Brainchip Holdings Ltd (BRN) Acting CEO, Peter van der Made

Brainchip Holdings Ltd (BRN) co-founder and acting CEO, Peter van der Made discusses themes driving Artificial Intelligence development and Edge Computing, how the Group’s Akida chips differ from other products and the importance of key relationships such as the one with NASA, in addition to the global search currently underway for a new CEO. 

1 Jun 2021

 

Transcript

[MUSIC PLAYING] TOM PIOTROWSKI: Thanks for joining us for the Executive Series today. I'm speaking with the Interim CEO of BrainChip, who is Peter van der Made. Peter, great to talk to you.

PETER VAN DER MADE: Thank you.

TOM PIOTROWSKI: Now, people who aren't familiar with BrainChip, how would you explain precisely what it is that BrainChip is focused on?

PETER VAN DER MADE: BrainChip is doing something unique in that when we set out in 2004, we decided that we needed to do something different from what everybody else was doing because neural networks were not working very well in 2004. So we set out right from the beginning to develop something that would advance artificial intelligence and do better than the technology. And because we started in 2004, and we got our first patent in 2008, we developed a technology that is very different from what other people are using in this field.

So when deep learning came in its own around 2011, we started looking at what deep learning actually did and implemented those technologies into our then existing technology. So now we have a chip that is very versatile. It can work both with spiking neural networks, which we were doing in 2008, as well as with deep learning networks that are very common these days.

TOM PIOTROWSKI: So do you use deep learning and machine learning interchangeably?

PETER VAN DER MADE: Interchangeable, our technology can do both. It can do machine learning in real time. So you'll show it a little elephant, and it'll recognise the elephant from there onwards. You show it a little tiger, it will recognise tigers in the wild. It can do that with one-shot learning, which is really unique in this field. All the other systems need weeks or months of training.

The thing is that we can also do, like older classification, in a purely spiking neural network. So the chip is very versatile. It can be configured in Python within TensorFlow. So it's a programming language that everybody knows and TensorFlows with all that data scientists are using.

TOM PIOTROWSKI: So you made an important point a moment ago, and that is your patent. Patents aren't necessarily what they used to be, historically. How important is patent protection, and does it have the value that it has done in years gone by?

PETER VAN DER MADE: Patents are very much validate a company. Without a patent, you have nothing. Anybody can use it. So patents is the first thing we do. And whenever we develop something new, like if you look at this new chip that we developed, the AKIDA 1000, we have two patents that are covering that technology and all the aspects of that chip. So our first patent is for spiking neural network and the technology that's involved in that. 2008, we were the first to do a digital sparking neural network in hardware.

The following patents that have been written are other implementations or other advancements on that technology that we developed earlier on. And at the moment, we have 14 patents that are in process or have been granted.

TOM PIOTROWSKI: So in the context of intellectual property theft in this era, given where we are with technologies moving ahead so quickly, is that an issue from your vantage point?

PETER VAN DER MADE: We are very careful with our trade secrets. Besides patents, we also have trade secrets. So we don't put everything in our patent description.

TOM PIOTROWSKI: OK.

PETER VAN DER MADE: So we're guarding those secrets very closely. We make sure that when we are delivering IP, for instance, things are encrypted so that people cannot steal our IDs. And, of course, it's very difficult to copy a chip.

TOM PIOTROWSKI: Some of the terms that get applied to the areas that you're operating in are things like edge technology, which we don't hear as much about at the moment, but the Internet of Things is probably a more common expression. These are the areas in which you're operating. That they're expected to grow quite quickly over the course of coming years, aren't they?

PETER VAN DER MADE: Yeah, it's expected by 2035 that there will be 40 billion units connected to the internet. Of course, if all those units are uploading images and sounds and everything else to the cloud to get processed in the cloud and then downloaded again, that's going to be a huge bandwidth problem. With Akida, that problem is solved because we do all the processing on the device itself, and that's what we mean by the edge.

TOM PIOTROWSKI: Yep.

PETER VAN DER MADE: The edge is completely independent from the cloud. So anything that is picked up by a sensor, like whether it's a smell or like odours, images, sounds, vibration, tactile information, all that information is processed right on the Akida chip.

And if you want to send something to a computer, all you need to upload to your computer is the information saying that, hey, there's somebody in the room, or I'm detecting this smell which is basically metadata. Very compact, no bandwidth issues. And with the low power consumption of Akida, you can power those processes off a small battery.

TOM PIOTROWSKI: That's a good point you make about bandwidth. How important is 5G in terms of being a catalyst for the growth in demand for something like your chips?

PETER VAN DER MADE: Well, 5G is not very important to us because we don't need to connect to the internet. We can process everything on the chip itself. So 5G is wonderful technology but not terribly important to us.

If you look at the way the market is at the moment, where like Siri and Alexa, devices like that, they upload everything to the data centre, and then they have to get an answer back, 5G becomes important. But even 5G would not be sufficient in a self-driving car. Because if you're driving in the [INAUDIBLE], there is no internet. If you need to wait one second for the data centre to respond to your information that comes from your sensors on your car, you're 50 metres further handled in 10 kilometres an hour. A lot can happen in 50 metres.

TOM PIOTROWSKI: So Peter, what do you see as the biggest catalyst that will drive demand for the Akida chip, for example, over the course of the medium term?

PETER VAN DER MADE: The biggest catalyst is the big advantage that we have over existing technologies. Existing technologies are what they call deep learning accelerators, which are basically math chips. They are taking information from the CPU, they process the math very quickly, and then give the math back to the CPU. But your neural network and all your smarts are still running on the CPU.

While in the Akida chip, everything is running on the Akida chip. And all the CPU needs to do is send information to the chip and get the information back from the chip. So there is no overhead on the CPU which is a huge advantage.

So I think we will see Akida being used in things like refrigerators that are smart and can smell what sort of food in the fridge and whether it's off or not. Cars that look at the driver and see if the driver is paying attention. A lot of traffic accidents are actually due to drivers either falling asleep or not paying attention. So beneficially, AI is our focus. Things that are good for society.

TOM PIOTROWSKI: Now, what stands out to me is that your R&D spend over the course of the last 12 months was just over $5.5 million. Compared to some of the sums that must be getting invested in AI in terms of R&D globally, that's a very modest sum.

PETER VAN DER MADE: Yes.

TOM PIOTROWSKI: Is that a risk that you're not investing quickly enough in your technology?

PETER VAN DER MADE: We have some fantastic people. We have nine PhDs on staff. Everybody within BrainChip is putting in an extra effort. We are a very hard working group of people. We have accomplished with 45 people what IBM and Intel can not do with their huge organisations and their deep pockets.

I think we are very efficient in spending our money. And we are currently working on next generations. The AKIDA 1500 is already working in our lab in Perth. So I believe we are, well, at least three years ahead of the competition. And we have accomplished this with a team of 45 or 50 people, which is really amazing.

TOM PIOTROWSKI: Speaking of the value of people, you've had a bit of a transition in terms of leadership in recent times. How is the job search going as far as finding a new candidate to head up the organisation?

PETER VAN DER MADE: Yeah, we have a shortlist. We entered this process for about two or three months now. We are looking at candidates that are in large organisations. We need a CEO, for instance, who's capable of taking this organisation from where we are at the moment to the next level up.

So we are thinking AMD, Intel, Google, all the big organisations. This is where we are looking for our next CEO. We need to entice our next CEO away from these big organisations, somebody who may be second in charge who has been passed over for the job, somebody who's eager and hungry for success.

TOM PIOTROWSKI: Indeed, there are some pretty impressive organisations that you've mentioned there. If there was a timeline for when you are hoping to have someone in the role, what date would you be shooting for?

PETER VAN DER MADE: We're looking at a time of about six to nine months to get this person on board.

TOM PIOTROWSKI: So the other thing I wanted to ask you about is, obviously, when you're growing an organisation, the need for capital is quite intensive. Are you thinking that there might be a moment in a future where you'll need to raise some more capital?

PETER VAN DER MADE: At the moment, we have an instrument in place with the LDA. LDA is eager to provide more money to us, which is a good position to be in. We have sufficient money in the bank. So we have a runway of at least a year. We will see a revenue kicking in within the next year. So we're very optimistic in as far as our financial position is concerned.

TOM PIOTROWSKI: Relationships are important for any organisation when they're at the beginning of that growth trajectory. You've got some important ones, specifically with your chip maker in Taiwan. Chip shortages have been a big issue globally in recent times. How are you going in relation to your target to getting the chips out by the end of the year?

PETER VAN DER MADE: We have a really good relationship with Socionext. Socionext is the second biggest user of chips in the world. The relationship between Socionext and TSMC in Taiwan is also really excellent. So with these contacts that we have, we have been able to book production slots. And we don't expect that the global chips shortage will have any effect on BrainChip.

TOM PIOTROWSKI: In another headline that got everyone's attention was when you struck up a relationship with NASA.

PETER VAN DER MADE: Yes.

TOM PIOTROWSKI: That is quite the coup. Tell me about the benefits and how that relationship is going at the moment.

PETER VAN DER MADE: The relationship is really, really good. Of course, it's a constant conversation with NASA. They have one of our systems of the Akida engineering samples that we had earlier this year. They have been playing with it.

They are very excited about the technology, which is great that NASA is validating our technology. It's a great validation of what we have been doing. And the fact that they are excited about the technology is great.

TOM PIOTROWSKI: Peter, it's such an exciting space in which you're operating. And it's been a delight to talk to you today. I know our customers will be really enthusiastic about listening to your conversation. So thanks for making the time to speak to them today.

PETER VAN DER MADE: Thank you very much.

TOM PIOTROWSKI: And thanks for joining us for the Executive Series.


Commonwealth Securities Limited ABN 60 067 254 399 AFSL 238814 ("CommSec") is a wholly owned, but non-guaranteed, subsidiary of the Commonwealth Bank of Australia ABN 48 123 123 124 AFSL 234945 ("the Bank") and both entities are incorporated in Australia with limited liability.
This information is directed and available to and for the benefit of Australian residents only and is not a recommendation or forecast.

This information has been prepared without taking account of the objectives, needs, financial and taxation situation of any particular individual. For this reason, any individual should, before acting on the information on this site, consider the appropriateness of the information, having regards to their own objectives, needs, financial and taxation situation, and, if necessary, seek appropriate independent financial, foreign exchange and taxation advice. CommSec, and its related bodies corporate, do not accept any liability for any loss or damage arising out of the use of all or any part of this information. We believe that this information is correct as at the time of its compilation, but no warranty is made as to its accuracy, reliability or completeness. 

 

Disclaimer

This site is directed and available to and for the benefit of Australian residents only. © Commonwealth Securities Limited ABN 60 067 254 399 AFSL 238814 ("CommSec") is a wholly owned, but non guaranteed, subsidiary of the Commonwealth Bank of Australia ABN 48 123 123 124 AFSL 234945 and both entities are incorporated in Australia with limited liability.

By clicking on the "Download the CommSec App" buttons above, you will be directed to itunes.apple.com or play.google.com. These sites are not affiliated with CommSec and may offer a different Privacy Policy and level of security.

Top