BrainChip Holdings (ASX:BRN) March 2018 Quarter Update Presentation

Company Presentations

 BrainChip Holdings Limited (ASX:BRN) CEO, Louis DiNardo, provides an update on the company.

Good afternoon everyone, this is Lou DiNardo. Hope everybody has good sound quality, got a little bit of feedback from last quarter's call or last month's call actually show some folks had a difficult time hearing. I hope it's improved, got a lot of good feedback from our friends at FNN and I'd like to thank Clive and his team in Sydney. It's about 4:00 o'clock in the afternoon here in California. West Coast US. I think it's about 9:00 a.m. Sydney, 7:00 a.m. Perth and about 1:00 a.m. in France though I do know that we have some participants from France that are likely dialling in. Saturday we're going to do the March 2018 quarter update, presenting this on May 1 in Australia. It's the day before that here in the U.S. and hopefully, this goes well. Next slide.

This is your standard disclaimer, you've seen it before and read it at your leisure. I'm sure many of you had seen this with many public companies so just move forward quickly. What we'll cover most of this is a little bit more colour a little bit more background on things that were presented in the quarterly update that we provided yesterday. So we'll cover a little bit more about the detail of the gaming par international development license and revenue sharing agreement. We've talked about that I think a couple of times now but certainly one of the most important steps forward that we've had within the original equipment manufacturer. We'll talk a little bit about Quantum Corp and the interoperability of our brain chips studio with their next large-scale storage platform.

And I think in the spirit of full disclosure some may know I was a director on the board of Quantum for several years. I have good insight as to the penetration and success they've had with that story next platform. We'll talk about Veritone. Veritone is listed on the OTC and the integration of Brainchip Studios as a cognitive engine in their software as a service and on premise solution. And we'll talk about what that means for the company in more detail on another slide or two. And that continued penetration that we've had with other original equipment manufacturers in the automotive industry, in the storage industry in a variety of market segments as well as the addition of better sales in Southern California which is our first manufacturer's rep here in the United States which is really how we do a go to market strategy for end users rather than building a direct sales force, incurring those fixed expenses. The announcement of Luis Coello as director of sales in Europe has been a rock star his first few months with us.

A little bit more about Akida where we are with respect to the development, the kernel that's been delivered to a European automobile manufacturer. We'll talk a little bit about cash receipts for the March quarter of 2018 which for US dollars this coincidentally amounts to 500000 exactly and approximately AU$650 000. Moving forward I want to spend a few minutes if I could here just not to diverge from the details of all those specifics but to give folks a bit of a better understanding of really what our business development sales model is. We generate revenue or will generate revenue, actually at this point, I guess we do already for each of these three segments. But this is where our growth will come from our sales model is to engage with OEM's and that really is the best leverage that we can get. This is business to business relationships.

Now just use as an example, we sell a product to a company like quantum, quantum then incorporates that into their system and they sell it to their customers. That could be HP, could be Dell, EMC, could be a wide variety of OEM's, which frankly gives us a great deal of leverage. Quantum corporate mission as I just mentioned, I know very well, hundreds or thousands of customers through one OEM. Similarly you'd find the same business model with Gaming Partners International, system integrators or those that are on the ground particularly for us in the surveillance business and both systems integrators, they pick where they have preferred suppliers of choice for specific parts of what they're installing, it could be the camera, it could be the electrical contractor. In our case, it would be the video analytics solution for a security system, could be law enforcement, it could be public safety, it could be school safety. And then we, of course, have end users. So if you look at this little bit of a graphic OEM's, their development agreements, there are sales and marketing cooperation agreements. And eventually, you get a design win, which is quantum says we're going to use Gaming Partners International, says we're going to use you, could be HP, could be Dell, could be a variety of OEM's. That design win is really the critical part of our process.

The wonderful part about the OEM part of our business is that it's a low cost of sales. Once you've won that design win with a Quantum or a Gaming Partners International or a Veritone, they have hundreds or thousands of customers and frankly they have a very large sales force, you know, feet on the street. You have to go through an integration process, where we introduce Brainchip's skill for example that was an off the shelf end user solution in order to make it appropriate for OEM so we had to add the APIs or the application specific interfaces. So there's a little bit of incremental engineering expense once that's a sunk cost, you have a very low cost of sales.

In the OEM space, we can sell a branch of studio cell branch if accelerator to the enterprise and large scale deployment space. And then of course as we move into Akida, that is really where the big bucks come from at the edge. Or at the enterprise, Akida will be either a single IC that gets designed in that edge, could be a camera, could be a car, could be an Internet of Things device or at the Enterprise sitting on a PCIe card plugged into a server. System integrators - we win the design, their sales and marketing, cooperation agreements. There are supplier agreements. This is where we would announce contracts in the OEM space that's far less likely and we'll work off of forecasts and we'll work off of what we know about, what they tell us about their customers. System integrators are a low cost of sales. You know we won the deal in France that we've talked about a few times, Evelyn frankly never called in a municipality. Evelyn came into us through a system integrator that system integrator is the largest in France. You could look at a system integrated like in that case Infocom where you look at a system integrator like Teleste which is the largest in European system integrators.

It so happens that Luis Coello came to us through Teleste that's a really great connection. Again it's a low cost of sales, there's minimal engineering expense. In some cases it would be an off the shelf product in some cases they may use the API version and we'd have to put some engineering resources in place system integrated so we'll predominantly be Brainchip Studio and Brainchip accelerator. OEM will be a target for a Akida, but as it relates to system integrators there are going to be using an off the shelf software package potentially with an API development that they have to go through with us or the accelerator card for large scale integration end users. It's a whole other story. Win the design, you get purchase orders and then presumably sometimes you'll get supply contracts. But it is a high cost of sales if you think about calling out every police force in Europe. Every law enforcement agency in Europe, with the United States or Australia or throughout Asia it's a very high cost of sales to really attack the end user base. That's why our focus, while end users are great we get educated we get some wins and we generate nice revenue. The high cost of sales somewhat moderate engineering expense and predominantly focused on Brainchip studio. End users are not going to be developing their own PC boards with IC's like Akida either for edge or enterprise applications.

So I just wanted to get a full view of really how we think about our go to market strategy, OEM's, system integrators and end users. Would also like to touch on, if you look to the left a bit, the sales process is that I know this is different than many other industries whether it's the natural resources industry, the banking industry, in the technology industry we generate leads. This is what Bob Beatrice spends a whole lot of his time on with our partners, generate leads. And we'll talk about where the sales funnel is today. And I mean we've committed to do that on a quarterly basis and frankly we're doing that more often because calls seemed to be almost monthly now. But you generate leads, figure out which ones are opportunities, qualify by those opportunities either with a face to face meeting could be a phone call where you qualify those opportunities and you size them and you prioritize them. At that point you try to wrestle to ground a design win where customers said we're going to use your product and you know how successful our customers are really determines how you convert design wins into revenue.

Important part of the qualified opportunity and design wins is that we prioritize who those customers are in the vertical markets where we think we have the compelling solution and we prioritize our engagement with customers that are most likely to succeed. I just want to step back a second and talk about our business development model, and sales and business development model and make sure that we articulate as best as we can, how it is that we go to market. If we step back again and talk about what happened in the last quarter the March quarter that we're talking about today gaming products international I think has been a tremendous success for us in the automated cable system that they will deploy for currency security in Asia potentially in the U.S. but Asia and primarily Macau and Singapore really brings to the casino their ability to secure currency from theft from collusion from cheating of all sorts.

But most importantly stealing from counterfeit chips being used in casinos. It's a little bit less about the theft itself but when theft occurs most casinos have to swap out their entire chip set and you know a 50 or 60 or 70 thousand dollar event could cost them a million and a half. In the counter bid could be maybe as much as two three or five million dollars to swap out that chip set and this is where gaming partners are skewed, gaming prop bets has been a tremendous win for us. This is maybe the epitome of an OEM relationship. They have massive market share in casino currency and that ability for them to provide us surety around a currency is very attractive and compelling solution for casinos across the board. That development is based on intellectual property that is existing at Brainchip and new jointly developed intellectual property where we're tuning systems with developing algorithms and tuning algorithms so that we can recognize chips. We can give counts and we can correlate those counts, in many cases to what will be a RFID or radio frequency identification. What the casino believes is on the table, we'll be able to give a visual correlation when there's a difference and that correlation means there's counterfeit chips on the table. We're going to be showcased with GPI, they'll be showcased, we'll be in the background.

Again we're a supplier to them at the G2E conference in Macau during the week of May 15, 15 May 2018. So I will be at our annual general meeting in Sydney, the prior week and then head off to Macau to make sure that the installation goes well. I don't do the installation, of course we have engineering guys that do that, but to be there for oversight and make sure things go smoothly. The commercial deployment is expected later in 2018, and that utilizes our spiking neural network and provides virtually instant permanency security and again, that correlation with radio frequency ID. That instantaneous correlation is really what's important for anybody that's set at a blackjack table or baccarat table. Things move very quickly if IG make it one count we get a different count that means there's something wrong at the table. That currency security is really the first instance of what GPI is intending to provide, they have many, many other features and benefits that they intend to lay out over the course of late 2018, 2019 all of which will be part of.

The next slide - a little bit more about GPI, I think we've described this before but there's been a few questions. I believe you all know that I try to incorporate answers to the myriad of questions that we get online before this call. This is a development agreement, a license agreement and a commercial agreement. The commercial agreement kicks in as soon as the development is complete. It's a five year revenue sharing agreement, can't be terminated. Once we're into a commercial phase, can't be terminated. It's renewable based on mutual consent of both parties and it provides us with 25 percent of all GPI revenue derived from the sale of the developed technology to Brainchip. We've looked at it, we've sized it and talked about this before. Somewhere, depending on which report you read, there's somewhere between 60,000 and 68,000 gaming cables in the world. Frankly, the majority of them are in Asia. US North America happens to be heavily weighted toward slot machines.

But when you get into Australia, Macau, Singapore or other parts of Asia it's very heavily weighted toward table games. Our ability to recognize chips, give counts, recognize cards that are played is really a compelling advantage for the GPI ATS system. Of course, this is an OEM relationship. GPI is responsible for product pricing. They're responsible for end user demand creation. In this case I think we could not have found a better partner, they have the dominant share of the currency on the table and the dominant share of the tables and what we call lay-outs, the felt on the tables. So the agreement is a very, very solid one. We've gone through Phase 1, we've done the proof of concept. We'll be in Macau on the 15th with a full blown prototype and then we expect commercial deployment sometime later in 2018. So I hope that gives folks that asked a bunch of questions about GPI and what it means to us a little bit more sense of what it means to us. If you size it, even at modest levels of penetration and heavily discounted pricing versus what I think or what I've heard may be the price, this is five to $15 million annually to a to us. Again because of the low cost of sales the sunk cost of the engineering. It is a very, very profitable business for us.

I want to move on to the next slide which talks about Quantum Corporation and Candace's and OEM engagement which I believe is the best business model for the company Quantum. I call them located in San Jose California. They're spread throughout the United States. They actually have some folks in Australia, they have folks in southern California, San Jose as well as up in Oregon. They are a leading supplier of large scale storage, that includes both discs, spinning media as well as tape backup. The company has been in business a very long time. About 400, maybe a little bit more, maybe 450 million US in sales. Again I know the company very well from prior experience.

Storage is fundamental to everything that was talked about, civil surveillance, public safety, commercial surveillance which include gaming, the internet of things called IoT, big data. And frankly have a very big business in media and entertainment, which is the storage of sporting events; could be NASCAR, the National Basketball Association, FIFA. You know so everywhere around the world, there are literally dozens if not hundreds of high definition cameras that are recording everything that goes on at a sporting event. If you think about civil surveillance, estimates are anywhere between three and 500 million cameras in the world, public safety, schools, banks and other places.

And public safety is a big concern. Commercial surveillance in the gaming industry, if you think about 24 hours a day, seven days a week, 365 days a year, we're talking about terabytes, gigabytes, not terabytes but petabytes of video that's been recorded that has to be at some point whether it's because of a crime, because they want to generate a commercial or do a replay at a sporting event or replay on something at a sporting event. They have to be able to quickly search an enormous amount of data and our Brainchips studio solution whether that is in the form of software running on their server, which may be the case which may be the case with Store Next on Quantum. Or it may be our accelerator card plug in so that they get more throughput, more density, the ability to quickly analyze and get to what it is that you're looking for. I think this is an important part of what we're learning about the storage space, potentially reduce the requirement for storage itself. So you can bring down the overall cost of those petabytes that have to be stored.

There are other leading suppliers in the storage market and frankly the big five, so HP, Dell, IBM, NetApp and Hitachi, it's in that order. NetApp might be bigger than IBM but somewhat in this order. And you know one of the wonderful things about the relationship with Quantum when we put that press release out that said we had demonstrated interoperability and they will be kind of co-marketing with us at this juncture, the Store Next platform that they provide with Brainchip Studio. Now within a day, we had inbound traffic from one of the majors in this space. No one's going to overlook the fact that a company like Quantum which has 400 or 150 million in revenue has found a compelling solution with our SSNM search capabilities. Generally, these numbers that are very, very modest penetration of a very, very large market, but 20 to 30 million dollars annually. Given what we've seen and what will be touched beginning late 18's with significant growth opportunities as a technology, as we expect is to go to widespread adoption.

I'm trying to size these things and give a little bit more insight to where we think the large revenue capability or the potential for the company, as we get into late 18 and 19. So it's just mentioned that these are based really on modest penetration of a very large market. Again these are OEM partners. So again I think it's maybe somewhat cultural, but we won't get contracts that say here's what we're going to buy. Here's how many things we need to remember, I think there's two things. Number one they're OEM partners. And number two this is software, having done hardware chips solutions which will be the case, prophetically will be the case when we get to Akida when you build a chip. It takes 16 to 20 weeks to build that device. So your OEM partners and of course, your end-user partners. But more importantly when you're building hardware your OEM partner gives you a forecast and they have to sign up for a supply agreement, because it is 20 weeks between the day they tell you they need something and the day you can actually deliver. When you're doing software, there's very little lead time. It's basically a software drop from an FTP site, so we get, we get a bit less visibility selling software that will change as a key to really comes to be the predominant solution that the company is marketing.

Going to move forward again, this is a little bit more about sales and sales development or business development. We've talked about Luis Coello joining us as director of a math that's basically Europe and the Middle East and parts of Asia. Been decades of wonderful experience in the surveillance market. He did come out of Teleste, which is the largest pan-European a system integrator, based on just a couple of months on the job, we've moved outside of what really was restricted to France under prior sales management and we're now with in law enforcement, homeland security, public safety in France, Germany, the UK, Sweden, Norway, Spain, and frankly a few others, which I just didn't bother to put on the slide.

Similarly here with Greg Ryan now having been in the saddle in Australia for a bit. We've moved beyond engagements, up in the Northern Territory in New South Wales, a little bit up in Queensland, some down in Victoria, but he's also now spreading his wings and we've got engagements in Singapore now. When you talk about the United States, Tom Stengel has been on board now for the better part of six months or so. I think also with the addition of Bagger, which was the last bullet on here were law enforcement groups throughout the United States, including New York City, New York state, Baltimore, many others, but I think the fault force multiplier effect of having the more traditional, at least North America manufacturer's rep who doesn't get paid unless they book an order, covers a wide geography. I think the announcement with bagger there went out a couple of weeks ago, we mentioned, 17 sales professionals, we don't want that fixed cost. When those 17 sales professionals start to generate revenue and they've already introduced us to companies like Exxon, which is one of the largest suppliers of body worn cameras by law enforcement. When those orders come in and we invoice them to a customer, they get paid, but having 17 people just in southern California compared to the fixed cost of us trying to cover those same customers. This is, this is a wonderful business model for us here, at least in North America.

Moving on, I want to give a little bit of sense of where we are geographically. I don't know that everyone has really a sense of just how widespread we are and what we're trying to cover and trying to accomplish. This is, I don't know how visible this is to you, but if you look to the left, you'll see the United States. You see we're here in San Francisco is primarily business development administration. It's myself, the CEO. Ryan is CFO. Bob is the VP of business, the senior VP of business development and marketing in southern California, and I tried to do a little colour coding here, southern California and red his research and development. Research is, Peter is located there. He has computational neuroscientists. He's looking for more. In Toulouse, France. If you jump over to Europe, the red, the red circle is too loose. That's again research as well as software design.

So southern California is focused on hardware design. That is the FP and implementation of Akida eventual hardware or a basic design of Akida, in France. It's really research as well as software design, algorithm development and all of the software layers that are necessary for a product to be easily used there by customers. Luis Coello is located in Barcelona, Spain and it's all sales and marketing and that gives him that wide swath to move from all the way from the UK out into Sweden and Norway dropped down into France and Italy, speaks to that most of those languages and he's been very successful in getting us inside. Really key customers very quickly in. If you go all the way to the right, and this is a little deceiving because that's not how I would fly from San Francisco to Australia and actually going around the backside, but in Sydney we, we have a Greg Ryan who we've talked about a few times, he deals with sales and marketing and then in Western Australia in Perth we still have finance which helps support Ryan and the rest of the finance team here in the US.

So I think that gives you a better sense of our global footprint. It's a lot of travelling, but we've got to 26 employees. I would say that in any given week, somewhere between at least five to 10 maybe on the road and maybe flying, maybe visiting customers, maybe going and dropping into Las Vegas for example, to move forward with the demonstrations and the completion of the GPI product heading off to Macau to help market that product. And certainly I spend as much of our team does the executive management in part of our board. We spent a lot of time going back and forth to Australia. So I think that gives you a good sense of just what our global footprint is. Moving onto a bit of the questions and I think they were very good questions and very appropriate where we are with the Akida development, the Akida device as we've described before, it's a complete neuromorphic, we're going to call it system on a chip and not a processor and that is again because we do pre-processing to take what our real world signals or they could be data patterns, zeros and ones coming from financial transactions or could be cyber security, could be data coming from a wide variety of IOT devices.

But our goal here is to provide a complete system to monitor chip, which is a conversion to spike the actual spiking neural network fabric as well as the bus interfaces, depending on the end market. In automotive, it will be a can bus and industrial it'd be I-squared-C bus and you've got a PCIE bus. If you're working in the enterprise and you're plugging into a server. I think there was a question about how Akida, will compare to something that Intel is promoting, which is called the low high and low high and ABR and presumably in the last week or so had some demonstration that you can never, you can never discount your competition and you certainly can never discount Intel, but frankly Intel doesn't have Peter van der Made, Intel it doesn't have the engineering team that has been working on this thing for many years. Low, high is somewhat different. I don't want to drag us down into the weeds on the technical nuance between how spikes are accumulated and process, but they are different, low, high and Akida are significantly different in different and that meant and they are significantly different. And with respect to low high is really a test chip. It really is a bit more like snap 64, which demonstrated that everything Peter had been working on, Peter and Neil and company had been working on, works.

But it's a test chip. It's about 100. And if my recollection is correct, it's about 130,000 neurons on a piece of silicon. It doesn't have the data to spike conversion implemented requires an external chip. Doesn't have bus interfaces implemented. Again, you can't discount that Intel, Intel could do these things, but that's a, that's a whole lot of engineering as a whole lot of work to get there. So my view, our view is that Akida, which is, moving from research into development, we expect to be on time for an FPGA implementation in the third quarter of this year and beyond the FPGA implementation, a development environment, you can build an FPGA and you can have it on a PCA card, but if you don't have the documentation and the tools and the development environment key customers will not be able to exercise the, the strength of the spiking neural network or all of the competencies or capabilities that the data could deliver. The last bullet here and then I have to jump back up to this European automobile thing. The last bullet is, is very important. The artificial intelligence or machine learning world has been developing convolutional neural networks for a very long time. And that is what is in use today and running on GPS from NVidia or even CPUs from Intel, depending on how much throughput speeds you need.

One of the opportunities that we have, one of the things Peter and research team and the development team is focused on is rather than exclusively looking to OEM customers saying, develop a new spiking neural network, it's lower power, it's lower latency and you can do things that you otherwise couldn't do with a CNN, which includes autonomous, supervised and autonomous unsupervised learning. If we can port CNNS to our SNN fabric, we open up an enormous opportunity to bring compelling advantages to customers without them have, without us having to do and it kind of evangelist excel for them to say, here's our benchmarks with our existing CNNS. We'd give them a translator to a spiking neural network. We show them what those benchmarks are. It's an easy path from one place to the other. I want to go back to the third bullet, which I missed. We've talked to a few times now. People all have their suspicion about which European or automobile manufacturer we're talking about. We did ship our first hardware accelerator card; I think it was a couple of quarters ago now.

Really the end game and the called it ADAS advanced driver assistance or the autonomous vehicle market really is key to know it's not using a PCIE card and in a car that's just never going to happen. You know what their intent was to get to understand the spiking network, that card and then into about where they go either point to us and it's important that the software research, it's been delayed to the European automobile manufacturer. They're playing with it and get up and giving us feed. Sending folks to help the kind of use cases who would like to employ a wonderful relationship opens up. So I couldn't say we couldn't please having this European auto manufacturer as a partner. We're doing similarly in the US and you know, certainly we can't discount the third party providers to the automobile industry. So you know, in Europe, while you've got any, just thinking about the top guys, think about VW, think about Mercedes, BMW in the US, you've got Ford, GM, Chrysler, all of them have suppliers, companies like continental, Delphi, Valeo, I think some of Valeo in Europe, they are, these are large, very large companies who provide a great deal of the technology and the it gets reduced to practice in the automobile industry.

So those engagements are going well also and moving forward just a bit on where our engineering resources are focused and I think this does relate to a few questions about spending and what's our burn rate, when do we get to profitability? Our engineering resources right now are focused on a couple of things, refinements of the pattern recognition technology for the GPI deployment. It's critical, it's near term revenue, it's, it's meaningful revenue and something that I think will really demonstrate the strength of this OEM approach to marketing our products and the sales effort. And also of course on the development of the Akida normal fixed system on a chip which is going very well. As I said, the colonel has already been handed off to one of our major partners in the automotive space and we've got other partners that we expect to hand it off too soon as well. We expect it to be a leader in the emerging neuromorphic market field applications, but these are just a few. I mean, I could make a list, an arm's length long robotics because of its vision capabilities as well as any repeating pattern for motion detection could be sensor inputs, temperature pressure flow, vibration, the gaming industry.

This will give us another step up with our partner in a TPI surveillance we've talked about many times. Cyber-security, financial technology, IOT is a potentially very large market for us. Again, IOIT is devices that are remotely sensing data and then sending that data back and there's you know, there's a couple of ways to cut this one. For us, it could be analyzing that data for patterns and trends where it could be recognizing the combination of IOT and cyber security where people are using IOT devices to hack into systems. It was a very large, what we call it, denial of service attack here in the US, I guess it was probably a year ago now, but it brought down some of the big names in the every, everybody from, the twitters of the world to other major players in the marketplace because IOT devices weren't effected and they launched, a million ships against a particular a service centre and brought down the system with a denial of service attack. Again, autonomous vehicles. We've talked about including driver assisted systems. What does it mean to us and generally accepted data at markets and markets of 2015 report and many others that you've read is that it's a $48,000,000,000 market in 2022. That's a 26 and a half, 26.3 % combined average growth rate between 2016 and 2022.

I think Akida, we think the Akida is really in the forefront when you're, when you're looking at the competitive landscape and you're, you're comparing yourself to Intel's low high, where you're comparing yourself to IBM's true north as company in its formative years. I think that really puts us well out in front of any of the other competition and none, none of which that we've seen is nearly as close to Silicon as we are. Again, what is Akida, to what is compelling advantages? It's really low latency and low power is how fast can I get my data out and what is my, what is the power to get that data out? And it is fractions of the power that's, more than fractions is 100 times less the power for a GPU for specific use cases and extremely low latency. Let's touch a little bit, and this is one of the last things that was in the, in the deck that we provide or the press release that we've provided.

I think we would all agree that there's a great deal of effort. There's a great deal of science that's speed read being reduced to practice, but we all know the truth of the matter is you cannot confuse effort with results and I think the company has made enormous strides in 2017 and we went from a situation where when I joined the company in late 2016, virtually 2017, we've had less than a million dollars in cash. We had just acquired Spike Net, which was a contract services organization. Customers were marquis, but it was custom software into France. We had no sales force, we had no infrastructure really. We didn't have a finance organization within the Human Resources Organization and again, we didn't have a sales or Marketing Organization in that 12 month period. And we're now sitting on an enough cash to carry us for, I don't know, somewhere between 15, 16 months maybe, maybe even longer. And I think we finished the quarter with $13.9 million in cash. We have extremely qualified sales organization. We've got a sales pipeline that I think is very robust and we've got a marketing position. We've got human resources in place with the finance organization, so when we talk about cash burn, I think for a company in its formative years with 26 employees that as accomplished as much as what has been accomplished in the last 12 months or so. I think we're actually doing quite well.

So we burned $1.8 million, in the March quarter of 2018 which was down from 1.9 million in the December quarter of 2017. The decrease in outflows, it was primarily due to a higher cash receipts and with 500,000 in cash inflows and you can do the math quite easily. We know, some of the, some of the increase of course was a cost was of course the launch or branch of studio and accelerator. If the people on the street you have to, you have advertising, you know, trade shows, all the things that go with launching a product. I do think it's important to recognize that and this was a question that I got as well, the $500,000 in cash inflows. There was a big question about what was that $500,000 from GPI because a quarter ago we mentioned in the press release since the $600,000 upfront cash before we go into the revenue sharing part and I will tell you that it was significantly less than half of the cash inflows this quarter and the cash inflows was really came from a number of customers in the US and France the next quarter. I hope we get a little bit more global and the other category, but the $500,000 come exclusively from GPI. In fact it was, I think it was about 40 per cent or so.

And again, we had $13.9 million in cash at the end of the March quarter. Next slide, just to give you a little bit of kind of the outlook. We do expect to see continued sales growth. OEM’s, system integrators and other partners and you could really maybe rephrase that as end users and companies that we've talked about, and I think we've covered another slide here just and if you just looked at the markets that we serve and these are discounted from the total available market and when you think about civil surveillance, it's much larger than 2.4 billion. The piece that we could serve as 2.4 billion and that is with existing products and potentially with the introduction of Akida. Commercial surveillance, which would include gaming as well as consumer behaviour another 500 million Internet of things could be, I mean it's an enormous market and well discounted for what we think we could do with respect to looking at the incoming data from pressure, temperature, all the things I've talked about. Two billion. You've got basically something over about $6 billion and served available market. That's not the total available market. It's already discounted for the things that we could serve.

And of course, over time we'll get a better understanding what our share of that market could be, but when you're servicing a market of $6 billion plus, you have a leading and compelling technology, we feel we're in pretty good shape for the, for the future. I want to move on to the next slide which talks a little bit more about what, again, don't confuse effort with results and the first part of results in a technology selling process is to generate customer activity. Now all our recent participation in the CTX anti-terrorism conference in Las Vegas alone generated 179 leads, good qualified leads. These are people that came to the conference and out of, I forget what it was, maybe 30 or 40 thousand people. We honed that down to 179 leads that we think are important. Our commercial sales pipeline is still very robust. Now you can take that 179 in leads and you edit to what we generated in the first couple of quarters since the introduction of brain shift studio and the accelerator, it's over 500 active leads, more than a 100 we've recognized as opportunities that is take out research, take out things that you don't think are kind of worth the whip or worth the effort and of those 100, there's well qualified 50 opportunities and you know this. This speaks to what our sales force has to do.

They have to qualify the opportunities that say, okay, it's a customer that is can be a priority for us. It's someone that we think if we win, then they have a higher chance of winning, then customer B, C, D, or E, and if you add up their estimated lifetime value at 50 opportunities, loan is US$300 million and US dollars potential. We've achieved 15 design ones that will accelerate over time as we come better and we have more people on the street, but those design wins along those 15 have an estimated lifetime value of about approximately is US$80,million. Now design wins don's always turn into revenue, your customer has to be successful and over time this sales pipeline or this funnel will have statistically significant data that said, from leads to opportunities. What's the percentage from opportunities to qualified, what's the percentage? And from design when to real revenue, what's the percentage, and then we can really have our arms around what the future of the company looks like. Right now we fill up the pipeline, what the best and priorities of those best where we're going to put our sales resources. So that's it, we always finished with this slide that says Q and A, but it's a little difficult on this type of environment. I tried to answer most of the questions.

We did talk about low high and what the advantages are and there's some very esoteric differences between grade encoding and sparse encoding and population encoding, but we're pretty confident that we have a far more integrated solution and a solution that will be able to maintain low power over a wide variety of use cases in demands. Think one of the things I didn't touch on was Baritone. Baritone marks, an interesting development for the company. We haven't been focused, whether it's with quantum or others that we've talked about on the enterprise side where you're plugging into a PCAE card and it's a dedicated system, whether it's for law enforcement because they can't go to the cloud for privacy concerns. Even gaining can't go to the cloud, but that's an on premise enterprise system. The software as a service systems, which Baritone is kind of pioneering degree. Viewing the video as a software as a service is a very nice environment for us to play. That could be our software running on their CPU. If there's enough demand, they'll be plugging in our hardware.

If it's in the cloud, that could be a relationship with AWS or the Amazon system but it's a business model. So our OEM business model, that could be someone like quantum where we were selling licenses, a OEM model with GPI is we're getting paid dollars per table per day, very nice annuity model. In the software as a service model it's really the pay as you go. So it will be, dollars or pennies depending on what the volume of video is that's processed, but it's dollars or pennies of video hours processed. So three different business models, but three models that allow us to leverage the strength of their sales organization, their account base. Their cost of sales and frankly the engineering costs or something done. We don't have the big sales force that's necessary to reach all of their customers. So I had missed. I miss touching on Baritone. There was a follow up question because I did mention last time that we had an inbound from a large cell phone manufacturer. I would just remind you that it's a cell phone manufacturer in China. They have their own semiconductor group, so yes, it's a very active dialogue and we're getting to know each other, but we're being very cautious as that relates to question on the next piece of paper, which is protecting our IP. Large cell phone manufacturer, China with their own semiconductor business.

We need to really be careful about what IP we exposed, but more globally we have taken very seriously protecting the IP of the company and both Peter and Neil as well as Bob and the rest of the team has spent a lot of time in the past quarter making sure that we understand what we call white space between patents and where we can put incremental rings around our IP to make sure it's well protected. It just hasn't been a few questions about system and I really don't want to avoid it. Cisco in Western Australia is an innovation centre and frankly the innovation centre seems to have been come a little tattered. People who spread, spread around, some have moved off to their Centerport innovation centre that so know the engagement is still very lively and there's a lot of dialogue and I don't want to distract our sales team terribly with a research centre that will never really be a customer. They provide support to their end customers. But the Cisco trials went well.

And it's research and it's not something that I think would percolate to the top of us delivering those results that we expect to need to do in the near term future. I'm talking about profitability. When did we get to profitability? I can only tell you this, we get to profitability when we hit about two and a half million dollars, US, in revenue, we had a nice quarter this quarter. We expect for the quarter and next quarter, letting as, as we approach that vector to two and a half million, that's when we would become profitable with a 95 per cent contribution margin and then take out the gross burn. Is another question. I think this one is going to be repeated over and over about the possibility of a table and I'll say the same thing over and over.

We're going to build the best company we can. Really get to the highest growth rates and larger revenue we can. Highest levels of profitability and we will see shareholder return on the fundamental performance of the company. Somebody comes in and they want to make an offer that shareholders are willing to accept. We'll entertain it at the time, but right now our focus is on building the best company that we can. There was a question about ABR and their involvement with Intel. I think I touched on that enough. ABR is not completely with Intel and what they do isn't exactly what we do. That could, we could still collaborate. It hasn't been something that's been kind of on the forefront of our engagement. It's really something that's more appropriate once we have Akida to implemented in an FPDA to determine whether ABR can provide some incremental benefits.

Okay, I think that's about it. I'm going to be with our team in Sydney, I guess it's next week actually. Certainly for the AGM. Bob is going to do a demo at the AGM or Brainchip Studio. Peter's going to give people look behind the curtain on Akida and we're going to have a couple of investor meetings, group meetings where we do the same thing. I was hoping to do demonstrations today, but we're all scattered around the world. I'm in one place, Peter's in another, Bob's in a third, but one way or another over the next week or so. Either we'll do something dedicated and put a note out that you can join or maybe we can webcast one of the events that we're going to do next week. So I think with that, we're about done and it was good to talk to you again and certainly we'll talk at the end of next quarter, but, I guess I'm going to see a bunch of guys at the AGM next week, which I'm looking, looking forward to it. Thank you very much.