I like to fossick through the market from time-to-time, digging through the Micro Caps to see what’s new. One company’s been on the radar for a while, but it’s always looked like it needs more time or capital and I’ve moved on. An analyst mate reminded me about this week though and maybe it warrants a closer look. It’s called Brainchip (BRN) and with a name like that it won’t surprise you to learn that it’s an Artificial Intelligence (AI) play. The only pure-play on the ASX no less!
There’s been a lot of hype around AI, so-much-so that it’s almost become another meaningless buzz word for management to throw around, showing how cutting edge they are. For major corporates, AI is more call centre than terminator though. The technology is most commonly deployed to automate otherwise mundane tasks, such as identifying abnormal purchases on a credit card. This is good news for profitability and efficiency and certainly something that will separate winners from losers across many different sectors. But it doesn’t feel like we need to start fighting back against the machines just yet.
Or does it? Having a read of Brainchip’s presentations kind of makes me think that we might. Now, I’m not going to pretend to be across the technical aspects of all this. But the company has patented technology that essentially replicates the architecture of a biological brain on a chip. It’s known as a Spiking Neural Network (SNN) – a network of neurons that trigger an electrical spike in response to an input. The chip uses synthetic neurons to recognise repeating patterns, which is the foundation of learning. This is different to the methods used by mainstream AI developments such as Deep Learning, which is a more traditional computing processes.
You’ll have to go elsewhere for the technical details, but the bottom line is that an SNN such as Brainchip’s is faster, less energy intensive and not reliant on massive computing power in the way that Deep Learning is. This means that it doesn’t need to be cloud connected, so there are less issues around lag (important for self-driving cars) and its light weight means it can be deployed in drones and/or in remote locations.
This being the case, it seems like it should be the standard for AI development moving forward. Maybe it would be if the technology had been developed by Google or Nvidia.
The issue for long-suffering shareholders has been that despite the apparent promise of the technology, there hasn’t been any traction on its commercial development. But the reason I’m more interested in the story again is that the company’s management ranks are now populated with commercial, sales types, rather than just the boffins that develop the product. The company also completed a $27.5m capital raising last year and is sufficiently cashed up to get well into next year, by which time the sales guys should have put some runs on the board.
So, a decent product, sales guys getting the message out there and a pipeline of potential customers to work through. If this product does what it says on the tin, then BRN may finally start building some revenue over the next year. If it doesn’t, then it’s really just a nice idea….A chip without a brain. Wait, that’d be a no brainer! Maybe it’s not that, but it’s worth a small allocation of your speculative exposure.
Call us and see if you can tell whether you’re talking to a robot or not!