Ambarella Inc (AMBA) 2021 Q2 法說會逐字稿

Ladies and gentlemen, thank you for standing by, and welcome to the Ambarella Second Quarter Fiscal Year 2021 Conference Call. (Operator Instructions) Please be advised that today's conference is recorded. (Operator Instructions) I would now like to hand the conference over to your speaker today, Louis Gerhardy. Thank you, and please go ahead, sir.

Thank you, Chris, and good afternoon. Thank you for joining our second quarter fiscal year 2021 financial results conference call. Our speakers will be Dr. Fermi Wang, President and CEO and Casey Eichler, CFO.

The primary purpose of today's call is to provide you with information regarding the results for the second quarter of our fiscal year 2021. The discussion today and the responses to your questions will contain forward-looking statements regarding our projected financial results, financial prospects, market growth and demand for our solutions, among other things. These statements are subject to risks, uncertainties and assumptions. Should any of these risks or uncertainties materialize or should our assumptions prove to be incorrect, our actual results could differ materially from these forward-looking statements. We are under no obligation to update these statements.

These risks, uncertainties and assumptions as well as other information on potential risk factors that could affect our financial results are more fully described in the documents we file with the SEC, including the Annual Report on Form 10-K filed on March 27, 2020 for fiscal year 2020 ending January 31, 2020, and the Form 10-Q filed on June 8 for the first quarter of the fiscal year 2021, ending April 30, 2020.

Access to our second quarter fiscal 2021 results press release, historical results, SEC filings and a replay of today's call can be found on the Investor Relations portion of our website.

With that, I'll turn the call over to Dr. Fermi Wang.

Thank you, Louis, and good afternoon, everyone. First of all, I would like to provide an update on how Ambarella is responding to the health pandemic. We are focused on ensuring the health and safety of our employees, their families and their communities, and to safely and effectively serve our global base of customers. The majority of our global workforce has returned to their offices, although our U.S. employees continue to work for home. Ambarella's workforce has always been globally distributed and collaborative, and we believe this culture and structure helped us adapt to remain highly engaged and productive during this global health crisis.

Our business is facing tremendous cross-currents. On one hand, AI computer vision has become pervasive. We are embedding AI in all of our new products, and we have strong and growing evidence of market acceptance. On the other hand, the pandemic and geopolitical risks are high. The economic constraints of the pandemic remains and we are managing the operational challenges.

Despite the challenges, in Q2 we delivered results slightly above the midpoint of the guidance we provided three months ago. Q2 revenue of $50.1 million was down 8% sequentially and down 11% versus the same period a year ago. A healthy customer and product mix, supported by strong operational execution, all contributed to gross margin above the high end of our long-term model.

Our AI business continued to show growing signs of acceptance with the CV revenue now representing a mid- to high single-digit percent of our total Q2 revenue, and we are still confident that 10% of our total revenue will be earned from CV products in fiscal year '21. Our CV SoC command an ASP more than 2x our non-CV ASP, and we are now seeing our overall blended ASP increase. With revenue growth returns, we expect the higher gross profit dollars per CV unit to drive positive operating and EPS leverage for shareholders.

I will now talk about our markets and the customers.

We have previously spoken about our ability to gain SoC share in the professional security camera market outside of China, and this continues to play out. With continued geopolitical uncertainty, we are now seeing additional opportunities to gain market share within China, in particular at the high end and the middle end of the market.Chinese IP camera makers had become concerned about continuity of supply of existing solutions and have restarted evaluating camera designs based on Ambarella solutions, including our CVflow AI SoCs.

Outside of China we are continuing to see strong design win momentum for our CVflow SoCs in professional IP cameras in all geographies and a very strong pipeline of new customer introductions planned over the next twelve months.

During the quarter, a leading Japanese camera supplier,PIPS, previously Panasonic, continued to introduce more new models based on our CVflow SoCs following their extensive product rollout in the previous quarter. This include a 4K Vandal Resistant Outdoor Dome Network Camera and a 4K Outdoor Bullet Network Camera, both based on Ambarella CV22 SoCs. AI features include Smart Coding Detection, Recognition and third-party application support.

And in August, Motorola Solutions announced the availability of an Ambarella-based Avigilon H4 Thermal Elevated Temperature Detection solution. It is a pre-screening solution that can be used to detect indications of elevated body temperature in a person. Composed of an infrared-spectrum thermal camera and featuring edge-based analytics, it provides a contactless alternative to traditional screening methods.

In the consumer IP camera market, we have seen a rebound in orders from the major home monitor camera makers, following initial push outs of orders earlier in the quarter as a result of the impact of COVID-19. Our customers offering monitored security services have seen an increase in the number of installations in the second half of the quarter, while demand from our customers in the self-installation market segment had now recovere., We are also winning new designs for our CVflow SoCs in future generations of home monitoring cameras as customers demand advanced AI features such as smart person detection.

And in July, Taiwanese baby monitor and sleep analytics company, Cubo, introduced its Cubo AI Plus monitor based on our S2L63 SoC. The camera's face detection technology alerts you if it recognizes that your baby's mouth and nose are covered or if they are stuck when rolling over.

During the quarter, we held two on-line webinars to promote our Janus access control reference platform, which enables contact-free access control based on 3D face recognition. The design has been developed jointly with ON Semiconductor and Lumentum, and features our CV25 SoC to enable 3D sensing capability using a single camera. The two webinars, hosted by Lumentum and IPVM, the leader in video surveillance reporting and research, attracted over 200 potential customers and partners, helping to secure our first design wins for the new platform.

And in the automotive market, we are seeing strong interest in our SoC solutions for a variety of new OEM applications including from ADAS, Level 2+ autonomous vehicles, electronic mirrors and car recorder applications. in particular our CVflow AI SoC winning designs based on their combination of low-power, advanced data processing, AI processing performance and ability to process multiple video streams. Since the beginning of the year, we have won now new OEM design wins that we estimate will generate approximately $200 million in revenue over the lifetime of their production, typically between two and seven years.

During the quarter, we won 2 designs for Japanese automaker, Nissan. Chinese Tier 1 Longhorn introduced a pre-installed single channel WiFi DVR for DongFeng Nissan passenger cars based on our A12A automotive SoC, while Tier 1 HangSheng also introduced its dealer-fit dual channel DVR for global Nissan models based on both our A12A and H22A SoCs.

And during the quarter, Shanghai MAXIEYE began shipping its front ADAS solution in Shanqi trucks, a leading commercial vehicle maker in China. This is a new customer for MAXIEYE, in addition to the Yutong win which we announced on March 3. The ADAS system is based on Ambarella's CV22A, CVflow SoC and runs MAXIEYE's deep learning algorithms to achieve high-accuracy visual perception of 3D sensing modeling using monocular vision.

In August, U.K. and European dash cam leader Nextbase launched its next-generation 622GW 4K Dash Cam, the first dash cam to offer life-saving emergency SOS alert capabilities. Based on Ambarella's H22 SoC, the dash cam provides 4K HD recording and image stabilization, while a Rear Camera module also supports 1080p video. Also during the quarter, Aoni introduced its dual channel car recorder for leading ecommerce company Sunvalley. The after-market car recorder features 4 Meg plus 2M resolution recording, and is based on our H22A automotive SoC.

As we mentioned in previous earnings calls, we are winning designs for automotive fleet management designs, where our CVflow SoC provides the required AI performance for driver monitoring and front ADAS applications.

During the quarter, EyeSight Technologies, an Israeli company providing driver and cabin monitoring systems and software, announced it has selected CV25 for its telematics fleet device. The device will implement EyeSight's Fleet Sense driver monitoring solution with EyeSight's algorithm being executed on Ambarella's CV25 CVflow SoC. The CV25 has been selected for its capability to run sophisticated AI algorithm at low power, best-in-class image processing and ability to support additional applications such as video recording and streaming.

In summary, we are addressing AI megatrends such as security, safety and efficiency, all enabled by the integration of our leading video processor and our new computationally intense AI processor. We continue to offer evidence of a broad and expanding customer adoption of SoCs. For example, during Q2, we had a record quarter of CV activity with more than 80 unique customers purchasing engineering parts, evaluation kits and/or development boards, with activity roughly split between our security and automotive businesses.

We are comfortably on track to achieve our guidance for professional security CV revenue, or wave 1, to become material this year and we are on track to reach our similar wave 2 goal for the smart home market in calendar year '21.

Our wave 3 guidance for auto CV revenue to become material in calendar year '22 and '23 is also on track. For example, during Q2, we had six new automotive CV production customers. Our long-term automotive revenue funnel is just starting to build with multiple projects either "won" or in the bidding pipeline. For example, year-to-date, we have won automotive projects representing approximately $200 million in lifetime revenue. Some of the more significant programs start production in calendar year '22 or calendar year '23, and we will be providing further updates as the overall sales funnel develops.

I would like to thank our employees worldwide for adapting to the changing conditions and delivering strong results, and I am thankful for the support of our customers, vendors and shareholders during these volatile times.

I will now turn the call over to Casey, who will give you more details about what we are seeing and expect for the business. Thank you.

Thank you, Fermi, and good afternoon, everyone. Today I will review the financial highlights for the second quarter of FY21 ending July 31 and provide a financial outlook for our third quarter ending October 31st.

During the call, I will discuss non-GAAP results and ask that you refer to today's press release for a detailed reconciliation of GAAP to non-GAAP results. For non-GAAP reporting, we have eliminated stock-based compensation expense adjusted for the impact of taxes.

Our revenue of $50.1 million was slightly above the midpoint of our original guidance. This represents a decrease of 8% from Q1 and a decrease of 11% when compared to the same quarter in the prior year. In Q2, revenue declined sequentially in automotive and other markets, with the security camera business increasing in the low single digits sequentially. Professional security was down and home security increased sequentially.

Non-GAAP gross margin for Q2 was 62.4% compared to 59.1% in the preceding quarter and was above the high-end of our guidance primarily due to a professional security concentration shift from China to other parts of Asia and North America. Margins were also supported by continued operational execution in a very difficult environment.

Non-GAAP operating expense for the second quarter was $30.2 million,compared to $31.9 million for the previous quarter. OpEx decreased and was below the low end of our guidance, primarily due to reduced employee expenses, lower depreciation from deferred capital expenditures and a onetime foreign subsidy related to COVID-19 of $700,000.

Other income of $1.3 million reflected the impact of lower interest rates.

Non-GAAP net income for Q2 was $2.1 million or $0.06 per share compared to non-GAAP net income of $1.3 million or $0.04 per share in the first quarter. In the second quarter, the non-GAAP earnings per share were based on 35.4 million diluted shares as compared to 35.2 million diluted shares in the prior quarter.

The non-GAAP net income for the second quarter of fiscal year 2021 includes a change in the non-GAAP tax rate calculation to exclude losses from jurisdictions where there is no tax benefits associated. This was done to improve the alignment of non-GAAP income tax to the non-GAAP profit before tax.

Accordingly, non-GAAP net income and non-GAAP earnings per share for the second quarter of fiscal year '20 ended July 31, 2019, have been adjusted for the change in non-GAAP income tax effect and presented consistently with the second quarter of fiscal year '21 presentation. A reconciliation of GAAP to non-GAAP calculation is included in the financial statement portion of our press release.

Total headcount ended for the second quarter was 758 with about 81% of the employees dedicated to engineering. Approximately 69% of our total headcount is located in Asia.

Cash and marketable securities were $410.7 million, down slightly from $411.3 million at the end of the first quarter. In Q2, we had negative operating cash flow of $2.4 million due primarily to increases in operating assets, while current liabilities remain relatively unchanged quarter-over-quarter.

Total accounts receivable for the second quarter were $23.3 million or 43 days sales outstanding. This compares to accounts receivable of $20.7 million or 34 days sales outstanding at the end of the prior quarter. Net inventory at the end of the second quarter was $23.9 million compared to $22 million in the end of the previous quarter. Days of inventory increased to 109 days in Q2 from 91 days in Q1.

We did not purchase any shares of common stock in Q2. In May, Ambarella's Board of Directors approved an extension of the current $50 million repurchasing program for an additional twelve months ending June 30, 2021. As of today, there remains $49 million available from the $50 million repurchase agreement authorized through June 30, 2021.

We had two 10% plus customers in Q2. WT Microelectronics, a fulfillment partner serving multiple customers, came in at 56% of revenue; and Chicony, a Taiwanese OEM who manufactures for multiple customers, came in at 20%.

I will now discuss the outlook for the third quarter of fiscal year '21. Forecasting continues to be difficult with geopolitical and pandemic risk continuing to create an environment of low visibility. It is a challenge, for us and for our customers, to confidently forecast government, business and consumer spending for our products.

The rate of order pushout and cancellations we previously discussed in the March 3rd and June 2nd earnings call has slowed, and design activity has begun to recover. However, not all customers' programs have returned to their pre-pandemic revenue trajectory and global political tensions remain high with a wide variety of risks.

In our prior earnings calls, we estimated 2 professional security camera customers in China had pulled in roughly $10 million of revenue from fiscal year '21 to fiscal year '20. We believe there has been some inventory reduction, but we continue to expect weak and/or volatile ordering patterns from these customers. During Q2, these two customers combined represented a mid-single-digit proportion of our total revenue.

Based on these factors and our best judgment at the current time, we expect total revenue for the third quarter ending October 31, 2020 to be in the range of $52 million to $56 million. We anticipate security camera revenue will be down sequentially, with automotive revenue flattish and "other" revenue expected to increase sequentially.

We estimate Q3 non-GAAP gross margin to be between 60% and 62% compared to the 62.4% in the second quarter, as customer, product mix and operational control help sustain gross margins in our long-term model range of 59% to 62%.

We expect non-GAAP OPEX in the third quarter to be between $31 million and $33 million due to continued lower travel, equipment and employee expenses related to the current environment.

Q3 other income should be modeled around $1.2 million. The Q3 non-GAAP tax rate should be modeled at approximately 15%. We estimate our diluted share count for Q3 to be approximately 35.7 million shares.

Ambarella will be participating in the Colliers Institutional Investor Virtual Conference on September 10th, the Deutsche Bank Virtual Conference on September 14th and Jefferies Hong Kong Virtual Conference on September 15th and 16th. Please contact Louis for more details on these events.

Thank you for joining us today. And with that, I'll turn the call back over to the operator for questions.

(Operator Instructions) And our first question comes from the line of Matt Ramsay with Cowen.

For my first question, I just wanted to follow up on something that you guys just said in the prepared script. I was surprised, I guess, to hear that the combination of Hikvision and Dahua, I think, Casey, you said mid-single digits of revenue, if you could confirm that that's right.

And I guess the follow-up question to that. For me, if you could talk to us about the makeup of your current professional security camera business in terms of customers within China, outside of China and what kind of customer concentration there is in that business at this point if, in fact, those 2 big guys are down to that level of revenue currently?

Well, I think, first of all, we confirm that it's mid-single digit of -- between Hikvision and Dahua. I think there are 2 factors around that. One is they definitely built safety stocks last year. I think they definitely draw on those safety stock at this point. And also, we talk about the -- there is a possibility of -- they are looking for a second source because they worry about the geopolitical situation. So that's definitely 2 things happening.

But also, in my script, I talk about another thing, which is not impacting our revenue yet. We talked about our CVflow SoC getting more design win opportunities inside China because -- also because of the geopolitical situation and our -- the -- our customer's current supplier might be facing some supply issues. So that's why we start seeing some momentum in the design win activity in China.

However, we do believe our technology, especially, has been proven outside China. We don't have really a concentrate -- a revenue concentration problem because most of our customers outside China, although there are some equipment in the others, but I don't think we have like a more than 10% customer in any account.

Outside China, we still continue to believe we have the major market shares for both CV and non-CV professional IP camera supply. But inside China, I think that the mid-single-digit revenue number reflect the 2 factors I just talked about. But at the same time, I believe that we will have more CV revenue inside China -- or more CV market share inside China in the near future.

Got it. That's really helpful. As a follow-up question, it was nice that you guys provided some revenue pipeline clarity within the automotive business, I think the $200 million. It sounded like that was a 2- to 3-year revenue lifetime, is that right?

And maybe you could talk to us about how many design wins are in that. What are the relative sizes and concentrations of them? I'm just trying to get a sense of do we have a few big deals that you could build upon in momentum? Is this a whole lot of smaller deals? So just any kind of color as to what that pipeline looks like would be helpful.

Yes. First of all, I think the pipeline, you talk about the $200 million design win, is only for the -- from the beginning of the year to now, all the design wins we have in the automotive space.

We think that there are some big deals, there are some smaller deals. When we say big deal, we are talking about the lifetime revenue close to $100 million in one design. And of course, there are multiple smaller deals for different spaces. And the applications cover ADAS, e-mirrors, recorders and also in-cabin monitoring. So there are multiple different applications in there. And also, the other thing I can say, that amount of $200 million, more than half of that is CV-related applications.

And our next question comes from the line of Tristan Gerra with Baird.

At this point, and I know you're only guiding a quarter away, is it fair to assume that we should see normal seasonality in your fiscal Q4 top line? And I'm basically asking the question in reference to the prior quarter commentary about the full year revenue guidance year-over-year.

Well, we haven't provided any guidance beyond Q3. But I would say that although pandemic, we start seeing less impact of the pandemic in the coming quarters, but we still -- the impact is still there. However, even with that, we still believe the seasonality is there, maybe not as dramatic as in the past years, but definitely, I think that our Q3, is that revenue will be higher than Q4, for sure.

And also, let's assume that the pandemic impact will be really minimized at the beginning of next year. We should see a positive growth next year also. So although we cannot give you a really concrete guidance yet, but I think that's a trend we're talking about.

Okay. That's useful. And then if I look longer-term, if you could remind us of the operating leverage that exists in the model a few years out, is it fair to assume that SG&A should grow at half the weight of your revenue growth? And then what about R&D? If you could maybe perhaps give some tidbits on the operating leverage beyond just the mix as you're back on year-over-year growth starting next year.

Yes. While there will be some growth in sales, marketing and SG&A, really, the majority of the growth in the OpEx area isn't in R&D. And that's not only continuing to grow the team and expand our effort from a headcount perspective, but it's also the -- although it's amortized, it's the increased tape-out costs, CAD tools, et cetera, that as we go down from the geometries that we're at now, down to the 5-nanometer that Fermi talked about in our last call, those costs also increase as well.

The leverage for the model that will offset that growth is really the 2x ASP for CV versus non-CV. While the margins are roughly the same or in the -- or our margin target is roughly the same, that allows us to have an operating leverage that should get us back to the levels that we were at, 25% plus, over the course of time as CV matures out in the model.

Great. And then last quick one. Are you guys planning, or maybe I missed it and if you could remind me, support for LiDAR in the future given that all your demos have centered around camera support for your CV chips?

Right. Although we are very a vision-centric company, but in the long run, we do believe we want to be in the central computers in a Level 2+ car or even moving forward with Level 4, Level 5. So with that goal, we need to support not only this vision-based processing, but also, we will be able to interface with other sensors like radar or LiDAR or even other potential sensors out there. And then we can take in those sensor data, we can do a -- helping our customer to do a sensor fusion algorithm on the CPUs on our chip.

I think that's our goal, and that definitely, we are planning to do that, and in fact, that the functional safety chip that we present, CV2 and CV22 functional safety chip, both of them can support taking in LiDAR data and a dual sensor fusion online.

And our next question comes from the line of Joseph Moore with Morgan Stanley.

It's actually Morgan Stanley. I wanted to come back to the China surveillance. I guess it seems to be really positive that those 2 customers are down around mid-single digit and you're seeing design win traction potentially there, when you think about how that business used to be, kind of 1/4 of your business. So am I being too optimistic about this? And I guess what is the alternative, if they -- if HiSilicon is compromised in their ability to service those customers, and you're starting to see some traction. I mean shouldn't you start to see a pretty strong potential ramp? Is there a third competitor? Are they enabling new suppliers to sort of offload their U.S.? Just how should I think about where that mid-single digits can go?

Yes. Absolutely, Joe. First of all, I'm not sure that I catched your firm, right. To answer the question, the -- from the CV point of view, from the high-end and middle-end products, out of the HiSilicon, we don't believe there is a competitor out there yet. However, the rumor is that a lot of customers, and also along with HiSilicon, they built a huge amount of inventory to try to ride out this uncertainty period. So that -- I think that's one thing you need to consider.

Then from the high end and the middle end, I think that we are very strong. You can see the evidence on the outside-China market. We are basically the leaders and probably have the biggest market share among all the possible customers.

On the low-end side, there's a huge amount of Chinese and Taiwanese companies trying to come in to sell $2 to $3, $4 of CV chip. And that's on the low-end side, we do see a lot of competition. But for the people who pay attention to the performance, power number and the best image processing on the middle- and the high-end products, I think we are probably in a very good position.

I would only add, Joe, that I think your interpretation is accurate. I think to be able to maintain our revenue and grow our revenue with the headwinds that we talked about and the change in market share there in the short run has been encouraging to us because we've got a lot of other activity going on, not only in that business but in other businesses. And as Fermi indicated, we look forward to hopefully having that come back and be a stronger piece. But the fact that we're able to maintain the revenue where it is with that level of business there, I think, is positive.

Right. And another point I would like to point out is in Taiwan, we start seeing that there are some shortage on the substrate. And the reason, we were told, was that HiSilicon is pulling inventory in a very, very huge way, and that caused a short-term disruption on the supply side. So that gives me also indication that, for some of the customers in China, they were probably building up huge inventories to try to weather this difficult period of time, and that's definitely something you should consider, too.

Great. Makes sense. And then a separate question. The thermal win that you talked about with Motorola, I guess we haven't seen you guys in a lot of these thermal cams before. It seems like a pretty big category, given everything that's going on. Like is that -- should we think of that as being a sizable opportunity? And just are you seeing -- is there potential for more customers? Is that -- could that be a decent-sized business for you?

We definitely believe that this is a market that we should target at, and we have worked with several customers and Motorola being the #1 out there to ship a product based on our solution. The way we're looking at this market is that a lot of thermal cameras today is just purely thermal information. But we do see huge market pools for asking both CMOS image sensor and the thermal sensor combined in a camera so that they can overlay the temperature on the object applying CV neural network algorithm on top of that. So you know the who -- when you see a high-temperature person, you know who they are and how to do contact tracing. So I think this combination of AI, plus thermal, really help us to -- help our customer to build much better contact tracing algorithm and a system, I think, as definitely, there's a big market for it.

And our next question comes from the line of Ross Seymore with Deutsche Bank.

One shorter-term question and one longer-term question. Maybe either Casey or Fermi. In the shorter-term side of things, I just wanted to walk through the segments just a little bit. It seems like in your guidance for the third quarter, if you have IP down and you have automotive flat, that the other portion has to be up significantly. Is that just because of seasonality? Or is there something else going on there? And also, part of that is why is auto flat sequentially? Most everyone else, at least because the shelter-in-place orders have come off, I know it's early days for you guys in auto, but I would have thought, just the world kind of turning back on in that end market, would have led to a little bit of sequential growth for you in fiscal 3Q.

Yes. We could, by the end of the quarter, see a little bit. But I think the guide to flat is reasonable and to your point, somewhat pointing to where we are in our life cycle in the automotive business, that maybe differentiates us a little bit from others.

The consumer business has been one that typically has grown pretty well in Q3 in front of the holiday season and into the year. And you're right, we are seeing that. That business today is dominated by DJI and a whole series of products that they have, that I think they're having pretty good success with. So we appreciate the relationship and look forward to them having a good season. But I think your assessment is fairly accurate around the consumer. It is really more the normal seasonality, along with maybe a few new product introductions that are looking pretty promising.

And then I guess my longer-term question, to create a bridge from the shorter to the longer, it's nice to see your gross margin upside so much. So any more color about what happened there, not company-specific, customer-specific? But how is that so good in the quarter? And then you mentioned that CV has 2x the ASP. Obviously, that will have higher gross profit dollars. Any sort of at least directional commentary on what that does to the gross margin as well? Do CV gross margins rise versus the prior gen as well? Or is it just the dollars we should think about?

Right. So for the gross margin, I think the main factor is that the Hikvision and Dahua combined have only mid-single-digit revenue this quarter, right? I think that's the main factor. The product mix is always the most important factor for our gross margin number. And moving forward, I think that for CV product, we continue to believe that our CV gross margin will be consistent with the video product gross margin. But however, like you said, twice higher ASP, so we believe it could provide leverage on the EPS, but not on the gross margin.

Got it. And one final one, if I could sneak it in and as a follow-up to your first answer there. For me, if you diversify outside to other Chinese vendors, like you talked about earlier in the call, is it the same dynamic where the gross margin pressure would be there? Or is it a more fragmented customer base regardless of the geography, something that could be less detrimental to your gross margin than Hikvision and Dahua or given their size?

I think that gross margin for -- with other customers will be better. Just Hikvision and Dahua have so much leverage because of their volume. They probably dominate more than 60% of the total security camera market combined. So that's why they have a much better purchasing power. But with other Chinese vendors, we do see that the gross margin with that will be higher than the ones with Hikvision and Dahua.

And our next question comes from the line of Quinn Bolton with Needham & Company.

I wanted to ask about the China market. Obviously, HiSilicon is no longer able to supply its security processors post September 14. Can you give us some sense how much market share does HiSilicon have in the IP security camera market, which I think you guys, in your slide deck, estimated about 200 million units?

So I would say that inside China, they have the biggest market share. And -- but however, I also want to point out, the majority of that is on the low-end side. In fact, if you look at the distribution, easily more than half -- a lot more than half is more on the low-end side, the $2 to $3, $4 of ASP chip. However, the most profitable business for us is really on the middle- and high-end side. And that's where, in the past, HiSilicon has had the biggest market share in China, we have the second.

Moving forward, I will say that will give us some competitive edge on the -- to gain some market share on the high-end and middle-end market. But however, I do want to point out one more time that although HiSilicon cannot ship after September 15, but there is a huge inventory that we believe had been built up to -- for the current customers, and as well is that for the low-end side, I think there is quite a few competitors out there competing purely on the price. So now we're going to see a price pressure on the low-end side.

Outside China, we don't see a similar scenario. I think that we are supplying through all the non-Chinese customers from the low end to high end because they appreciate using a Zen system software across their platform. I think that's one key design requirement for almost all our non-Chinese customers.

I guess if I could get you to maybe even guesstimate for me, I mean, if it's 200 million-unit market, you said earlier, Hikvision and Dahua are 60%, so it sounds like just those 2 customers alone are 120 million units a year. I assume most of that's probably for domestic China. I mean is the mid-range and the high end, I mean, is that 30 million or 40 million units that you share with HiSilicon? Or I mean, I just -- it feels like there's tens of millions of units, that once those customers burn through the HiSilicon inventory, that they're going to be up for grabs. And it sounds like that's a meaningful unit opportunity. Just wondering if you could help us at all try to size what the mid- to high-end market in China might be on a unit basis.

Well, we don't have that number. It's hard for us to estimate how big this is. I really don't want to give you a wrong impression on that. But maybe we should do -- we will -- hopefully, our customer, Chinese customer, going through the production with our CV solution, we can give you more visibility how we look at our CV revenue generated through -- in China market in the next quarter. I think that's probably the best time for us to give you more clear indication at that time.

Okay. Appreciate it. And then, I guess, just a follow-up on the automotive design win pipeline that you talked about being $200 million. I was just wondering, I know you had automotive designs prior to the beginning of this fiscal year. I was wondering, could you give us sort of what the total aggregate design win pipeline is now, or if you can't give that number, is the $200 million the significant majority of the total pipeline. I'm just trying to figure out how important that $200 million is. Is it in relation to the entire pipeline that you've built up since you started sampling CV?

So first of all, in my script, I talked about the sales funnel exercise inside the company. And we are just doing what you're asking and have really tried to evaluate the category. One is design wins that are wholly confirmed with the customer, or design inside the pipeline that I think that we have RFQ or answering questions or doing evaluation, and then that's the secondary category. The third category is the meaningful customer that we think, in the long run, we want to target and we think there's a chance we can win.

So among those 3 things, 3 categories, we put the first and second category into our sales funnel, and we'll start building models based on that. And we are not ready to disclose the number, but I think, in the near future, very near future, we should be able to talk about that number. And I would think that the $200 million number is just a part of it, it's not a majority of it.

Our next question comes from the line of Kevin Cassidy with Rosenblatt Securities.

Just around the professional surveillance market outside of China, you're seeing some weakness there. But is it -- do you think there's pent-up demand? And what kind of -- is your customers building up inventory? And with your inventory going up, are you holding inventory for them for when these products get deployed?

Yes. Hey, Kevin, first of all, welcome back, and nice to hear from you. And from the outside China, we don't see that there's any inventory buildup. In fact, because of the pandemic, people cut back so severely at the beginning of the quarter, and we start seeing people trying to pull in orders from our customers. So I would say there's no -- really no inventory buildup anywhere outside China. And also, we continue to believe that the design win activity, design win momentum as well as that many customers will start introducing more and more CV models in the next 12 months will be helping us to build a significant business outside China.

And also, we are very comfortable with our guidance. And we talked about wave 1, meaningful revenue from professional security camera by the end of the year. I think that's definitely there already. And we believe we can build up on that.

Well, the only thing I would add there is while there are significant challenges, our relationship with Samsung and our package and test partners have been very good. Zemo and Vincent and his whole team have done an excellent job of really managing this. And there's always a chance that you can have issues. Everything, so far, they've worked through very, very well, and Samsung and most partners have supported us very well.

Okay. Great. And maybe if I can turn to the 5-nanometer products. Can you talk a little more about how that schedule is going? And then also, what's your performance and power consumption improvements are you expecting with 5-nanometer?

First of all, we only tape-out the 5-nanometer test chip to Samsung. And we're expecting to receive the test chip in maybe in 2 to 3 months, and then we can start testing it. We believe our first tape-out of 5-nanometer chip will be roughly the end of the year or early next year, and that's our target. And of course, that's definitely another CV chip for our key customers.

And the power consumption saving for 5-nanometer is significant. And we believe, for example, that to do an 8K video, we can reach 2 watts power, which is nobody there to talk about in the market. I don't think that's even reachable for any other people's design. But we do believe that the potential 5-nanometer product can continue to help us to strengthen our position, not only with performance improvement, but also power improvement -- power consumption improvement.

Our next question comes from the line of Tore Svanberg with Stifel.

You first of all, I know you typically don't give out bookings in backlog. But can you talk about your relative visibility for it, the October quarter? And it does sound like where there's the most uncertainty is the professional security market. Would that be a fair assessment?

Yes. I think that we have a pretty good visibility. And like I say, the quarter is, I think, moving very well for us. As far as the visibility going forward, I think it's improving, as Fermi commented on, but there still is challenges. As I mentioned in my script, there still is challenges. We're trying to balance all of this as we go forward. So right now, I think we've got the right inventory level. I think we're doing the right things to meet our customers' needs. And we feel pretty good about the way we've managed that. But it's not like there isn't challenges throughout the quarter that -- or come up from time to time, no question.

That's great. And as my follow-up to Fermi, on the competitive landscape, I understand the geopolitical stuff. But could you maybe elaborate a little bit on your partnership with Samsung, the move to 5-nanometer, and how that kind of sets you up against especially that big competitor you have in China for sort of the battleground next year?

So I really think that our competitor in China, meaning is HiSilicon, they're moving very aggressively on the process now also. So however, like we said before, we believe our power consumption advantage comes from our design, not from the process node, moving fast with process just being defensive so that we don't lose those advantage because we fall behind on the process now. So with the 5-nanometer, we believe we'll continue to be a leader on the performance per watt as well as continue to provide differentiated features, with potential more what die size that we can play with providing actual function and features to our customers. And our competitive landscape, from a security camera, high-end side, didn't change much. On the lower-end side, I talked about many times that there are many smaller companies that are trying to compete.

On the automotive side, I think our main competitor is Intel/Mobileye, NVIDIA and also that NXP and TI and the FPGA company, Xilinx. So I think they are some of our key competitors in this space right now.

Our next question comes from the line of Suji Desilva with ROTH Capital Markets.

So I'm trying to get a profile of the security camera business now versus maybe kind of prior peak or -- how far below prior peak is it, just to understand currently? And what's the mix now looking like versus typically professional consumer? Just trying to understand if consumer can hold up here even if professional is weak next quarter.

Yes. Again, the -- on the professional side, the biggest dynamic change is, over the last couple of years, we've added several customers, albeit not as big as those 2 major customers in China, and a tech win and others that I think have helped give us a little more balance. But the -- in the short term right now, as we referenced, with those 2 customers being a bigger part of revenue in the current situation, we have seen a decline in revenue over the last couple of quarters there. And as we look forward, with the adoption of CV and others, that's all remains to be seen.

But for us, the mix has changed a little bit because we've gotten a broader acceptance across. And even in China, I think, Fermi, you referenced several new customers, albeit smaller, that are starting to engage with our technology. So we're going to, I think, continue to partner with everybody. The 2 customers from China are still very good partners of ours, and we have a relationship where we're trying to be as supportive as possible in any way we can. But that's really the change in the short-term dynamic. And the longer-term is, again, something we'll talk to as it develops out.

Okay. And then my other question is on the automotive CV pipeline that you talked about, the number and the customers. I'm just trying to understand the geographic mix here, and particularly, what portion of that maybe is targeting the China market versus non-China? Is it geographically diversified?

Well, maybe one thing I can say is the -- in terms of design win activity on number of production, I think there are several in China. But if you look at the $200 million, the revenue that we -- the design win we got from the beginning of the year to now, majority is outside of China. So it's -- we get a lot of momentum in China with a smaller design, but also, we have momentum, for example, in the U.S. and in Japan that continue to help us to build up our pipeline.

Our next question comes from the line of Charlie Anderson with Colliers Securities.

Congrats on all the CV progress. A couple of questions on the year-over-year comparisons. That $200 million number that you're disclosing on the pipeline, I wonder if you had a year ago number. So as you folded in all the CV progress, I wonder if you could speak to how much that's improved at the same point in time from a year ago. And then on the mid-single-digit number for the 2 large Chinese OEMS in enterprise security, I'm curious what that number would have been in Q2 a year ago as well. And I've got a follow-up.

Right. So first of all, compare $200 million to the previous year, obviously, $200 million is higher. But like I said, that in the near future, we're going to disclose our sales funnels for both design win and also design pipeline, and that we will give you more indication on our progress we make with our customers in the automotive space.

In terms of Hikvision and Dahua, I think that a year ago, I would say, easily, they are probably 20% to combined total -- of our total business. So basically, that was probably, yes, roughly $40 million last year or more than $45 million last year. And that -- but at the same time, I need to point out one more time, that they pulled in some orders last year from this year to protect themselves from the potential geopolitical situation. Just a reminder.

Yes. Absolutely. And then for my follow-up, we noticed that FedEx has adopted the Daimler delivery van concepts that your CV is involved in with CoROS. I wonder if you could speak to the potential of that deal and then maybe just that end market broadly. And then I'm sort of curious if you could also talk about maybe the content per delivery vehicle as CV enables some of those use cases.

Great. So I think what you're referring to is really, last week, I think FedEx was disclosed as a lead customer for CoROS. And the 2 companies are working on a pilot program, which has our CV2-based system for delivery vehicles and warehouse. We announced this in -- at the CES, but we didn't mention Federal Express, I mean, because that's really our customer, really, our customer mentioned this time, not us.

But however, I think the significant part of this is really showing that our CV chip is mainly a design focused on ADAS and an AD type of application. But in automotive, particularly for delivery trucks, that also helped our customers through managing inventory as well as other possible applications. So in terms of total content, we talk about there are 2 CV2 in there, and the content is roughly between, $80, $90 to $100 a car. And however, I think that the production date is -- I should not talk about production date because, really, it's our customer's information. But I really think that we'll leave the production day of the system -- of this vehicle to our customer. But I definitely say that's part of our pipeline -- not pipeline, I should say, our sales funnel numbers that we will disclose in the future.

And our last question comes from the line of Richard Shannon with Craig-Hallum.

I guess 2 questions for me. First one, if you look at your professional security market with OEMs outside of China, you've talked about having the largest share and I think it's increasing here. To what degree do you see the potential for even more share coming? Have we reached the end of that? It seems like you would have a bit more room to run from that perspective. But I would love to hear what you think about that.

[There's additional growth for us outside China, our customers may start gaining market share against their Chinese competitors outside China. I think that's probably the -- one of the major areas that I think our (non-China) customer is hoping for that they can grow market share there] (corrected by company after the call).

However, at the same time, the -- in terms of total design win available out there for us to win, we're working hard on that. But if you look at when we mentioned probably all of the major professional IP and camera outside China already as our customer for CV, and so the biggest issue for us is helping them to start rolling out more CV models to replace a non-CV median camera they have and that's another way we can grow market share in terms of the total revenue. So I think that's 2 areas that we're working on.

Okay. That is helpful. My second question is, and Fermi, I may have caught your comment in your prepared remarks incorrectly. So set me straight if I'm wrong here, but I think you said you had a very good design engagement quarter with, I think, 80 engagements or so. Wondering if you can characterize that and to the degree, to what you're seeing it in kind of emerging markets like, say, robotics. Can you give us a sense of where you're seeing all the activity and when you think that might come to fruition, particularly in markets like robotics?

Right. So among the 80 companies, majority of them is in our current market with security camera as well as auto, but there are several of them, is working on the robotic applications. And the size of the market is still relatively small compared to our current market. However, again, I said that many times before, I truly believe, 10 years down the road, the robot with the computer vision capability is going to be a major market for us, and we continue to invest on that. Although we don't have a visibility when this is going to be a huge market for us in terms of revenue, but that doesn't stop us from investing in marketing and engineering resource on that so that we can continue to understand the requirement of the market, get an update about a customer's patient, and I think that's definitely meaningful for us that we continue to do that.

The other thing I would add to that is we're starting to see more and more activity in the access control area and we have a couple of announcements that we highlight this quarter, but we've had engagements in the past too, and that's another, I think, interesting area as well.

Thank you. And ladies and gentlemen, this does conclude today's question-and-answer session. I would now like to turn the call back to Dr. Fermi Wang for any closing remarks.

Thank you. And thank you all for your attendance today, and looking forward to seeing you next time.

Ladies and gentlemen, this does conclude today's conference call. Thank you for participating. You may now disconnect.