The company famous for robots that dance, backflip and open doors just handed the keys to a man who built a talking speaker. On October 7, 2026, Rohit Prasad, the former Amazon head scientist behind Alexa and the Nova AI models, became the new Boston Dynamics CEO. It’s the biggest bet yet that the future of robots will be decided by their brains, not just their bodies.
Prasad replaces interim chief Amanda McMaster, who had run the company since longtime CEO Robert Playter stepped down in February. And he arrives at a tricky moment: the electric Atlas humanoid is finally in production, but output is tiny, losses are big, and Hyundai’s targets are huge. Here’s who he is, why he got the job, and what it could mean for you.
| Key fact | Details |
|---|---|
| New CEO | Rohit Prasad |
| Announced | October 6, 2026 |
| Effective | October 7, 2026 |
| Previous role | SVP and Head Scientist, Alexa and AGI, Amazon |
| Replaces | Interim CEO Amanda McMaster (also CFO) |
| Last permanent CEO | Robert Playter, left February 27, 2026 |
| Owner | Hyundai Motor Group (controlling stake since 2021) |
| Flagship project | Electric Atlas humanoid, unveiled at CES in January 2026 |
| Board seat | Expected, subject to approval |
Who is the new Boston Dynamics CEO?
Rohit Prasad is an AI scientist first and an executive second. According to Boston Dynamics’ official announcement, he spent 12 years at Amazon, where he helped turn Alexa into a consumer AI service used in hundreds of millions of households. Before that, he spent nearly 14 years at Raytheon BBN Technologies, leading machine-learning research for government and commercial customers.
His last big job at Amazon was even more ambitious. After ChatGPT launched in late 2022, Amazon put him in charge of its artificial general intelligence team, and he oversaw the Nova family of foundation models that businesses now use. Fortune reported in December 2025 that Andy Jassy announced Prasad would leave Amazon at the end of that year, as the company reorganized its AI work under AWS veteran Peter DeSantis.
There’s also a local angle. Prasad has lived in the Boston area for more than 25 years, so he isn’t parachuting in from Seattle. He knows the region’s universities, startups and talent pool, which matters for a company headquartered in Waltham, Massachusetts.
“Boston Dynamics is uniquely positioned to advance Physical AI through its world-class robotics expertise.” – Rohit Prasad
The key takeaway here is simple: Boston Dynamics didn’t hire a robotics engineer. It hired someone who knows how to ship AI to millions of people.
What happened to Robert Playter?
Let’s back up a little. Robert Playter joined Boston Dynamics in 1994, just a couple of years after Marc Raibert spun it out of MIT, and became CEO in 2019 after serving as chief operating officer. He steered the company through its sale to Hyundai and the switch from the old hydraulic Atlas to the all-electric version.
On February 10, 2026, The Robot Report reported that Playter was stepping down, with his last day set for February 27. The company said he planned to retire. In a staff email he called it “the incredibly difficult decision to step down as CEO of Boston Dynamics” after more than 30 years there. CFO Amanda McMaster took over as interim CEO while the board searched for a permanent replacement.
Timing raised eyebrows. Playter left barely a month after he stood on stage at CES calling the new Atlas “the best robot we have ever built.” Gizmodo notes that Semafor described the period as a C-suite exodus. So the search took roughly eight months, and the board landed on someone from outside the robotics world.
Boston Dynamics CEO history at a glance
| Leader | Role | Period |
|---|---|---|
| Marc Raibert | Founder, led the company | 1992 to 2019 |
| Robert Playter | CEO (previously COO) | 2019 to February 2026 |
| Amanda McMaster | Interim CEO (CFO) | February to October 2026 |
| Rohit Prasad | CEO | From October 7, 2026 |
Who owns Boston Dynamics today?
Boston Dynamics has changed hands more times than most startups raise funding rounds. Google bought it in 2013, then sold it to SoftBank in 2017. Hyundai Motor Group agreed to buy a controlling stake in December 2020, a deal valued at roughly $880 million, and took control in June 2021.
Fast forward to 2026, and Hyundai wants the whole thing. A Reuters report, republished by The Star, says Hyundai announced plans in July to buy out SoftBank’s remaining stake of about 10%, with media putting the price around 500 billion won (about $371.5 million). Hyundai Glovis holds about 11%, alongside Hyundai Motor, Kia, Hyundai Mobis and Executive Chair Euisun Chung.
| Year | Owner or event | Price |
|---|---|---|
| 1992 | Spun out of MIT by Marc Raibert | N/A |
| 2013 | Acquired by Google | Undisclosed |
| 2017 | Sold to SoftBank Group | Undisclosed |
| 2020-2021 | Hyundai buys controlling stake (about 80%) | About $880 million |
| 2026 | Hyundai plans to buy SoftBank’s remaining stake (about 10%) | Reportedly about 500 billion won |
That ownership matters for you because it explains the strategy. Hyundai isn’t a venture fund waiting for an exit. It’s a carmaker that wants robots on its own factory floors.
Is Boston Dynamics actually profitable?
Not even close, at least not yet. According to the same Reuters report, a Hyundai Glovis filing shows Boston Dynamics lost 528.4 billion won in 2025 alone, and nearly 1.7 trillion won between 2021 and 2025. That’s well over a billion dollars in losses under Hyundai’s ownership.
Here’s the catch. Investors still see huge potential. IBK Securities estimates a valuation of 141 trillion won by 2030, while a senior Hyundai executive told Reuters an IPO is unlikely in 2027, and one analyst thinks 2029 or 2030 is more realistic. In other words, the new Boston Dynamics CEO has to turn a famous research lab into a business that makes money.
The key takeaway here is simple: Prasad’s job isn’t to make robots cooler. It’s to make them pay.
Why an AI leader for the Atlas rollout?
Here’s the thing. Building a robot body is no longer the hardest part. Boston Dynamics has had world-class hardware for decades. What humanoids still struggle with is understanding messy, changing environments and learning new tasks quickly without months of custom programming.
That’s exactly the problem large AI models are starting to crack. It’s the idea behind physical AI, explained in why 2026 is the ChatGPT moment for robotics. Boston Dynamics is already leaning in: at CES it announced it would integrate Google DeepMind foundation models into Atlas, starting in the automotive sector.
Let’s break it down. Here’s what Prasad brings that a pure roboticist might not:
- Scale experience: he helped run an AI product used in hundreds of millions of homes, with all the reliability and safety headaches that come with it.
- Foundation model know-how: he led Amazon’s Nova models, so he understands the training, data and compute that robot brains will need.
- Product discipline: Alexa had to work for regular people, not just engineers in a lab.
- Partnership fluency: deals like the DeepMind tie-up need a leader who speaks the language of AI labs.
That said, there’s a real trade-off. Gizmodo points out that his background is mainly AI, not robotics, even if he has put AI software into physical products before. Alexa also struggled to keep pace after ChatGPT arrived, and a 2024 Wall Street Journal story cited by Gizmodo describes him racing to catch up with a team of thousands.
“The robot body is solved enough. The race now is for the robot brain.”
Where the electric Atlas stands right now
Atlas is the product that will define Prasad’s tenure. If you read our breakdown of the new four-fingered Atlas hand, you already know the hardware is getting more capable. The numbers below show both the ambition and the gap.
| Atlas detail | What we know |
|---|---|
| Unveiled | CES, January 5, 2026 |
| Degrees of freedom | 56 |
| Reach | Up to 7.5 ft (2.3 m) |
| Lift capacity | Up to 110 lbs (50 kg) |
| Operating temperature | -4°F to 104°F |
| Power | Swaps its own batteries |
| First customers | Hyundai’s RMAC and Google DeepMind |
| 2026 deployments | Already fully committed |
| Reported output (May 2026) | About four robots per month, per Semafor |
| Plan by 2028 | 30,000 humanoids, per Gizmodo |
| Estimated unit cost | $130,000 to $140,000 (expert estimates) |
See the problem? Going from roughly four robots a month to tens of thousands by 2028 is a massive manufacturing jump. Hyundai is building a factory with capacity for 30,000 robots a year as part of its $26 billion US investment, and Gizmodo reports that about 25,000 of the planned humanoids would work in Hyundai and Kia factories.
Competitors aren’t standing still either. Tesla is pushing hard with Optimus, and our look at whether a $20,000 humanoid can fold your laundry shows how aggressive the price talk has become. BMW and Xiaomi are also testing humanoids on real lines, as covered in our piece on humanoid interns.
What the new Boston Dynamics CEO means for you
So, will an AI chief put a humanoid in your living room sooner? Probably not soon. Every Atlas detail Boston Dynamics has shared points at factories and warehouses, not homes. Expert estimates reported by the Korea Herald, via Gizmodo, put the cost of one Atlas at $130,000 to $140,000, dropping to about $30,000 only after 50,000 units are made, which is well after 2028.
Here’s how the change could reach you, depending on who you are:
- Factory and warehouse workers: you’re the first in line. Hyundai plans to start deploying humanoids at its Georgia plant in 2028, so robots doing heavy, repetitive lifting is the near-term story.
- Business buyers: extra Atlas customers are planned from early 2027. If you run operations, watch for software and AI updates, not just hardware specs.
- Consumers: a home humanoid from Boston Dynamics isn’t on any announced roadmap. The AI-first approach could make future products easier to teach, but that’s years away.
- Investors: there’s no listing yet, and Reuters reports a 2027 IPO is unlikely. Hyundai stock is the only indirect way in.
1. Watch the first Atlas fleet results
The robots heading to Hyundai’s RMAC and Google DeepMind are the real test. If they handle varied tasks reliably, the 2028 targets start to look believable.
2. Track how Atlas learns new jobs
Under Prasad, expect more talk about foundation models and faster task training. That’s the signal that the AI strategy is working.
3. Compare prices as production scales
Unit cost is the number that decides when humanoids leave the factory. If estimates fall sharply, the timeline for wider use moves up.
The road ahead for Boston Dynamics and its new CEO
Boston Dynamics has spent three decades proving robots can move like living things. Now it has a leader whose whole career is about making machines understand people. That’s a promising combination, even if the balance sheet and production numbers show how much work is left.
If you’re curious where this is heading, keep an eye on the first Atlas fleets in Georgia and at DeepMind over the coming months, and on any word about Atlas pricing or new customers in 2027. Want more context in the meantime? Start with our explainer on physical AI, then come back as Prasad’s first moves unfold. The next chapter of robotics is about to get a lot smarter.
Frequently asked questions
Rohit Prasad, the former Amazon head scientist for Alexa and AGI, became CEO on October 7, 2026. He took over from interim CEO Amanda McMaster.
Hyundai Motor Group has held a controlling stake since 2021, and in July 2026 it announced plans to buy SoftBank’s remaining stake of about 10%, according to Reuters.
No. A Hyundai Glovis filing cited by Reuters shows a 2025 loss of 528.4 billion won and nearly 1.7 trillion won in losses from 2021 through 2025.
Boston Dynamics said Playter planned to retire. He announced his departure on February 10, 2026, after more than 30 years at the company, and his last day was February 27.
Hyundai agreed to buy a controlling stake from SoftBank in December 2020 for roughly $880 million and took formal control in June 2021.
No. Atlas is aimed at industrial customers, with 2026 units already committed to Hyundai and Google DeepMind. No consumer version or price has been announced.
