When I first started using computers every day, I was working in the 50-story HQ of one of Australia’s big four banks. Hardware failed so often that each floor had a storeroom full of spare machines, and IT staff spent their days swapping out cereal-box-sized computers and flickering screens. Software and networks were prehistoric, and even desktop publishing required consultants charging A$240/hour (A$580 today).
Back then, the joke was that software was the stuff you couldn’t smash. I learned that first-hand in 1990 when turbulence over Sydney turned a $10,000 Toshiba laptop into a bag of bits. Hardware really was the hardest part of IT.
It still is, especially in edtech. Over 25+ years I’ve seen promising products come and go: the Apple Newton, CueCat, TiVo, Google Glass, OLPC laptops, PiTop, AR/haptics, and classroom clickers. Even genuinely great ideas, like Cubetto (an early programmable robot I advised and invested in), eventually hit the same wall.
The most recent example of the rock upon which edtech hardware gets wrecked is Promethean. Founded in 1996, it rode the Blair “education revolution” and billions in government spending to an £224m IPO on the London Stock Exchange in 2010. By 2012 it was posted losses of £166m losses, and in 2015 was sold for £97m to China’s NetDragon Websoft Inc. In 2023 NetDragon merged Promethean backing with Gravitas Education Holdings, in a complicated deal to create NYSE-listed Mynd.ai. Originally the shares traded at $285 but last week it was $0.53, giving it a micro cap of just $26.8m. Latest numbers: revenue down 39% to $89m, gross profit down 54%, margins slashed to 21.7%. In other words this was looking like an edtech hardware implosion.
Mynd’s strategy to get out of this predicament looks like a ‘rope a dope’ where a losing boxer’s strategy become tiring their opponent out by pretending to be trapped on the ropes and when the opponent’s exhausted,, execute a devastating counter offensive. This was (in my view at least) why last week Mynd last week made an NYSE announcement saying they had ‘as entered into a definitive agreement to acquire all of the software, patents, trademarks, know-how and other intellectual property related to an award-winning AI voice and remote-control technology’, so they can ‘incorporate AI into our Promethean ActivSuite’.
They didn’t say who, but I can: Merlyn Mind (although that’s just part of the story). I met some of the Merlyn Mind team at BETT earlier this year where were spinning their corporate narrative that ‘voice control of technology saves time, eliminates tech friction, and allows teachers to focus on what matters most: improving student outcomes’. It was a very solid pitch, but they faced three big hurdles: money, manufacturing and distribution. Having raised a $100m Series B in 2024 led by Learn Capital money looks like the smallest challenge, but that really depends on the burn rate and the markets shrinking appetite for edtech (especially hardware) even amongst specialist investors.
Manufacturing and distribution are very different and that’s where NetDragon’s China connections and existing international distribution network becomes very compelling. Except they aren’t getting the real value in the Merlyn Mind business, which is the agentic AI (autonomous systems that can make decisions and perform tasks with or without human intervention) business Emergence AI.
At first I wondered why Mynd and Merlyn Mind had been so cagey in their NYSE announcement? It turns out it’s really a sideshow fto their valuable non-edtech business Emergence AI. So we get some Merlyn Mind tech shoe-horned into Mynd.AI/Promethean, but what happens to the actual stuff schools use today? Right now no one is saying but I’d watch this space.
Back to how tough it is to have a successful hardware-based edtech business. Mynd’s move may buy them time, but I doubt it will save investors much pain in the long run. Hardware remains the rock on which most edtech dreams, and a hell of a lot of money have been wrecked.
Leave a Reply