Ambarella Inc (AMBA) 2022 Q4 法說會逐字稿

Good day, and thank you for standing by. Welcome to Ambarella's Fourth Quarter and Full Year 2022 Earnings Call. (Operator Instructions) Please be advised, today's conference may be recorded. (Operator Instructions)

I'd now like to hand the conference over to Louis Gerhardy, Corporate Development. Please go ahead.

Thank you, Liz. Good afternoon, and thank you for joining our fourth quarter and fiscal year 2022 financial results conference call. On the call with me today is Dr. Fermi Wang, President and CEO; and John Young, VP of Finance.

The primary purpose of today's call is to provide you with information regarding our fourth quarter and fiscal year 2022 results. 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're 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 that we file with the SEC, including the annual report on Form 10-K that we filed on March 31, 2021, for the fiscal year 2021 ending January 31, 2021, and the Form 10-Q filed on December 10, 2021, for the third quarter of fiscal year 2022 ending October 31, 2021. Access to our fourth quarter and fiscal year 2022 results press release, historical results, SEC filings and a replay and prepared transcripts of today's call can be found on the Investor Relations portion of our website.

Fermi will first provide a business update for the quarter and the full year. John will review the financial results, and then Fermi, John and I will be available for your questions.

With that, I'll turn it over to Fermi.

Thank you, Louis, and good afternoon, everyone. Thank you for joining our call today. Fiscal year 2022 represented an inflection year for Ambarella with numerous milestones achieved. Record revenue of $332 million was up 49% year-over-year, with CV revenue more than tripling to exceed 25% of total revenue and driving blended ASPs higher. It was the first year where CV delivered a material contribution to our operating results, enabling non-GAAP operating margin to reach 19% versus 4% the prior year. CV revenue wave 2, smart home, became material during the year as expected, and we are pleased to announce CV revenue from wave 3, automotive, also became material during the year. This occurred more than a year ahead of the guidance originally provided in June 2019.

Turning to products. We expect CV5, our first 5-nanometer SoC, to begin mass production in the second half of fiscal year 2023. While CV3, our second 5-nanometer SoC, a 10 billion transistor automotive domain controller, taped out and is expected to sample this year.

There is a strong and broad-based demand for our CV products. At the end of the year, we have cumulatively shipped more than 6 million CV SoCs to more than 275 unique CV customers across many verticals.

Looking to fiscal year 2023, geopolitical and public health headwinds persist and the market forecasters predict real global GDP and the semiconductor industry growth rates will decelerate. There is a continuation of supply chain challenges. And in February, we were informed our 14-nanometer supply from Samsung will be constrained. At this time, we anticipate an adverse impact to our video processing revenue of approximately $5 million in Q2.

Despite these headwinds, we see the secular forces from the digital transformation remaining very strong, in particular, demand for our deep learning AI processes for IoT end points. A majority of our customers' design activity is now on CV, and we expect CV revenue to reach 45% of our total revenue in fiscal year 2023, driving a further increase in our blended average selling price as fewer of the lower ASP video processors are shipped.

Now I will provide some representative market development activity during the quarter. We hit the road running in the first week of January, hosting our annual technology showcase held during CES in Las Vegas and announcing our new CV3 AI domain controller. We held over 35 live technology demonstrations, including our latest EVA autonomous vehicles, Oculii radar technology running both automotive and security applications, and numerous IoT, robotics and the partner demonstrations. The automotive partner demonstration included ADAS implementation from software partners, Autobrains and Helm.ai; as well as driver monitoring demos from Seeing Machines and Cipia, all running on our CV4 SoC.

On January 4, we launched CV3, our AI domain controller SoC family for single-chip multi-sensor perception, fusion and path planning in ADAS to Level 4 autonomous vehicles. This scalable, power-efficient CVflow family of SoCs provides the automotive industry's highest AI processing performance at up to 500 eTOPS, representing a 42x increase over Ambarella's CV2 and 160x CV22. The first CV3 family member is fabricated in 5-nanometer process technology and enables centralized, single-chip processing for multi-sensor perception, including high-resolution vision, radar, including Oculii software, and other sensor modalities. The family's unique hardware scalability allows automakers to unify their software stacks across their entire product portfolios while reducing the cost and the complexity of software development. We have received significant interest in CV3 from leading automotive OEMs and the Tier 1s following announcement.

In addition to automotive applications, we will be developing CV3 SoC derivatives targeting other markets, including security camera and robotic applications. This new CV3 family SoCs will leverage the increased performance of the CV3 third-generation CVflow architecture, which provides over 3x the power efficiency of the previous CV2 generation.

In January, we announced our new Artificial Intelligence Image Signal Processor. This new ISV (sic) [ISP] architecture use neural networks to augment the image processing done by the hardware ISP integrated into our CV SoCs. We demonstrated it live during the show, showing full frame rate, HD video captured in under 0.03 lux lighting conditions or almost complete darkness. We believe this technology will be game-changing in all of our markets as the higher-quality video data not only improves visibilities in human viewing applications but also improves the accuracy of AI processing algorithms under challenging lighting conditions.

During the quarter, we made a number of automotive partnership announcements. We announced our collaboration with Autobrains, a developer of self-learning AI for assisted and autonomous driving to deliver a scalable range of AI solutions ranging from ADAS to higher-level autonomy for automotive mass market. At the show, we demonstrated Autobrains' ADAS software running on single CV2 automotive SoC. The collaboration will deliver high-resolution front ADAS solution targeting compliance with 5-star NCAP standards.

We also announced our partnership with Seeing Machines, the industry's leading computer vision technology company for driver monitoring applications. And we demonstrated Seeing Machines' software running on our CVflow AI platform. The partnership will enable the delivery of a complete, integrated DMS and ADAS solution.

We're also seeing new opportunities in automotive applications employing cameras for security, access control and customization of personal settings. During the quarter, Hyundai introduced its Genesis GV60 SUV with a face connect feature that recognizes driver's face, opens and closes vehicle doors and provides customized setting for the driver using near-infrared cameras and deep learning technology. The face recognition module is based on Ambarella's CV25AQ automotive SoC.

Also during the quarter, Rivian began shipments of its first consumer SUV, the R1S; and its first commercial vehicle, the new EDV 700 delivery van, designed in close collaboration with Amazon. As announced by Rivian, the R1 family, comprised of the R1S and the R1T, has 71,000 vehicles preordered, while Amazon has placed an initial order for 100,000 of the vans. Both R1S and EDV 700 have a camera platforms based on Ambarella's solutions, with multiple Ambarella CV2 SoCs being utilized for Rivian's autopilot called Driver+, gear guard and surround view systems. The EDV 700 delivery van also includes stereo forward-facing cameras enabled by CV2's built-in stereo processing capability. The rapid design and deployment of these vehicles, in addition to our other automotive production wins, reflects the maturity of our CVflow SoC and development tools.

In February, the U.S. Department of Transportation's National Highway Traffic Safety Administration issued a final rule allowing automakers to install adaptive driving beam, or ADB, headlights on new vehicles. Adaptive driving beam headlight systems are useful for distance illumination of pedestrians, animals and objects without reducing the visibility of drivers in oncoming vehicles. The use of cameras with AI-based perception to intelligently control the headlight beams represents a new opportunity for Ambarella's CVflow SoC.

In December, we announced our early win in this area with Chinese automotive retail technology company, HASCO Vision; and IM Motors, a new automotive technology company jointly created by SAIC Group, Zhangjiang Hi-Tech and Alibaba Group. At the 2021 Shanghai Auto Show, IM Motors announced its new L7 electric vehicle, which began sales in February. It includes an intelligent DLP lighting system based on our CV22AQ automotive SoC that can perceive the driving environment and provide visual warnings to pedestrians as well as automatically adjust the width of the headlight beams under narrowing road conditions.

Also in China, during the quarter, Hycan, an EV vehicle joint venture between GAC Group and the electric car maker, NIO, introduced its Z03 SUV. The Z03 includes Hycan Pilot 2.0, an intelligent driving assistance systems based on our CV2AQ SoC and supplied by Tier 1 MAXIEYE.

Now I will talk about some of our IoT customers' product introduction during the last quarter. In the enterprise camera market, Motorola Solutions announced the expansion of its license play recombination portfolio with the introduction of the L6Q camera based on our CV25 SoC. The L6Q can accurately scan vehicle moving at up to 75 miles per hour and up to 70 feet away. It can be powered by battery, solar panel, or AC/DC power and can operate in complete darkness.

Also, in the enterprise and public market, Verkada introduced its CD42 and the CD52 dome cameras based on our CV25 SoC. And Panasonic iPro sensing solutions introduced its iPro mini, a versatile network camera based on our CV22 SoC.

In the smart home market, Comcast introduced its first video doorbell, Xfinity Video Doorbell. Based on our SoC, it supports 1080p HD video, crystal clear night vision and a wider 4:3 aspect ratio that allows you to see more of your doorstep. In the same market, Ubiquiti introduced its G4 Doorbell Pro, a WiFi-enabled video doorbell with primary 5 megapixel camera and a second 8 megapixel package camera. Ubiquiti also began shipping its AI 360 PTZ camera, which provides panoramic, overhead surveillance and is based on our CV25 SoC.

We are also seeing opportunities in new AI camera IoT applications, including robotics and video conferencing. During CES, HMS, an AI factory automation company based in Japan, announced and demonstrated its new SiNGRAY AI 3D camera designed for robotic and industry 4.0 applications. In this application, Ambarella's CV25 SoC does the camera perception and fusion, combining the RGB camera and the ToF dual sensors. The product is being supplied to Japanese robotic leader, Yaskawa, for robotic arm application.

In the video conferencing applications, an example of one of our first wins is from Minrray, a China-based leader in communications and video conferencing. Their new UV430A Series, an Ultra HD PTZ video conferencing camera, based on our CV22 SoC, offers 4Kp60 video, 25x optical zoom and a pan tilt zoom operation.

In summary, these engagements indicate we are successfully leveraging our state-of-the-art human viewing video processor expertise into markets for high-bandwidth AI processors in machine sensing IoT edge end points. Much of the early CV growth we have experienced has been driven by new product cycles in existing markets like security cameras, but we are extremely excited about the penetration we are achieving in multiple new market verticals, like the robotic arm, automotive headlamp or automotive access control mentioned earlier. The megatrends for security, safety, automation and eventually complete autonomy, or robotics, are key drivers in our market, and it is a key driver for our SAM expansion, increasing from about $4 billion this year to approach $10 billion in calendar year 2027.

With the breadth of our SAM expanding, the market opportunities are now much larger. We are engaged in discussions at multiple customers where the lifetime revenue of any one program could be more than 20x higher than what we have experienced in the past.

With the evidence mounting, we are more convinced in the positive secular trends of our AI products. In order to sustain and grow our leadership position, we have and will continue to make a premium investment in our differentiated semiconductor and software R&D. Our organic R&D investment leads the way and has been augmented with our synergistic acquisition of HD radar leader, Oculii. Our R&D investment includes new perception technology development, deep fusion for cameras and radars, automotive functional safety, next-generation processors on 3- or 4-nanometer technology, and software modules higher up the stack.

And also, I would like to provide an update on Oculii about 4 months after we announced the acquisition. We are very pleased with customer interest and activity with our proprietary HD radar products and many cross-selling opportunities were discussed with customers during CES this year. Since the acquisition, we have seen a significant pickup in the number of radar module customers. In fact, during Q4, we shipped a record number of 77 gigahertz radar modules. We understand many of these customers are evaluating the technology for their production programs, and we plan to continue to aggressively hire to support the strong interest and outlook for our HD radar technology.

In conclusion, after 5-plus years of edge AI investment, fiscal year 2022 represented a firm inflection for Ambarella, and while decisively established, we're still in the very early innings of the AIoT market development. We are very pleased with our progress, evidence continues to build and we remain convinced the market opportunity is very significant. We are committed to sustaining our strong investment to capitalize on our leadership position. And along the way, we are committed to deliver positive operating leverage at the $500 million and $1 billion annual revenue milestones, as indicated during our Capital Markets Day in January.

With that, John will now provide our prepared financial comments.

I will review the financial highlights for the fourth quarter and the full fiscal year 2022, ending on January 31, 2022, and provide a financial outlook for our first quarter of fiscal year 2023, ending on April 30, 2022. I will be discussing 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 and acquisition-related costs adjusted for the impact of taxes.

Fiscal year 2022 revenue increased 49% year-over-year to $331.9 million. Security camera revenue was close to 65% of total revenue, up more than 50% year-over-year. Auto revenue more than doubled year-over-year and represented almost 25% of total revenue. Our Other business declined roughly 20% year-over-year to represent slightly more than 10% of revenue.

For fiscal year 2022, non-GAAP gross margin was 63.4%, up from 61.4% in fiscal year '21. Despite some higher costs and expenses, our customer mix and product mix improved and the pricing environment was relatively stable. Non-GAAP operating expenses increased 15% for the year, with the majority of this increase in our organic R&D investment. For the 14th consecutive year, we reported positive annual operating cash flow, which was $38.8 million in fiscal 2022. With no debt, net cash on hand totaled $171 million at the end of the year.

Q4 revenue of $90.2 million was slightly above the midpoint of our guidance range, with CV ending the year on a strong note. Q4 revenue declined 2% sequentially, well above the 5-year average of down about 10% sequentially. With several new video processor and computer vision programs commencing production, auto revenue increased more than 30% sequentially. Security camera and Other revenue declined about 10% sequentially.

Non-GAAP gross margin for Q4 was 64.8%, above the high end of our guidance range of 63% to 64%. Despite some increased costs, the more diversified nature of our business, a relatively stable pricing environment and a richer customer and product mix all contributed to the strong gross margin performance.

Non-GAAP operating expense for the fourth quarter was $40.3 million compared to $35.6 million in Q3. This was slightly above the midpoint of our guidance range of $39 million to $41 million.

Non-GAAP net income for Q4 was $17.9 million or $0.45 per diluted share compared to $22.2 million or $0.57 per diluted share in the third quarter. In the fourth quarter, the non-GAAP earnings per share were based on 39.7 million shares.

Total head count at the end of the fourth quarter was 899 with about 82% of employees dedicated to engineering, most of whom are focused on software. Approximately 65% of our total head count is located in Asia.

In Q4, we generated positive operating cash flow of $20.6 million. Total accounts receivable at the end of Q4 were $44.5 million or 45 days sales outstanding, in line with the 45 days outstanding at the end of the prior quarter.

Net inventory at the end of the fourth quarter was $45.2 million or down about 4% in dollars from the $47 million at the end of the previous quarter. On a days basis, inventory increased to 128 days in Q4 from 118 days in Q3.

We had 2 customers represent 10% or more of our revenue in Q4. WT Microelectronics, a fulfillment partner in Taiwan, who ships to multiple customers in Asia, came in at 59% of revenue; and Hakuto, a distributor for automotive customers in Japan, came in at 12% of revenue. Revenue from Dahua and Hikvision declined more than 60% sequentially and more than 75% on a year-over-year basis and represented about 2% of total revenue in Q4 and 6% for the year.

I will now discuss the outlook for the first quarter of fiscal year 2023. Underlying demand remains solid, but supply-side conditions are highly dynamic. First, as Fermi indicated, our lead times remain extended and we are now facing new challenges and uncertainty with regard to our suppliers' timing of deliveries for our 14-nanometer video processors. Second, some of our customers have experienced significant delinquencies from other component suppliers. As a result, some customers may choose to defer our shipments. Other customers may build inventory of our SoCs as they await a complete kit. We are also prepared for a potential public health lockdown in Hong Kong, where our main warehouse is located, by arranging alternate delivery routes to our customers.

To the best of our knowledge at the current time, our guidance contemplates these new and existing supply chain challenges. Considering all these factors, we expect our Q1 results to be better than the 9% average sequential decline experienced in the last 5 years. We estimate our revenue to be in the $88.5 million to $91.5 million range or approximately flat sequentially at the midpoint. After a surge in demand for several new programs in Q4, auto revenue is anticipated to decline sequentially, with IoT revenue increasing sequentially. As indicated at our January 4 Capital Markets Day, given our strategy and the related changes to our business, going forward, we will be reporting in 2 revenue categories: automotive and IoT.

We estimate Q1 non-GAAP gross margin to be between 63% and 64%. We expect non-GAAP OpEx in the first quarter to be between $41 million and $43 million, with the increase from Q4 primarily coming from increased engineering head count, payroll tax accruals and other engineering expenses. The Q1 non-GAAP tax rate should be modeled in the 3% to 6% range. We estimate our diluted share count for Q1 to be approximately flat sequentially.

Ambarella will be participating in the Morgan Stanley TMT Conference on March 9 and 10, the 33rd Annual ROTH Conference on March 14 and Bank of America's 12th Annual Global Automotive Summit in April. During Q2, we will also be hosting virtual demos from our recent CES acquisition. Please contact us for more details.

Thank you for joining our call today. And with that, I will turn the call over to the operator for questions.

(Operator Instructions) Our first question comes from Matt Ramsay with Cowen.

Guys, my first question is, obviously, there's some sensitivities and changes with the supply situation from Samsung on 14-nanometer, so a couple of questions around that. I guess the first one is, in the prepared script, it sounded like there was -- you guys found out pretty recently, there's going to be sort of a $5 million impact, I think, you said to Q2. Maybe you could clarify that, how long you feel like that type of or magnitude of impact might continue to be a headwind into the back half of the fiscal year. And I guess the other part of the question is, does that mean that you're not yet seeing any of that impact in the guidance for April?

Well, I think the guidance we just provided for Q1 has considered all of the potential impact. We noticed this problem after Chinese New Year. We have communicated with Samsung many, many times throughout the last few weeks to identify potential impact. And our conclusion is our Q2 revenue will be impacted by approximately $5 million, and we're still working hard on Samsung to secure our second half supply. And we believe that Q2 might be the worst case for the whole year, but we still need to confirm the supply situation with Samsung for Q3 and Q4.

Got it. Supply is more around the longer term. I mean you guys have worked with Samsung really since the birth of the company Ambarella. Is there any change that happened on their side that caused the supply disruption? Or do you get the feel that you're getting the priority with your supply partner that you would like? Are there any changes to that dynamic? And are you considering, going forward, any other sources of foundry capacity?

Well, I think our relationship with Samsung is still very strong. I think you probably noticed that the 14-nanometer supply that we talked about is the same product line that we had a problem last year when Samsung Texas Foundry had a problem. So I think one of the reasons that has popped us is if there's any supply change, we really have no inventory to cover the -- to bridge the problem. So I think that we still get a high priority from Samsung. But I think this shortage of 14-nanometer in Samsung is severe, and we'll just continue to see the result of that.

In terms of considering a second source, we are still working on 5-nanometer right now. And any tape-out, regardless of other engineering costs, any tape-out is $50 million. It's going to be a very costly proposal if we want to consider a second source. So at this point, I think we would like to understand what happened in Samsung and try to work out the catch-up plan so that we don't face a more severe situation in the future. But at the same time, I feel that Samsung is working hard to work with us to reduce the impact.

Got it. Just for my last question, you mentioned that as you look into the April quarter, after some stronger results in the last couple, that the automotive revenue might be down sequentially slightly. Maybe you could talk about the drivers of that sequential revenue decline and how do you think the automotive business is going to trend for growth for the full fiscal year.

Well, first of all, I think our automotive business continues to be strong. This quarterly reduction's more like seasonal but also reflects last year, we had multiple products ramping up at the end of the year. I think that the ramping up really triggered a big surge on the demand, and now we start seeing the regular run rate. So I think the combination of seasonality and this is the reason. But I still believe that our automotive business continue to perform strong.

Our next question comes from Gary Mobley with Wells Fargo.

I wanted to focus a minute on the demand side of the picture. Fermi, you seem to be hedging a little bit in your prepared remarks talking about decelerating industry trends. And so what I'm curious to know is, to what extent is sort of the flattest revenue trends you're expecting to see for 3 consecutive quarters, albeit better than seasonal, impacted by supply and to what extent is impacted by demand? In other words, I guess the most direct way I can ask the question is, did backlog grow during the quarter?

We still see a very strong demand. Particularly on the CV side, we feel strongly that we have a great momentum on CV, and we continue to believe that our CV revenue will be more than 45% of this year's total revenue. And from that point of view, I think we feel confident that we invest heavily on that direction, and we believe we're going to get paid off from that investment.

From the supply point of view, 14-nanometer has been a problem for us, between Samsung and us, for more than 18 months now. So I hope that when the overall industrial short supply become better, this problem will go away with us. For any other process nodes or any other product lines, we have not seen any supply issues from Samsung.

Okay. On the thread of supply chain issues, I wanted to ask about what you're getting produced at Samsung Korea? I presume almost everything CV is in Samsung Korea. Are you seeing any constraints on that particular fab?

Just like any other fab, the older fabs around the globe are very tight, and we believe that the Samsung Korean fab is tight, too, but we are convinced that, at this point, we will not have any shortage on our CV product supply chain.

Our next question comes from Ross Seymore with Deutsche Bank.

It's nice that you sized that $5 million impact. It's not fun to have to deal with that. But any sort of sizing about the supply side impact, if any, on either your fourth quarter that you just reported or the first quarter you're guiding to?

I don't think there's any impact on Q4 and Q1. We just realized this problem just a few weeks ago. So we are dealing with it immediately by disclosing to our investors. We give the potential impact, the size of impact as well as the timing of impact, so our investors can understand the situation.

Great. And I guess as my one follow-up, the non-CV revenue, looks like it grew about 25% last year, which is really impressive. Can you just talk about what the drivers of that growth would be? And do you see them continuing this year ex the supply limitation? I just want to figure out what sort of growth rate we should think about for the part of the business that's not included in the 45% CV.

Right. So last year, the non-CV revenue growth mainly come from the security camera and automotive, particularly a lot of automotive growth came from the non-CV product line. However, last year, I also want to make a note, we continue to lose business from Hikvision and Dahua. We have been documenting the unit number drop. In fact, on John's script, he talked about that we're going to see from 6% last year total revenue from Hikvision and Dahua go down to an even lower level. So I think from that point of view, the non-CV revenue will continue to have pressure because of Hikvision and Dahua. But also, I think that our CV product line also replaced some of the video product line. So I don't expect that the video product line growing another 25% this year. But the shortage from 14-nanometer is all video processor, which is non-CV, and they'll continue to put pressure on our non-CV revenues.

Do you expect the non-CV revenues to grow this year?

I don't think so.

(Operator Instructions) Our next question comes from Andrew Buscaglia with Berenberg.

Maybe you could talk a little bit about, on the margin side, you had a great gross margin quarter but operating margins were a bit lighter than I was expecting only because it looks like some of your operating expenses were a little bit higher. How much was that more so sort of self-inflicted versus anything supply chain related?

Well, there are some supply chain-related issues because I think the cost has continued to go up not only just on wafer side, packaging, testing, even delivery, all the cost is going up. But however, I think the main driver of our operating expense growth is always on the R&D side. In fact, when we tape out our CV3 chip, that R&D expenses on the IT side as well as the contractor side is definitely higher because that's such a huge chip, and we work very hard to put a resource arm to solve that problem. So definitely, our operating expense will be driven by our continued investment on the premium R&D, including the 5-nanometer, maybe even 4-nanometer tape-out and also tools, and also continue to hire further engineers in the future.

Okay. And then maybe one more on automotive, just with the guidance for Q1. You made some comments about customers maybe choosing to defer shipments. Was that specific to automotive? And then just kind of what's your confidence in that, the demand versus a few weeks ago, being there this year than it was prior to some of the stuff that's come to light?

I don't think the demand changed in the last several months or several weeks. The biggest change is -- I think everybody is seeing that the automotive industry continued to have a supply problem, not only our supply problem but also the other materials, right? In fact, we continue to see our customers cannot build a product because of the lack of other components, microcontroller for one. So I think that's probably the biggest variable we are dealing with. I don't see a demand change. I still continue to see the uncertainty of supply is the biggest problem.

Our next question comes from Joe Moore with Morgan Stanley.

To the extent that you guys talked about these incomplete kit type issues where they may have excess inventory of your product, what markets do you see that as being more likely than others? Does it span all the markets? And is there anything you could do to help us kind of quantify what that impact might be?

Well, we start seeing it, we see pockets of the inventory, and some of it is automotive, some is security. It really depends on the customers' purchasing criteria, right? Some customers continue to buy our chip until they realize that they cannot get other components. Some people stop earlier, so they don't build inventory of our chips. So we see small pockets here and there across all the markets.

Okay. And any sense for the magnitude of it? Like is this extreme by historic circumstances, do you think?

I don't think it's a huge amount yet. Otherwise, we should make it visible to our investors. But we want to talk about what we're seeing in this, and because we start seeing people pushing out the demand because of that. But I don't think that's a severe problem just yet.

Our next question comes from Kevin Cassidy with Rosenblatt Securities.

Maybe just another way to ask Joe's question, in the security camera business that was down 10% quarter over-quarter, was that mostly professional? Or was it the home security camera?

I think it's both. I think both went down. But also, you have to remember, seasonality played a major role in there. If you look at our last 5-year trend, both professional and the consumer market went down. So it's not a surprise to us.

Okay. And maybe with the shortage of supply, are you getting any manufacturing cost increases?

In the last 18 months, we continue to get increase, yes, and we continue to see that.

Okay. And can you pass along that cost increase to your customers?

To a certain extent, yes. In the past, we did pass some of the cost increase to some of the people when they try to expedite the delivery. But this time, with the 14-nanometer problem, we did, because of the cost raised to a point that we have to reflect it to our customers.

Our next question comes from Tore Svanberg with Stifel.

Just a clarification question on the foundry situation, so is it purely just capacity? Or are there some other technical yield issues? Because it sounds like you've had some issues for the last 12 months. Is it just purely capacity?

I think right now, what we're seeing is that we don't get enough allocation. But what's the real reason behind it, I don't want to speculate because Samsung didn't tell us. But I think it's become clear in very early January that they either got overbooked or they have other issues that we are not familiar with that they had to reduce the allocation to us in Q2.

That's very helpful. And can you just elaborate a little bit more on the timing for CV5 and CV3? So I think you mentioned revenue contribution to CV5 late this year. Are you expecting CV3 revenue maybe second half of next year still?

First of all, yes, you are right, CV5, we're expecting second half revenue from customers and, in fact, both from the security camera as well as some of the traditional consumer product. For the CV3, we are sampling to our customer, I would say, in second quarter this year and the revenue will come much later than that, right? This is really an automotive customer we're talking about.

Our next question comes from Quinn Bolton with Needham.

I wanted just to ask on the video processor side. Do you think that, that demand is perishable, meaning that if you don't get the allocation from Samsung in the second fiscal quarter that, that demand goes elsewhere? Or do you think that it simply pushes into fiscal Q3 or fiscal Q4?

Quinn, my gut feeling is this, if this is a 1-quarter problem, I think our customers probably can wait and their customers can wait. But if this drags along further, I think some of the revenue will become perishable. So it really depends on how fast that we can work with Samsung to control this problem and provide much better visibility to our customers for Q3 and Q4.

Got it. And then Hik and Dahua are down to 2% of sales. Do you think that they recover over the next couple of quarters? Or do you think that they just continue to represent a very low single-digit percentage of revenue going forward and effectively become immaterial at this point?

Yes, I think they will stay at that level. But last year, if you look at the whole year, they are still probably 6% of total revenue. And this year, they will be a low single-digit and probably stay there until it becomes nonmaterial.

Our next question comes from Vivek Arya with Bank of America.

The first one I actually have is on your balance sheet. When I look at the cash level, it's strong, but it's kind of at the lowest level since fiscal '15. And I'm curious, Fermi, will you need to raise more, whether it's to execute on the R&D road map? Or more importantly, will you need to provide long-term supply agreements with your foundry suppliers? Because we are seeing many of the computing payers kind of prepay for capacity. Is that something that you will also need to consider, so you are assured of supply going forward?

Right. So they are 2 separate questions. So first question is, because of acquisition of Oculii, our cash flow dropped to $150 million. But I think you also noticed that in this quarter, we generated $20 million cash flow. So from this point of view, if we don't need the extra cash in the near future, we don't need to really raise any funding. But you made a great point that there are many companies out there trying to commit cash to secure their foundry. That's something we are considering to do. But however, that if we want to do it, really, we can get a line. We talked about this last time that we can easily get a line to help us to bridge that demand. So I think worst case is we're going to get a line, but we are not considering any secondary at this point.

Got it. Very helpful. And for my follow-up, just a quick one on Q2, the $5 million impact. Does it kind of mean conceptually, Q2 sales closer to $85-ish million? Or are there kind of positive offsets from the CV side? And are there any gross margin implications because you had very strong margins in Q4, but you're kind of getting back to a trend line in Q1. What happens when you have this additional supply issue? How should we think about just the puts and takes around the sales and the gross margin, the best insight you have today would be very helpful?

Right. I think when we said $5 million, it really directly impacts our top line. So it's really just revenue. We believe we lost $5 million revenue in Q2. So that's probably answer to your first question. From a gross margin point of view, there is a combination of cost increase as well as we talked about we have to raise the price to certain 14-nanometer customers. And that combination, we are factoring in our guidance. So that's why you see it on our guidance on the gross margin line. So I think we continue to believe that long term, our gross margin is 59% to 62%, but in the next quarter it is based on the 63% to 64% guidance right now.

Our next question comes from David Kelley with Jefferies.

Maybe just a question on customer and product mix, clearly, it's been a tailwind. How are you thinking about the sustainability of that benefit into '23?

Well, I think that from a computer vision point of view, I think all the indications that we have really makes us to believe that we have very strong support on not only customer base but also revenue growth. We talked about from 25% total revenue last year to 45% total revenue this year. And also, we talked about the customers that we engage and a number of customers that we take into production. So we continue to believe that story is very strong, right? So now I think this 14-nanometer product line shortage definitely will impact on the video processor product line and also put impact on the revenue but also some design win activities. People probably want to look at in Q3 and Q4, if we cannot deliver, how they can satisfy their revenue. So I think that's definitely an impact that we need to evaluate quickly and work with Samsung to solve the problem.

Okay. That's helpful. And then maybe one more, following up on the earlier automotive deferral conversation. I was hoping if you could talk a bit more about the timing of what you're seeing there. Did your customers start to pull back in January as some of the production disruptions were picking up again? Or is this something that you were seeing earlier in Q4 as well?

Yes. Just to be clear -- David, this is Louis -- we didn't talk about an automotive deferral. In fact, our automotive revenue increased 30% sequentially in Q4. As we had a number of new programs kick in, in multiple geographies for both video processors as well as computer vision. And so in Q1, we do see our automotive business down a little bit sequentially, but we do expect auto to be a key driver of growth in the next year.

Okay. One more, if I may, just to follow up on that. The Q1, is that more of a seasonal pullback? Or is there anything from a customer standpoint where they're pushing out order books with some of the OEM production disruptions?

Yes. No, I think it really comes down to some several new programs that kicked in, again, for both video processors and computer vision. But now that they've established their initial inventory, now their order rate comes back to a normalized run rate. So those are the dynamics we think caused it.

Our next question comes from Tristan Gerra with Baird.

Any way you could help us quantify the revenue contribution from the CV5 ramp in the second half? And then embedded in that question, given the commentary that inventories are not a huge problem yet and suggesting it gets worse, any attempt to gauge when the inflection point quarterly is going to be this year?

So for CV5, we're comfortable with the ramp occurring in the second half of the year. There's multiple customers, multiple product areas where that's going to occur. So there's no change there and no change with CV3 that we expected to sample in the first half of the year. I didn't understand your question about inventory. Could you just rephrase that or repeat it?

Yes. Fermi, during the Q&A, mentioned that inventories were not a huge problem yet, as part of an answer to question. So it suggests that it might get worse in terms of inventory corrections. I'm just trying to look at the timing of where we should expect an inflection point.

Well, that's a question that we are trying to find an answer for because right now, there are all kinds of supply chain issues that are out of our control. A lot of our customers are pushing out their demand of our silicon because they cannot find other silicon like a microcontroller, WiFi chip or PMIC. So until they get better visibility, we will not get a better visibility from that. So I think that it's really reflected in this industrial problem that until the shortage of other components are getting better, we are going to have a very limited visibility on how much our inventory is sitting on the supply chain.

Our next question comes from David O'Connor with BNP Paribas.

Great. Maybe one or 2 on my side. Maybe firstly, a question for John on how we should think about the OpEx growth for this fiscal year ahead, should we kind of take what the trend was last year? That's my first question. And then secondly, just a clarification on the seasonality for Q2, typically up, but given the Samsung issue and also the deferred shipments, can you give us any more on kind of the puts and takes there of how we should kind of model that for looking out towards Q2?

Yes. So as far as Q2 goes, I think, really, the only guidance that we're giving at this point is just the $5 million issue. And so beyond that point on revenue in Q2, we don't really want to make any additional comments at this time just because we're looking to get more visibility as we get closer to that.

I think with regard to OpEx, as I think we said in the prepared remarks that our expectation, our plan, is to continue to invest in premium R&D that we've been developing in-house year after year. And so that's going to take on several different initiatives: developing modules for the complete automotive stack, higher up the stack, that's one of the initiatives that we have; developing chips that fully take advantage of the Oculii radar that we recently acquired their technology; continuing to hire. And so I think while we are at an inflection point for the CV revenue and we expect that to be very favorable for us on a long-term basis, I think in the near term, we are expecting to continue to invest. And I think the level of investment that we've seen in fiscal '22 is indicative of what we might expect going forward.

Our next question comes from Martin Yang with Oppenheimer.

My one question is on your business development activities in China. It seems that there are a lot more nonautomotive, non-IT security camera projects going on. Has anything changed as the result of lower revenue exposure to your IT customers in China or that's basically reflective of the normal design activities among the Chinese customers?

Well, in our script this time, we didn't talk about any Chinese security camera customer. We did mention several Chinese automotive customers. There is a video conferencing customer. And one of the interesting automotive application is this automatic beam control for the lighting. So I don't think there's any major change on the Chinese security camera market per se. I think that market today is Hikvision dominated. And we talked about our supply situation to them. And I don't think that we talked about any major changes on the Chinese security cameras.

Maybe if I may clarify my question a little bit, I was asking if your overall business development [if, in fact, the structure on] resources have changed. So I think are you putting more resources into non-security customers or applications?

Yes, that's obvious. Because we realize that security camera in China is a very sensitive industry, and most of our Chinese customers are looking for non-U.S. suppliers. However, in other business, which is not as sensitive as security camera like automotive, like other industry we talk about, we continue to put effort in China because that's still a big market that we can continue to gain market share from. So that's why you are absolutely right, we continue to have a strong team in Shanghai and Shenzhen, helping us through this development in China. In fact, we also see Oculii radar that China can be a very, very good market for them, too.

Our next question comes from Brian Ruttenbur with Imperial Capital.

If I dig down a tiny bit on the security business, can you tell me what percentage of your revenue was security in the fourth quarter?

Brian, this is Louis. Yes, security in the fourth quarter, let me just check, around 60%. And for the year, security overall grew a little more than 50%. It was around 65%.

Okay. So 65% on this fiscal year that just ended. Can you talk then about where you see it at least in the first quarter or for fiscal '23 would be great. But at least the first quarter, where do you see that percentage going? I assume down, from 60% to 55%. How much of a stair step are we talking about in terms of the security revenue as a total?

Yes. So we see that IoT category, where security is the largest part of it, growing sequentially in Q1. But as you know, with automotive growing to be 65%, 70% of our SAM over the next 5 to 6 years, over time, our revenue mix should shift in that direction.

Okay. Do you have a percentage that you're targeting for first quarter that's going to be IoT/security?

No, we don't have a figure for that other than it will grow sequentially. That's our expectation. And auto, we expect to have a slight decline in Q1 given the factors we've talked about earlier.

Okay. And then just finally, on that same point, is this more a demand issue or a supply issue on the slowing on the security side?

I think it's seasonality, right? If you look at the seasonality in the last 5 years for professional and consumer security camera, it's always that seasonality always play a role from Q4 to Q1.

Yes. I mean there's also been the factor in the last 2 to 3 years where Hik and Dahua were more than 25% of our revenue, and you heard now they're in the very low single-digit range combined. So that's been a factor over this period of time.

That concludes today's question-and-answer session. I'd like to turn the call back to Dr. Wang for closing remarks.

Thank you, and thanks to all of you for joining the meeting today. I'm looking forward to seeing you at some other opportunities in the near future. Thank you. Goodbye.

This concludes today's conference call. Thank you for participating. You may now disconnect.