Overview
Anki made robotics familiar through racing games and expressive characters. Founded in 2010 by three Carnegie Mellon graduates, it shipped Drive, OVERDRIVE, Cozmo, and Vector. Its challenge grew with its product range: each device needed manufacturing, retail distribution, software, and support. Founder Boris Sofman described forecasting mistakes before the company closed. His account supports a diagnosis of cash and operating risk, although public evidence does not isolate the final balance-sheet shortfall. [1][2]
The original company closed in 2019. Digital Dream Labs bought its assets that December, and the current ANKI operator continues product support. This history joins a failed company with products that retained value after an asset sale. [3][4]
Image 1 / 2
Founding Story
Boris Sofman, Mark Palatucci, and Hanns Tappeiner met when they began graduate study at Carnegie Mellon in 2005. Sofman and Palatucci earned doctorates; Tappeiner earned a robotics master's degree. They founded Anki in 2010. Their shared research background supplied the team and the technical starting point. [1]
The founders chose entertainment to introduce robotics to households. Sofman told Carnegie Mellon, “an entertainment product was a fantastic way to familiarize people with the products”. Drive appeared at Apple's Worldwide Developers Conference keynote on June 10, 2013. It gave the company a prominent demonstration of autonomous physical play. [1]
Selling beyond early enthusiasts required different skills. In January 2018, Sofman told SVB's Matt Trotter, “We’ve massively underestimated demand, and we’ve also over forecasted”. He described the distinction between retailer orders and purchases by households. Those measures can diverge: shipping stock into stores creates revenue opportunities while leaving unsold inventory in the channel. [2]
Anki's history therefore requires two questions. Did people want particular products? Could their purchases support the development and service obligations of the whole company? The reviewed evidence supports the first more clearly than the second.
Timeline
- 2010: The three Carnegie Mellon alumni founded Anki. [1]
- June 2013: Drive appeared at Apple's developer keynote. [1]
- October 2016: Cozmo launched in the United States at $179.99. [5]
- April 2017: Universal and Anki announced a Fast & Furious OVERDRIVE edition. [6]
- 2019: The original company closed. Its asset sale to Digital Dream Labs closed in December. [3][4]
- 2026: Current ANKI support documents Vector memberships and a successor OVERDRIVE app. [7][8]
What They Built
Drive and OVERDRIVE put software-controlled opponents into physical racing cars. Players built a track and used a mobile device to race. Anki combined a familiar game with autonomous movement that people could watch on a table. [1][6]
Cozmo made personality the main attraction. The small robot remembered faces, played games with interactive cubes, and received new activities through its companion app. Its $179.99 launch package included a charger and three cubes. Anki also introduced a free software development kit for programmers, educators, and researchers. Buyers could treat the same device as entertainment or a programming tool. [5]
Personality needed explanation before purchase. Vehicle, Anki's retail design partner, adapted existing Cozmoments videos and added captions around features identified through consumer research. It also adapted messaging for Amazon. The display case study documents a deliberate sales effort across several countries; it provides no measured sales lift. [9]
Universal's Fast & Furious partnership offered another route to recognition. Its April 2017 announcement described two film-based cars, characters, a Power Zone track piece, and compatibility with existing OVERDRIVE equipment. The planned September launch price was $169.99. Familiar entertainment supplied a reason to try the system, while adding licensed content to its continuing software obligations. [6]
Vector moved toward a home companion. Hilco's asset-sale account describes a robot that recognized people and objects, avoided obstacles, took pictures, and expressed the weather. This broadened the range beyond racing and Cozmo's structured games. [4]
Market Position
Target Customers
Anki sold premium household entertainment through retail. The programming tools also addressed educators and hobbyists. These audiences shared hardware but needed different proof of value: a gift buyer could respond to personality, while an instructor needed predictable setup and repeatable activities.
Market Size
No reliable addressable-market calculation was established. Anki's reported holiday rankings and usage offer narrower evidence. Applying the value of the entire robotics industry to these products would overstate their reachable market.
Competition
Anki competed for purchases and time among toys, games, and consumer electronics. Physical character distinguished its products from screen-based play but added manufacturing and compatibility costs. The founder's account describes sustained awareness spending after the initial launch, limiting what a prominent keynote could accomplish alone. [2]
Educational robotics remains an occupied market. Sphero currently lists its Mini Activity Kit at $99, including a robot, construction pieces, and activity cards. Its Edu app offers drawing, blocks, and JavaScript. Current Cozmo support also describes Code Lab and Education Mode. A new lesson product would compete with maintained manufacturer tools. [10][11]
Business Model
Anki sold hardware with continuing software value. Cozmo's launch promised free app content and software updates. Future service costs therefore followed the original sale. That promise could improve the purchase while creating work that later hardware sales or other funding had to cover. [5]
Anki said a late-stage investor deal failed. Roughly $200 million had been raised. [12] Publicly reviewed material does not establish audited product margins, inventory losses, return rates, or the funding shortfall. The operating interpretation should remain narrower than an accounting conclusion.
Traction
In January 2017, Anki said Cozmo and OVERDRIVE ranked among the four best-selling premium toys above $75 during the holiday period. The release attributes the ranking to NPD's retail tracking for October 2 through December 31, 2016. It also reported four million hours of OVERDRIVE play and 130,000 track configurations. These are company-reported ranking and usage measures, rather than profit or recurring retention. They nevertheless weigh against a story of universal consumer rejection. [13]
Outcome and evolution
Sofman's earlier account explains a plausible operating mechanism before closure: uncertain demand required inventory decisions before household purchases were known. Launch publicity also had to be followed by repeated sales work. In 2018, he acknowledged, “we’ve also over forecasted”. Expanding products and countries could spread demand risk, but also increased the range of equipment and software to maintain. The evidence supports these pressures; it does not establish which expense determined the investor's decision. [2]
Hilco marketed patents, trademarks, and anki.com for an assignee acting for creditors. The sale to Digital Dream Labs closed in December 2019, with terms undisclosed. Digital Dream Labs announced the asset purchase in January 2020 without assuming liabilities. Its SEC-filed history also dates the acquisition to December. Product continuation thus followed an asset sale rather than continuation of the original operating company. [4][14][15]
As of October 23, 2026, the current operator lists Vector memberships at $11.99 monthly or $99.99 annually with ChatGPT. Its Claude option costs $14.99 monthly or $139.99 annually. Each membership covers one robot. Autonomous behavior and Alexa commands remain available without membership; the operator reserves Vector voice commands and model connectivity for paid plans. These are successor-service prices and claims. [7]
The current OVERDRIVE FAQ documents both continuity and limits. Original apps reached end of life, while the paid 2.6 app supports older cars. Two Fast & Furious cars now have different names following license expiry. The Power Zone track piece remains unsupported and can crash the app. Those details show why maintaining a connected toy includes software, operating-system compatibility, and licensing work long after purchase. [8]
Key Lessons
- Anki made difficult robotics understandable through racing and character. Its documented holiday rankings support that product judgment.
- Retail orders and household purchases carry different cash risks. Sofman's forecasting account explains why headline sales alone could mislead an operator.
- Cozmoments and the Universal partnership supplied specific ways to explain products. Initial launch attention still required continuing retail work.
- Free software updates attached future work to hardware revenue. The successor's subscriptions and compatibility limits make those obligations visible.
- Anki's assets retained value after the original company closed. A history should distinguish the company, its products, and their later operators.
Sources
- Carnegie Mellon founding and launch account
- Boris Sofman's SVB interview
- Current operator's company history
- Hilco asset-sale record
- Cozmo launch release
- Universal partnership announcement
- Current Vector memberships
- Current OVERDRIVE compatibility FAQ
- Vehicle retail-design case study
- Current Sphero Mini Activity Kit
- Current Cozmo programming and Education Mode
- Axios closure reporting
- Anki holiday-sales release
- Digital Dream Labs asset announcement
- Digital Dream Labs SEC-filed history