Source: AI-generated illustration.
Sprout vs. the Field: Hardware Comparison
| Spec | Fauna Sprout | Figure 02 | Tesla Optimus Gen 2 | Agility Digit |
|---|---|---|---|---|
| Height | 107 cm (3 ft 6 in) | 1.70 m | 1.73 m | 1.75 m |
| Weight | 22.7 kg (50 lb) | 60 kg | 57 kg | 65 kg |
| DOF | 29 | 65+ | 28 | 30 |
| Onboard Compute | NVIDIA Jetson AGX Orin 64GB | Custom AI chip | Tesla FSD hardware | Intel NUC |
| Battery Life | 3–3.5 hrs (swappable) | ~4 hrs work | ~2–4 hrs | ~4 hrs |
| Target Market | Developer / Consumer | Industrial (Deployed) | Industrial (Expanding) | Logistics (Deployed) |
| Price (2026) | $50,000 | Undisclosed | ~$20,000–$30,000 (target) | RaaS pricing |
Design Philosophy: Safety-First Scaling
• Tiered safety system: Sprout yields to external forces on contact rather than holding position, reducing injury risk in unstructured environments.
• Low center of gravity: At 107 cm, Sprout's mass distribution reduces tip-over momentum compared to full-height humanoids.
• Soft exterior: Engineered to minimize pinch points and impact force — critical for deployments around children and pets.
• Modular autonomy: Navigation, mapping, and human awareness are separate SDK modules — developers can add capabilities incrementally rather than deploying full autonomy on day one.
The Founders and the Talent That Amazon Bought
The acquisition value isn't just the Sprout hardware — it's the team behind it. Both co-founders bring specific, hard-to-hire expertise that Amazon couldn't easily develop internally.
Rob Cochran — Co-founder and CEO
Former Head of Product at CTRL-labs, the neural interface startup acquired by Meta in 2019 for up to $1 billion. Before CTRL-labs: Goldman Sachs and Harvard Business School. Cochran brings the product and commercial instincts to translate research capability into a product people will actually buy and use.
Josh Merel — Co-founder and CTO
Research Scientist and Manager at Google DeepMind, where he focused on AI models for robot locomotion. His publications have accumulated over 10,000 citations. After DeepMind: CTRL-labs and Meta. Merel's work directly contributed to the locomotion control research that underpins modern humanoid walking — he's now inside Amazon.
The 50-person Fauna team as a whole represents a concentration of robotics AI talent from some of the most competitive research environments in the industry. Amazon's robotics division — which has historically focused on purpose-built warehouse robots like Proteus, Sparrow, and Cardinal — gains a team with direct consumer and general-purpose humanoid experience.
The Personal Robotics Group, the Amazon unit that Fauna is joining, was formed to pursue longer-horizon consumer robotics research. That group produced Astro, Amazon's home robot prototype, which has been in limited release since 2021. Sprout represents a more capable platform to build on than anything Amazon had internally.
What Amazon's Humanoid Push Means for the Industry
Amazon's two robotics acquisitions in March 2026 — Rivr for last-mile delivery and Fauna for consumer humanoids — mark a significant shift in the competitive structure of the humanoid robotics market. Until now, the dominant players have been pure-play startups: Figure AI, Agility, Boston Dynamics, and 1X Technologies. Big Tech companies have mostly participated through investment (Microsoft in Figure, Google in various startups) rather than direct product development.
Amazon changes that calculus. The company has four specific advantages that pure-play startups don't:
• Distribution at scale: Amazon serves 200 million Prime members. A consumer robot that clears regulatory approval could reach consumers through a channel no startup can replicate.
• Home data: Amazon has decades of in-home product data from Echo, Ring, and Astro deployments. That data is directly relevant for training a home robot's spatial reasoning and human interaction models.
• Logistics testing environment: Amazon's fulfillment network is the world's largest real-world robotics testbed. Sprout variants could be tested in actual warehouse conditions at a scale no startup can afford.
• Capital depth: Amazon's 2025 operating income exceeded $68 billion. The company can sustain a consumer robotics program through years of development at a cost that would exhaust any VC-backed competitor.
Key Insight: The Acqui-hire Calculus
Fauna raised only $30 million before being acquired. For Amazon, the deal cost a fraction of a single year's robotics R&D budget while acquiring a shipping product, a validated hardware design, and a research team with direct ties to Google DeepMind's locomotion work. Acqui-hires of this type — small, technically elite teams with working prototypes — are likely to become Amazon's preferred entry into robotics sub-categories.
The competitive implications extend to existing humanoid startups. Amazon's entry raises the probability that other Big Tech companies — Google (which already has DeepMind robotics), Apple (which has been quietly hiring robotics engineers), and Meta (which has a robotics research division and prior relationship with Cochran via CTRL-labs) — will move from research to product more aggressively. A market that was mostly startup-vs-startup in 2025 now has a Fortune 1 company building toward the same consumer use cases.
For Fauna's early customers — Disney, Boston Dynamics, NYU — the acquisition creates an immediate question about continued access to the Sprout platform and SDK. Amazon has not indicated whether Sprout will remain available as a developer product post-acquisition or be folded into a proprietary Amazon robotics effort.
$68B+
Amazon 2025 Operating Income
200M+
Amazon Prime Members
2
Robotics Acquisitions in March 2026
The 12-Month Outlook
Amazon has not announced a product roadmap for the Fauna team. The most likely near-term outcomes are: (1) the team continues developing Sprout as a developer and research platform under Amazon's robotics umbrella, with possible hardware revisions; (2) Sprout technology is folded into a next-generation Astro product targeting the consumer market; or (3) the platform is redirected toward Amazon's fulfillment operations, where a smaller, lighter humanoid might handle tasks too delicate for Proteus or Sparrow.
Key milestones to track through Q4 2026 include whether Amazon makes any formal announcement about Sprout's future availability, whether Fauna's founders take public-facing roles in Amazon's robotics organization, and whether Amazon files any robotics-related patents that build on Fauna's locomotion work. On the broader market front, watch for Google, Meta, or Apple to respond with acquisition or product announcements of their own — the consumer humanoid race just got its first Big Tech entrant.
Frequently Asked Questions
How much did Amazon pay for Fauna Robotics?
Amazon did not disclose the acquisition price. Fauna had raised approximately $30 million from Kleiner Perkins before the deal. Given the team size of 50 people and the 59-day gap between product launch and acquisition close, analysts estimate the deal was in the $100–$300 million range, though no public figure has been confirmed.
What is Sprout and what can it do?
Sprout is a 3-foot-6-inch, 50-pound bipedal humanoid robot designed for human-shared environments. It can walk, dance, kneel, crawl, sit, and wave, with 29 degrees of freedom and rubberized grippers for picking up objects. It runs on NVIDIA Jetson AGX Orin hardware and includes a ZED 2i stereo camera, four time-of-flight sensors, and a four-microphone array. The Creator Edition, priced at $50,000, ships as a developer platform with a full SDK — it's a research and application development tool, not a consumer product.
Is Sprout still available to purchase after the Amazon acquisition?
Amazon has not clarified whether Sprout will remain available to external developers and researchers. Existing early customers — Disney, Boston Dynamics, UC San Diego, NYU — were already shipped units before the deal closed. New orders are in an uncertain state. This is a common pattern in acqui-hires: the acquiring company typically pauses external product availability while integrating the team.
Why did Amazon acquire Rivr and Fauna Robotics in the same month?
The two acquisitions target different points in Amazon's business. Rivr (acquired March 19) makes stair-climbing delivery robots that address the "last 100 yards" problem — getting packages from a delivery van to a doorstep, including up stairs. Fauna addresses the home and consumer interface layer. Together, they suggest Amazon is pursuing a full-stack robotics strategy: autonomous delivery to the door, then autonomous assistance inside the home. Both acquisitions were completed within a week of each other, indicating they were likely in parallel negotiation for some time.
How does Amazon's robotics acquisition strategy compare to its competitors?
Amazon has historically built warehouse robotics through acquisition (Kiva Systems in 2012 for $775 million) and internal development (Proteus, Sparrow, Cardinal). Google has pursued a research-partnerships model through DeepMind, working with Boston Dynamics, Agility, Agile Robots, and Apptronik rather than acquiring them outright. Tesla is entirely vertically integrated, building Optimus internally. With the Fauna acquisition, Amazon is now the only major tech company to have acquired a shipping consumer humanoid product, putting it ahead of Google and Apple on the product timeline — if Amazon chooses to commercialize what it just bought.
Amazon's Consumer Robotics Bet
The Fauna Robotics acquisition is a compact deal with large implications. Amazon paid an undisclosed sum for a 50-person team, a shipping developer product, and — perhaps most importantly — a co-founder who built locomotion AI at Google DeepMind. The deal closes in the same month Amazon acquired a last-mile delivery robotics startup, the same month Apptronik closed a $935 million Series A, and the same month Figure 03 walked into the White House. The industry is past the point of debate about whether humanoid robots are coming — the question is who will be selling them to consumers at scale by 2028.
Amazon has a credible path to that outcome. Its distribution network, home device footprint, and capital reserves give it structural advantages that no pure-play startup can match. Whether it moves on that advantage depends on what the Fauna team builds next — and whether Amazon chooses to commercialize a consumer humanoid or keep Sprout inside the walls of its fulfillment network.
The Bottom Line: With two robotics acquisitions in five days, Amazon has shifted from robotics investor to robotics builder — and the consumer humanoid market's competitive structure will never look the same.
The next 12 months will reveal whether Amazon treats Fauna as a research asset or the foundation of its first consumer humanoid product line. Given the company's track record of scaling what it acquires, betting on the latter is not unreasonable.