The Sad Story of Softbank’s Aldebaran Robotics and its Emotionally Intelligent Robot

Dec 15, 2014
Vote on Hacker News

atelier 02 - Crédit François Glevarec ©

The following article contains information obtained with the authorization by current and former employees of Aldebaran Robotics, who wished to remain anonymous due to the nature of their employment contract and its non-disparagement clause; however, all sources were vetted thoroughly). 

It’s been a good year to be working in robotics & artificial intelligence. While technology leaders like Elon Musk & Stephen Hawkings may not be so bullish on the pace at which we’re advancing in AI, the opportunity to bring our science fiction dreams to life – iRobot, Terminator, Bicentennial Man, need I say more? – seems to be inevitable, and close.

In France, no company has better embodied the nation’s desire & opportunity to become a market leader than Aldebaran Robotics. Acquired by Softbank Group in 2013 for $100 Million, the company’s line of programmable, voice-, speech- & facial- recognizing small-scale humanoid robots have dazzled in recent years. Videos of the company’s flagship model, Nao, dancing to Daft Punk or driving its own car propelled it to fame, and the announcement earlier this year that Softbank & Aldebaran would release a larger, closer-to-real-life model, Pepper, seemed to signal that the company was on its way to ushering in a revolution. Pepper is currently available in retail stores for Softbank, and Nestlé has announced Pepper will help sell its coffee capsules in Japan starting next year, as well as in Softbank’s recently acquired US Telecom Sprint’s US stores.

However, behind what has been hailed as one of the most provocative ventures in the world of robotics, continuous murmurs of worry and despair, coupled with rumors about the company’s ailing status have stacked up and threatened the future of Aldebaran and the robots it has spent the last 10 years building.

Internal Turnover

The first sign came from within, namely that, following the public announcement that Softbank had acquired Aldebaran (there were rumors of the deal in 2012, when the company acquired 75-80% of Aldebaran), turnover of early employees showed signs that things were changing.

Current and recently departed employees inside the company have told Rude Baguette that, since the acquisition, Softbank has appropriated all software development inside of Softbank, meaning that much of the software-side of the Aldebaran team has left, including Software Release Director Chris Kilner (with the company since 2007) & Director of Software Engineering Cedric Gestes (with the company since 2008).

In addition, Aldebaran has gone through 3 CTO’s this year alone. Alec Lafourcade-Jumembo left in February, after having been CTO since 2009, due to what unnamed employees close to the executive team called disappointment with Pepper’s development status given the empending launch date (June 2014). Aldebaran recruited Julien Simon from Criteo, who lasted just 6 months. The current CTO, Nino Sapina, was formerly Head of Studio.


Employees made it pretty clear that there was a disagreement between parent company Softbank and French subsidiary Aldebaran. Some point to Softbank’s culture differences (though the company manages 700+ subsidiaries), while many point the blame at Aldebaran’s executive team (namely, CEO & Founder Bruno Maisonnier). Respected in his field, Maisonnier has been criticized as having his heads in the clouds with respect to Human-Computer Interfaces (HCI), focusing more on the possibility of humans becoming friends with robots than on the practical applications of humanoid robots.

“I’ve never seen a company so big with such little organization.”

“I’ve never seen a company so big with such little organization,” one member of the Aldebaran executive team told us (they requested to remain anonymous, given the nature of their non-disparagement clause). Maisonnier has his hands in almost every part of the company, and split decisions can erase weeks of R&D from 15+ person teams – the kind of decisions that a small lab can afford to make, but that a commercial subsidiary can’t.

This cultural difference, sources tell us, grew to a point where, in 2014, the tone shifted in the relationship between Softbank & Aldebaran – email exchanges between the Softbank employees in charge of managing the Aldebaran joint-project & Aldebaran’s leads became heated – especially if you consider the typically polite nature of Japanese businessmen.

Up until the release of Pepper, Softbank still believed in Aldebaran’s ability to become “the global leader in robotics,” a source from Aldebaran’s executive team tells us; however, Aldebaran had been losing money on R&D since day one, with their flagship product Nao only available for schools & labs, and it was only a matter of time before projects hit the chopping block.

The first Pepper demo was done in 2012, when Softbank purchased Aldebaran – one source told us that Softbank’s acquisition of Aldebaran sparked Google to play catchup, acquiring the likes of Boston Dynamics in 2013. With the influx of acquisition money, Aldebaran hired too quickly, according to our sources, shooting up to 130 people, with Julien Simon – being the first real software CTO – arriving this year.

“If we were paranoid, we’d say that FoxConn stole the hardware, Softbank stole the software, and Aldebaran was sucked dry.” – Aldebaran anonymous executive


From inside Aldebaran, employees began whispering about conspiracies to drain Aldebaran of its resources, pointing to the June launch of Pepper in Japan, which featured, in addition to Softbank CEO Masayoshi Son  & Aldebaran CEO Bruno Maisonnier, Foxconn Technology Group chairman Terry Gou, which some felt was a hint that Softbank was leaking hardware development plans to Foxconn, in order to take its own subsidiary out of the development cycle (remember: the software had already been insourced to Softbank).

Next came the Budget Cuts

Beyond a hiring freeze (see this post from CTO Julien Simon in March, filled with new positions, and the company’s currently empty Job Board), the internal layoffs have been big: 25% of the team was cut in July.

The big cuts came down from Softbank with respect to, well, anything other than Pepper. Not many were irked when Aldebaran announced they’d discontinue development of their first ever robot, Rabbit; however, many will be upset to learn that Nao will soon be getting the same treatment, according to inside sources. Softbank never intended to keep Nao around – it was a necessary cost, during the development of Pepper; however, now that Pepper is out (well, sort of), it’s time for Nao to say goodbye.

Aldebaran still does great demo’s – in fact, its engineers say that the company is especially good at programming Pepper and other robots to do demo tasks; however, Pepper, today, still only performs a fraction of the functionalities it would need in order to become a standalone, commercially viable product. Likely, we’re told, Pepper will only ever be used within Softbank stores and a hand full of stores with which Softbank is friendly (like Nespresso in Japan).

It’s easy to look at this story as a case of company mismanagement, or an acquisition gone wrong, and write it off as such; however, it’s clearly more complex. On the one hand, you have a company developing cutting edge technology – the likes of which have not been seen anywhere else in the world – which has the potential to dethrone Google as the master of your home, and Apple as the next big App Store (imagine an app store where you could buy “Do the Dishes” or “Iron my Shirt”). On the other, you have a French company which spent money recklessly, was unable to convert R&D and academic ambitions into a commercially viable product, and a parent company which bought more than it bargained for, mostly within the executive team.

Softbank is writing off its losses – with investments in Alibaba and Yahoo! Japan, the company can well afford it – and France’s star robotics company is being sucked dry. We asked several employees if they though tAldebaran could make a turnaround. The answers ranged from “No” to “only with a major change in (Aldebaran) management.”