Good day. My name is Chloe, and I will be your conference facilitator. I would like to welcome everyone to AVA Technologies' First Quarter 2026 Earnings Conference Call. During the opening remarks, all participants will be in a listen-only mode. Following the opening remarks, we will conduct a question-and-answer session. As a reminder, today's conference call is being recorded and simultaneously webcast. I would now like to turn the call over to Andrew Fung, Senior Director of Investor Relations and Corporate Development. Andrew, please go ahead.

Thank you, and welcome everyone to AVA's first quarter 2026 earnings conference call. Joining on the call today are Suresh Salahian, AVA's co-founder and CEO, and Saurabh Sinha, AVA's CFO. Ahead of this call, we issued our first quarter 2026 press release and presentation, which we will refer to today and can be found on our investor relations website at investors.eva.com. Please note that on this call, we will be making forward-looking statements based on current expectations and assumptions, which are subject to risks and uncertainties. These statements reflect our views only as of today and should not be relied upon as representative of our views as of any subsequent date. These statements are subject to a variety of risks and uncertainties that could cause actual results to differ materially from expectations. For further discussion of the material risks and other important factors that could affect our financial results, please refer to our filings with the SEC, including our most recent Form 10-Q and Form 10-K. In addition, during today's call, we will discuss non-GAAP financial measures, which we believe are useful as supplemental measures of AVA's performance. These non-GAAP measures should be considered in addition to and not as a substitute for or in isolation from GAAP results. The webcast replay of this call will be available on our company website under the investor relations link. And with that,

let me turn the call over to Suush. Thanks, Andrew, and good afternoon, everyone. Q1 was a strong quarter at AVA, where we achieved another new quarterly revenue record as we continued executing on our growing commercial momentum. With the rise of physical AI and more industries looking to leverage new levels of perception, AVA is uniquely positioned with our FMCW LiDAR on-chip technology to tap into a broader and more diverse range of applications than what is possible with conventional LiDAR. This quarter, it was especially exciting to see our commercial momentum progressing and expanding to more real-world deployments. This is not only helping drive our strong revenue trajectory, but we believe also further positions Ava to continue our commercial growth with a growing list of customers. In automotive, we achieved important milestones with our customers on their path towards commercialization and production. With Diamond Truck, we began deliveries of our production intent AdBest product, which highlights the maturity of Ava's technology and is a critical step for the OEM's work towards serious production. And in passenger vehicles, we successfully integrated First Atlas Ultra Sensors in the top European OEMs' development vehicles that are jointly working on the AV stack development with this passenger OEM. Outside of automotive, we are growing in defense across multiple fronts, including expansion with Fortera to use Ava's technology in a second autonomous ground vehicle as they look to leverage our long-range and velocity measurements, and notably our wavelength undetectability by night vision systems. We're also seeing interest in other applications in defense, such as drones, for further expansion. In smart infrastructure, we recently unveiled AvaCityOS, our AI-powered platform for real-time intelligent traffic management. Perception has been very positive, and we have already won our first large-scale deployment in Georgia. And on factory automation, we are seeing growing momentum with multiple customers. Our multi-year collaboration with Nikon is entering its next phase of commercial deployment, with Nikon recently launching its next-generation robotic inspection laser radar system called Aptis MV5, which is powered by our EVE high-precision technology. This is part of a multi-year production agreement to use AVUS technology for automated inspection in factories for automotive, aerospace, and energy industries. Beyond all of this, we continue to advance on new opportunities to use our technology platform across multiple physical AI applications. Our focus remains on delivering on existing programs, further solidifying a leadership position with additional wins, and scaling manufacturing to support the expanding demand for AVA's differentiated technology. Now let's dive more into AVA's recent business developments. Starting first with Daimler Truck. As the exclusive long-range LiDAR supplier and primary detection sensor for Daimler Truck's production program, AVA played a critical role in bringing their Highway Level 4 solution to market. And we continue to work very closely with Daimler Truck as well as with Torque, which are developing the AV stack. The start of Atlas C-sample deliveries to Daimler Truck this past quarter marks a major step forward on the OEM's path to commercialization. This is our production intent sensor for DymaTrucks series production trucks and will be used by the OEM to finalize the AV stack validation and operation ahead of their target launch in 2027 and the following production ramp-up. For AWOB, it also represents a major achievement that demonstrates the maturity and readiness of our technology for mass-scale automotive-grade deployments. We are on track to scale deliveries of our Adla sensors to down the truck over the course of this year, which will support the OEM's vehicle fleet rollout and additional milestones ahead of serious production. Moving now to the latest developments in automotive with the global top 10 passenger OEM based in Europe and other automotive engagements. First, on the top European passenger OEM program, where we are the exclusive lighter supplier globally outside of China for the OEM's next-generation Level 3 production program. We have been collaborating closely with both the OEM and the OEM's AV stack provider and in the past quarter began integration of Atlas Ultra in the OEM's development vehicles. This has been going on track and will be used to jointly develop and mature the AV stack. We expect to deliver additional sensors this year to support the OEM's ongoing development and fleet rollout ahead of target starter production in 2028. The top European passenger OEM selection of AVA for its large-scale Level 3 program continues to serve as a strong vote of confidence in our technology and our growing maturity to the rest of the automotive industry. We are encouraged by our ongoing engagements and Oprah Q1 continue to grow our pipelines. This includes our progress on the development program with a global top five passenger OEM that we announced last quarter. The work is focused on the configuration, integration, and validation of our AdLas Ultra sensors for the OEM's next-generation global vehicle platform. We have successfully completed the initial set of milestones and will continue working with this OEM toward the next-generation vehicle program. Separately, we have kicked off our collaboration with NVIDIA following their selection of AVA as the reference sensor for the Drive Hyperion platform. AVA is the reference LiDAR sensor globally outside of China, which has the potential to effectively make AVA a core LiDAR supplier to many leading passenger and commercial vehicle OEMs using the NVIDIA platform globally. As part of this, we are working together on one common platform comprising of the same sensor suite, meaning one common set of cameras, radars, lidars, and NVIDIA compute and autonomy software to offer to these leading OEMs and AV players. And in the past quarter, our teams have made good progress on the integration of our FMCW technology into Drive Hyperion stack, including working together to implement our velocity data path in the Drive Hyperion platform. So in summary, we continue to advance on multiple automotive engagements, including other passenger programs, and high-volume ADAS Level 2 for commercial vehicles. We continue to believe that AVA is well-positioned to secure additional wins given our differentiated performance, balance sheet, and commercial momentum, bringing further validation of AVA's capability and maturity. Searching gears to other physical AI applications, AVA is quickly expanding in defense with Forterra, a leading provider of autonomous ground systems for defense and other complex operational environments. Since announcing our win with Forterra last quarter, I'm pleased to share that Forterra is expanding use of AVA-40 LiDAR to its newest autonomous ground vehicle called MESA. MESA integrates four ATLAS sensors for surround view and leverages our long-range and velocity detection, vehicle positioning, and stealth operational capability in GPS-denied environments. Beyond the growing demand for AGVs, we are seeing new interest in drones and working on opportunities to expand further with existing customers as well as new engagements with defense companies and organizations on both autonomy applications. Over the past quarter, we have also been expanding deeper into the smart infrastructure market, particularly around intelligent transportation systems, or ITS. This is a growing market opportunity as municipalities across the country look to modernize infrastructure to be safer and more efficient. In just the U.S. alone, there are around 15 million intersections and more than 300,000 traffic signals. This is why we launched AvaCityOS, a full-stack intelligent traffic solution that combines 4D LiDAR with edge AI processing and analytics in collaboration with our partners. Compared to traditional ITS solutions which rely on cameras, radar, or inductive loop sensors, CityOS leverages the advantages of Ava's 4D LiDAR to enable operation in all lighting conditions and deliver more advanced detection while preserving privacy. We are very encouraged that CityOS is quickly gaining traction with DOTs and municipalities across the U.S. We have already secured our first large-scale deployment in Georgia, with an expansion to 30 additional intersections in the greater Atlanta area. This expansion comes after a successful initial rollout across multiple intersections surrounding Centennial Olympic Park and others. The area is one of Atlanta's busiest pedestrian corridors where we believe CityOS can help improve roadway, safety, and traffic operations. AVA's ITS team is also actively working with other programs and municipalities on new opportunities, and we believe that a differentiated solution will drive additional deployments over the course of this year. In precision sensing, we are incredibly excited to see Nikon's first commercial laser radar product powered by our EVE precision sensing platform. With AVA, Nikon's next-generation Aptis laser radar system is capable of faster measurements in a smaller, more flexible size, which enables Nikon's major automotive OEM customers, aerospace, and energy production partners to shorten production times, cut costs, and improve quality for volume production. This product is the start of a multi-year production agreement to use AVA's technology in Nikon's products. More broadly, the flexibility of our EVE precision technology is driving new interest to use AVA to cross-manufacturing and factory automation for a diverse set of industries, from automotive to energy production and semi-capital equipment manufacturing. We are engaged with multiple customers on additional opportunities and working toward converting those to design wins as we expand in precision. With that, let me tell the call over to Saurabh, who will discuss our Q1 financial results.

Thank you, Sourouj, and good afternoon, everyone. Consistent with how Ava is delivering on our commercial objectives, our Q1 financial results also reflect our growing momentum. This includes achieving a new record revenue quarter of $6.3 million in Q1, which represents an increase of around 90% year-over-year, driven by scaling sensor shipments across multiple markets and progression on development milestones for major customers. The non-GAAP operating loss was $25.8 million in Q1, which is about flat year-over-year and highlights our ability to maintain operating expenses at similar levels versus the prior year while continuing to scale the business. Gross cash use, which we define as operating cash flow less capex, was $28.1 million in the quarter. Our total available liquidity at the end of Q1 was $224.5 million, which consists of $99.5 million in cash, cash equivalents, and marketable securities, and $125 million in an undrawn facility that is fully available to draw at management's sole discretion. We continue to believe that our performance and liquidity position differentiates us from peers, and together with our ongoing financial discipline, enables AVA to support ongoing programs as well as secure new wins. With that, I will turn the call back to Sourouj for closing remarks.

Thank you, Saurabh. In closing, I am really proud of how AVA is expanding its leadership position with increasing real-world deployments of our unique perception platform across a wide range of industries. Looking ahead, as we continue to see growing commercial momentum and an increasing list of opportunities to pursue, we are keenly focused on execution, both with existing programs and new engagements, while also scaling our manufacturing to support more customers and the increasing demand for our products and differentiated technology platform. And with that, let's now turn to Q&A.

Thank you. If you'd like to ask a question, press star 1 on your keypad. To leave the queue at any time, press star 2. Once again, that is star 1 to ask a question. We do ask that you limit yourselves to one question to allow time for everyone to ask theirs. And we'll take our first question from Colin Rush with Oppenheimer & Co. Your line is open.

Thanks so much. Guys, you know, can you just give us an update on the progress with SOA and CPO solutions for Data Center? We continue to see data management and transport expense ramping pretty aggressively for all applications. Just want to see where you're at from a commercialization perspective with that and how we can think about that coming to market over the next few years.

Hey, Colin, this is Seroosh. Yeah, happy to answer that. So obviously, the AI data center market is a massive market and a number of opportunities. I think maybe just a quick background here. I mean, as most of us know, the first wave, you know, of the AI data center market really started with all the semiconductor companies and the GPU and compute processing driven by faster compute needs. Then the second wave and bottleneck rose from, you know, memory and high bandwidth storage, right? And I think as we're now going to higher and higher speeds, what's become clear is that there's a certain limitation on how much we can transfer data between data centers and RACs. And I think this next bottleneck really is going to be relying heavily on optical interconnects because that's where copper hits a physical limit to transfer data within data centers. So I think, you know, this is obviously a significant, I think, next wave that's coming on. It's a massive opportunity. And why this relevant for AVA, I guess as you mentioned, is because we've spent the past decade creating and perfecting some of the best high-power sources in proprietary silicon photonics technology and making it so that it works to meet the harsh requirements for automotive. And we did it because it simply did not exist. And I think what we are seeing now is those components, Our proprietary technology for high-power sources and silicon photonics has a significant advantage potentially for both performance and cost efficiency compared to what's on the market. So we're seeing some strong, you know, very strong interest for high-power sources in silica platonics from some of the big players in this space, you know, from folks that are making the GPUs like the obvious, you know, NVIDIA and AMD to some of the hyperscalers, Amazon, Meta, others. And I think AVA has a unique technology there in how we do those sources and the silica platonics. We are looking now to take those investments we've made in the past number of years and apply it to the CPL market. But I think initial data is really promising on the performance of those high-power sources. And, you know, we're looking to apply that as we can talk more about this in the next year as we will. But I think it's overall a massive opportunity for us where we are definitely going to be taken advantage of.

Perfect. And then just moving on to other physical AI applications, you know, we're seeing factories as a prime target for optimization. and given the fact that you've got a couple of partners and started delivering with Nikon now, can you talk about the potential acceleration in that market segment with metrology solutions and how we should think about new build versus retrofit applications for the sensors?

Yeah, yeah, happy to. I think in general we are seeing for the industrial market and general physical AI significant demand across multiple segments for us. I mean, as you saw in the earnings today, you know, it's not just about automotive right now where we are seeing significant traction, but in the other physical AI, including both in the industrial market for EVE sensing, we have now the first commercial product of Nikon coming online into the real world and shipping. Nikon already does a few hundred million dollars in just robotic inspection and metrology alone. And obviously, you know, we're taking higher speeds there than automotive with a mix of volumes. But also we're seeing interest and demand from others, importantly, in the e-sensing market for, for example, semi-capital equipment manufacturing, where there's significant investments. It kind of relates back to some of the infrastructure on the IS that we talked about. But folks are using our sensors already in some of those semi-cap factories for various things, wafer measurements, quality control, all that. So I think that opportunity is definitely picking up. You know, we're already shipping in the 1,000-plus type of sensors on the E side. You know, there's orders, you know, coming in for that as well with a much higher volume. So this is definitely an area that we're going to continue to grow. And then separately on the other side of physical AI, I guess, on the, you know, ITS and smart infrastructure, we're seeing growing demand for CityOS. And within the past quarter, we've already had some wins. We were starting to deploy some of the large-scale deployments Georgia. So, multiple segments. And, of course, defense is the other one that we're seeing double-digit growth pretty quickly there. So, we're excited by all the progress. All that means, though, obviously, we're focused on scaling and manufacturing in the next phase as we bring up the products to massive.

Awesome, guys. Thanks so much.

We'll move next to Suji Da Silva with Roth Capital. Your line is open.

Hi, Soroush. Hi, Saurabh. Congratulations on the progress here. Just, you know, with getting closer on the passenger auto OEM and moving toward working on the software and the stack development, I'm just reminded earlier in the auto industry where there was challenges of the software development between the auto OEM and large programs, maybe Volkswagen and Cariad, those kind of concepts. I'm just curious how you think it's happening differently now that it's going to be more likely to hit production schedules and move forward versus having challenges any color that would be helpful.

yeah uh so obviously on automotive we are you know firing on multiple founders right and we within commercial vehicles we uh shipped our asc samples to diamond truck we just announced that today uh it's a critical milestone for the industry because it's you know these are the production intent um sensors and products and i think it's going to be one of the first ovms that we use you know um you know redundant chassis with production intent hardware and sensors for the AV stack. On the passenger car side, in the last quarter, obviously, you know, we just announced the win with the top 10 European OEM, top 10 global OEM, which is based in Europe. And we have made, you know, very good progress across the teams. We are working together with them and all the AV stack partners. We've delivered the first AdVest Ultra samples for integration. I think the key right now is in the next number of months, we're going to be working together on, you know, implementing the sensor data into the stack and also doing fleet operations, fleet runs on the vehicles. And I think all that is pointing to the program being really progressing well on track. You know, if you also look at the timeline, you know, we're not that far away from SOP, right? You know, by 2028, the timeline target is in the SOP, and we're progressing all pretty well to that. So it's kind of around the corner. So the teams are working very intensely together with on-site support and, you know, multiple times weekly engagement. So all that, I think, is progressing good. On the pipeline side, you know, we are also seeing growing interest on the pipeline with both passenger, and we talked about the top five, but also on commercial vehicles and on ADAS applications. It's not only Level 3 autonomy, but also high-volume ADAS Level 2 as well.

And then my other question, which is on the defense market and drones, the defense market, I'm curious, you know, your go-to market strategy, your partners there, are you a sub-con, how are you tapping that opportunity? And then specifically on drones, it sounds like it's a very interesting opportunity, but I'm curious, does the Ava product translate to the flying vehicle drone market easily, or are they kind of changes that we made, any color that would help?

Yeah, so I think in defense, it's definitely an area that we're seeing significant growth and demand in the market. I think in the past consecutive two quarters since we announced even our first win in defense, defense has been a significant contributor of our shipments and also revenues, some of them, you know, a double-digit percentage of product revenues. And I think Forzura is obviously one customer, our first win in there, and, you know, we talked about that. they're expanding on multiple vehicle platforms. These are more ground vehicles or AGVs where we have significant advantage with our technology as well as our wavelength compared to time of flight. That's why this is accelerating pretty quickly. If you recall, this kind of moved on from obviously engagements and then to wind to shipments and deployments with a matter of less than 60 days or so. So it's moving pretty quickly. But beyond that, as you mentioned, And this is obviously not the only area on defense. We are acting as the tier one supplier to these defense companies and innovators. So obviously we're not the prime, but we are the tier supplier to them. But we are seeing significant interest and growth also in the drones applications. We already have some engagements there. Our technology with its long range has some significant advantages. and the velocity with the fact that also it's, you know, stealth in terms of detectability by night vision. So we are seeing that, and we are having some engagements with some of the, I would say, larger, you know, prime organizations that have a significant investment and budget for a drone's application. So as we can, I'll also talk about that.

We'll move next to Matthew Paciuli with Canaccord Genuity. Your line is open.

Hi, guys. Congrats on another great quarter, and thank you for taking my questions. Maybe just to start, you know, I think on the last call you guys had mentioned four commercial wins in 2026 you were targeting. Is it safe to assume that CityOS is, you know, one of those wins, and could you just provide us an update on, you know, how those conversations are going?

Yeah, Matt, happy to. So, I mean, to be honest, we have had multiple wins since the beginning of the year. I think the way also how we count that matters, like for us, you know, with Forterra as the first win in defense, if you look at also NVIDIA and CityOS, we have had multiple wins. You know, we are counting right now, I think, Forterra and NVIDIA on that. And so we're already ahead of track in terms of the four goals. And, you know, that was obviously aggressive because it was, you know, 100% or double from last year, I think, in terms of the targets for goals, but we're well ahead of the schedule. And I think more importantly, as we are growing and maturing also as a company, you know, we see that it's going to be less about just the number of the wins. Of course, these are targets we hit for the set for this quarter. But, you know, really responding to the growing demand across multiple segments and focusing on those main leaders with high volume and near-term potential as well to expand that. But I think we're overall progressing very well on track on the four targets we have.

Great. I appreciate the clarity there. And maybe just as a follow-up, you know, OnCityOS, appreciate, you know, the commentary on just the market size behind that. Do you guys kind of foresee any potential bottlenecks associated with, you know, getting this product out globally? It seems like it could be very dependent on kind of timelines and funding of municipalities and such. You know, if you could just provide some colors to, you know, how those deployments are going and how those timelines and sales cycles work.

Yeah, happy to. I think as a data point of timeline, we started entering this market really towards the end of last year, and it's only been a few months since that timeframe with the team that we have brought on and the ecosystem capabilities that we have, we've been able to introduce this new solution, which is importantly a total solution. It's a comprehensive solution, not just LiDAR. It's sensors, compute, and perception software and analytics software, working together with our partners to deploy that. And obviously, ASP is there for the solution. It's much, much higher than automotive. So I think in the span of a few months, we've had multiple wins there. We talked about already our first win with Georgia and deployments. You know, it's one of the first large-scale deployments in the state, over 30 intersections, which is significant. But also, you know, the team is continuing to grow this, I think, fairly quickly in the space. And part of it is because of the experience and the conversations that have been happening for some time. The other part is there is this growing interest and demand from both at the state level and at municipalities level to modernize traffic management and traffic flow. It's, you know, these are things that have been around for a number of years with very basic technologies like inductive loops in the ground to detect vehicles. So, obviously, that's very ripe with better perception and sensing right now. And there's, you know, significant, I think, budgets and resources allocated for that. So, from a timing standpoint, we're seeing that already. And, you know, right now our focus you asked about internationally is in the U.S. Obviously, each country has different rules and regulations. So over time, I think we may expand to other areas, but U.S. alone, you know, has over, you know, 15 million intersections and, you know, a few hundred thousand signalized traffic signals today. And, you know, you do the math, these are multi-billion dollar opportunities, and that's why it's one that we are going after. So we do expect that this is going to continue to contribute also to our growth in the year term as well.

We'll move next to Richard Shannon with Craig Hallam. Your line is open.

Richard, you may need to check the mute function on your device.

How's that? Is that working now? We can hear you. Okay. Sorry about that. I wasn't on mute, but I'm glad it's working now. Thanks, guys, for letting me ask you a couple of questions here. My first one, Soroush, is regarding kind of similar questions here regarding both the top 10 OEM for which you have a win and the top five, which I guess I'd characterize as an advanced engagement here. But I'd love to get a sense from you what to expect from announcements and updates in the next couple of earnings calls here, either in terms of finalizing designs and hearing about forecasts with the top 10 OEM or getting to and announcing a win with the top 10 OEM. Just want to get a sense of what we should hear in the future.

Yeah, yeah, happy to talk about that, Richard. So obviously in the top 10 OEM, you know, this is a production program that we have already won. And, you know, we have already started the first, you know, phase of development there that's been going and progressing very well. I think in terms of what's coming down, I think, I guess as I mentioned a bit earlier, the key focus right now is enabling the OEM and the AV partner to go and build the fleet with the full AV stack, including the FMSW technology, and do the validation across different regions and make sure that all the KPIs are met and get that ready for launch. And, you know, we don't have – it's a short kind of timeframe between now and 2028, so it's an automotive world that's pretty much lightning speed. So that's moving pretty quickly. I think in terms of what to expect, obviously, as we make those progress, I think, into those milestones for the fleets, and we can talk about that, we will. And I think at some point, that's expected in the next number of months that's coming online, actually. So that's important. And then on the other opportunities, top five and others, we are engaged. Obviously, we delivered on top five the ADMES ultra sensors as well. And we're working through the integration and the validation and testing. I think there, obviously, that's not a one program yet. And we have to see how that progresses with the final decisions towards RFQ, but we're working towards a next-generation vehicle. But we're also, beyond these, are engaged in multiple other Level 3 automated driving, but also ADAS, which is really Level 2, which typically LiDAR doesn't really penetrate, where we see as some interesting opportunities where these will be very high-volume opportunities across passenger and commercial vehicles. So multiple OEMs, they're engaged. Some are in the RFI and some more in the RFQ. but we expect that in the next few months some of those will make decisions as well so and you know we are feeling good and well positioned and I think our goal is that obviously as part of our goals for the four wins this year that we have automotive wins included in that so that's what we also expect that at least you know additional win in automotive for you know hopefully the rest of this year that's come online and I think overall all indications have been pretty positive for the developments we have at

okay uh to your last uh uh comments there solution you've you mentioned this briefly in the press release as well here about looking at l3 even a dash level two here which is interesting because you've typically talked about uh and seem to be targeting more advanced levels of autonomy um and you typically i think have talked about a you know a bit of a higher asp than what other solutions might offer here so being able to hit that pricing envelope is pretty interesting there. So I'd love to get a sense of whether you see the pricing looking attractive for you and any way that you would characterize or quantify the number of programs you're looking at for these lower levels of autonomy. Thank you. Yeah, sure. I think you're definitely spot on

there, Richard. I think in the space, level two and ADAS typically and historically has been enabled more by vision and maybe radar solutions. I think one of the advantages, as we have, obviously, is we kind of have a laser radar product, right? So we call it 40 LIDAR, so with addition of velocity. I think we have multiple engagements there on the ADAS side. What I'm excited by is the fact that our solution, beyond just price, which I think is an important piece, And as we are, you know, getting into, you know, towards the automated production, we have always said that economies of scale will enable us to go after, you know, higher volume markets and lower level ADAS is part of that. So I think from a price point structure standpoint, it's definitely made possible by our investments we've made on the core technology, our core vision, as well as AdVest and Ultra Lines. But I think, you know, we see that both, you know, both on commercial vehicles and passenger. And I think at least, you know, between the programs that we have, I expect one of those to make the decisions this year coming up. So I think those are, would be for, you know, basically think of it as more advanced automatic emergency braking scenarios where cameras suffer like nighttime, pedestrians, automatic braking. and typically those are higher volumes in the 100,000 plus type run rate that we talked about.

And we've reached the end of our Q&A session for today's event. Thank you. This brings us to the end of today's meeting. We appreciate your time and participation. You may now disconnect.