Peraso Inc. (PRSO) Q2 2026 Earnings Report, Transcript and Summary

Good afternoon, and welcome to Peraso Inc.'s Second Quarter 2026 Conference Call. [Operator Instructions] As a reminder, this conference call is being recorded today, Tuesday, August 11, 2026. I would now like to turn the call over to your host for today's conference call, Mr. Jim Sullivan. Please go ahead.

Good afternoon, and thank you for joining today's conference call to discuss Peraso's Second Quarter 2026 Financial Results. I'm Jim Sullivan, CFO of Peraso and joining me today is Ron Glibbery, our CEO. Today, after the market closed, we issued a press release and related Form 8-K, which was filed with the Securities and Exchange Commission. The press release and Form 8-K are available on Peraso's website at www.perasoinc.com under the Investor Relations section. There is also a slide presentation that we will be using in conjunction with today's call that may be accessed through the webcast link on the IR website. As a reminder, comments made during today's conference call may include forward-looking statements within the meaning of Section 27A of the Securities Act of 1933 as amended and Section 21E of the Securities Exchange Act of 1934 as amended. All statements other than statements of historical facts could be deemed as forward-looking. Peraso advises caution and reliance on forward-looking statements. These statements include, without limitation, any projections of revenue, margins, expenses, non-GAAP gross profit, non-GAAP gross margin, non-GAAP operating expenses, adjusted EBITDA, non-GAAP net loss, cash flows or other financial items, including anticipated cost savings as well as any statements concerning the expected development, performance and market share or competitive performance of our products or technologies. Any statements regarding the sufficiency of the company's capital resources and its ability to continue as a going concern, any statements regarding customer demand forecasts and concentration risk and any statements related to prospective future financing arrangements or capital transactions and the evaluation or pursuit of strategic alternatives. All forward-looking statements are based on information available to Peraso on the date hereof. These statements involve known and unknown risks, uncertainties and other factors that may cause Peraso's actual results to differ materially from those implied by the forward-looking statements, including unexpected changes in the company's business. More detailed information about these risk factors and additional risk factors are set forth in Peraso's public filings with the SEC. Peraso expressly disclaims any obligation to update or alter its forward-looking statements, whether as a result of new information, future events or otherwise, except as required by applicable law. Additionally, the company's press release and management statements during this conference call will include discussions of certain measures and financial information in terms of GAAP and non-GAAP. With respect to remarks on today's call involving non-GAAP numbers, unless otherwise indicated, referenced amounts exclude stock-based compensation expense and the change in fair value of warrant liabilities. These non-GAAP financial measures, definitions and the reconciliation of the differences between them and comparable GAAP measures are presented in our press release and related Form 8-K, which provide additional details. For those of you unable to listen to the entire call at this time, a recording will be available on the Investor Relations page of our website. I'll now turn the call over to our CEO, Ron Glibbery, for his prepared remarks. Ron?

Thank you, Jim. Good afternoon, and welcome to everyone on the call and webcast. We appreciate you taking the time to join today's call. As reported in our press release this afternoon, revenue for the second quarter increased approximately 36% sequentially and was above our previous expectations. The sequential growth primarily reflected our successful fulfillment of a previously delayed order shipment to a new fixed wireless access customer. Despite the better-than-anticipated top line results this quarter, our overall business continues to be impacted by a combination of macro-related headwinds and irregular order patterns across our existing customer base with limited ability to influence certain underlying market dynamics, particularly those affecting our leading fixed wireless access customers, we remain focused on the elements of the business we can control. These include operational efficiencies, ongoing cost management and enhancing the resilience of our supply chain, whilst continuing to cultivate extensive customer engagement across our target end markets. Turning to Slide 4. Across each of our target end markets, we continue to be encouraged by the growing recognition of 60 gigahertz mmWave technology and its inherent performance benefits over conventional wireless communication solutions that operate in lower frequency bands. Historically, the most prevalent commercial application for 60 gigahertz technology has been fixed wireless access. While this continues to be the largest and most mature end market for Peraso, the rapid proliferation of AI and associated demand for hardware in support of AI infrastructure has led to significant market dislocations. With respect to fixed wireless access, this includes customers navigating persistently elevated memory prices and the constrained supply of other key components. We believe this has and is continuing to contribute to subdued purchase order activity and limited near-term visibility into future customer demand. While these current market dynamics are expected to extend into the fourth quarter, the value proposition of 60 gigahertz mmWave technology and fixed wireless application is its commercially proven ability to overcome interference and capacity limitations in congested wireless environments. As such, we believe we are positioned to benefit once the current supply dislocations moderate and order patterns normalize. Having said that, it's become increasingly clear over the past 12 to 18 months that the inherent benefits of 60 gigahertz mmWave technology extend beyond traditional fixed wireless access applications. We continue to engage with prospective customers and partners seeking to utilize our 60 gigahertz solutions to address real-world challenges encountered in other end markets, namely in autonomous Edge AI, unmanned aerial vehicles, and defense communications applications. Today, we are actively engaged on a combination of new and existing opportunities to leverage Peraso's proven 60 gigahertz technology in both of these high-value end markets. Turning to Slide 5. I want to start with the substantial emerging market opportunity that we see for our technology in emerging edge AI and associated autonomous systems application, whether it's operating a fleet of autonomous vehicles, a swarm of drones or a factory floor full of robots, implementing Edge AI at scale creates a fundamental bandwidth challenge. These platforms typically require the continuous upload of massive volumes of sensor and camera data as well as the download of large AI models and software updates. This often entails not only multi-gigabit data transfers, but doing so while operating in close physical proximity to many other independent autonomous devices. Traditional 5 gigahertz WiFi technology struggles in these types of environments due to limited spectrum. Additionally, omnidirectional coverage creates widespread co-channel interference or overlapping of wireless signals. As the number of devices sharing the same channel increases, the throughput collapses and latency becomes unpredictable. With 60 gigahertz technology, we utilize highly directional narrow beams. That directional nature significantly reduces co-channel interference, allowing numerous simultaneous multi-gigabit links to operate reliably inside the same physical footprint with deterministic low latency performance. This isn't just conceptual. In June of this year, we completed a customer demonstration with our partner, Virewirx, that put these advantages on full display. Together, we created a simulated drone wireless environment and successfully operated a high-capacity network cluster that achieved a peak aggregate data rate of more than 50 gigabits per second. Within the combined space of a large room, we simultaneously transmitted several independent multi-gigabit streams, something conventional sub-6 gigahertz technologies struggled to do at that density without severe interference. The demonstration also underscored 2 additional attributes that are critical for autonomous and mission-critical systems, the inherent low probability of intercept and detection of 60 gigahertz links and the ability when combined with Virewirx's network synchronization technology to maintain coordination without reliance on GPS. That GPS independence is particularly valuable in contested or GPS-denied environments. Looking ahead, we see applications in both defense, especially emerging drone swarm tactics and commercial settings such as automated factories where large fleets of robots require high-capacity interference-free connectivity. The Virewirx collaboration is a strong example of how complementary 60 gigahertz technology can address real-world applications that traditional wireless architectures were not designed to solve. As we've introduced on previous conference calls, many of these fundamental characteristics of 60 gigahertz are equally, if not more, critical in defense and tactical communications applications. Turning to Slide 6, I want to highlight one of the most unique applications of our technology within the tactical communications space, IFF, which stands for Identification Friend or Foe system. It goes without saying today's increasingly modern battlefield is a complex and congested environment. The ability to rapidly and reliably distinguish friendly assets from hostile ones is mission-critical. Traditional IFF systems operate by transmitting what's known as an interrogation signal and receiving authenticated responses. These signals or links can be detected, jammed or spoofed, exposing troops and assets to devastating consequences, including friendly fire incidents. 60 gigahertz technology is designed to meaningfully upgrade the performance of an IFF system in 3 ways. First, the highly directional narrow beam makes the communication links inherently difficult for an adversary to detect or intercept. Second, that same directional beam combined with the physical properties of the 60 gigahertz band substantially reduces the risk of spoofing or successful jamming. Third, we have integrated our solutions to deliver these capabilities in a compact low-power form factor. This makes it ideal for size, weight and power-constrained platforms such as unmanned aerial systems, or UAS. In partnership with our previously announced lead defense contractor, InTACT, we are continuing to make solid progress on the refinement and testing of next-generation drone IFF solutions. This system is purpose-built for highly contested electronic warfare environments and is designed to enable secure real-time mutual authentication between friendly drones and ground forces. Following our delivery of initial limited production modules earlier this year, feedback on this jointly developed solution continues to be positive, which we believe is meaningful validation of Peraso's 60 gigahertz technology for IFF applications. More broadly, the combination of stealth-oriented design, anti-jam resistant architecture and compact integration is generating expanded interest from additional prospective customers and partners seeking to evaluate the benefits of 60 gigahertz for their future product road maps. With that foundation, let me turn to the wider opportunity we see in unmanned aerial systems communications. Moving to Slide 7. Building on the initial foundation from our ongoing work on IFF applications, we have identified and are beginning to target broader opportunities for our 60 gigahertz solutions in unmanned aerial systems. In fact, the rapid growth of UAVs and related systems, especially in defense and security applications, has helped to increase the awareness and the limitations of traditional sub-6 gigahertz radio links. These legacy systems are increasingly susceptible to wireless congestion, relatively easy to detect and vulnerable to jamming in contested electromagnetic environments. Additionally, modern drones have become highly data-intensive with large arrays of sensors. They are also frequently required to support real-time coordination, autonomy and swarm operations. Today's existing drone communications infrastructure is challenged to meet all of these demands. 60 gigahertz mmWave technology is designed to enable a fundamentally different solution for modern UAS platforms. Instead of broadcasting energy broadly, it utilizes highly directional, high-capacity data links. Atmospheric oxygen absorption further limits long-distance propagation, reducing the probability of detection or interception outside the intended path. Adaptive beamforming is the foundational capability that makes this possible, enabling precise energy steering and spatial filtering of interference and high spatial reuse of multiple independent links can operate concurrently without overlapping interference. As seen on the slide, the narrow point-to-point paths of 60 gigahertz beams are virtually impossible to detect and extremely difficult to jam. The highly directional beam makes alignment by a jammer difficult, while oxygen attenuation further weakens potential interference signals. Since by definition, the band is unlicensed, 60 gigahertz also avoids interference with licensed spectrum. Bringing all of these elements together, this week, we expanded our product portfolio with the formal introduction of Peraso's PRM2145. This new jam-resistant communications module is purpose-built for unmanned aerial vehicles, autonomous systems and defense applications. The PRM2145 leverages our proven 60 gigahertz platform to deliver secure, ultra-low latency gigabit class performance in a lightweight compact form factor optimized for airborne platforms. It is designed to support robust command and control links, high-definition video, ISR data transport, Identification Friend or Foe communications and drone-to-drone networking. This new module is expected to be available for customer evaluation in the fourth quarter, supporting integration across a broad range of unmanned and autonomous platforms. We believe the PRM2145 further strengthens our early entry position in these high value-add markets while also reinforcing our strategy of diversifying Peraso's future growth beyond our core fixed wireless access business. Turning to Slide 8. Before my concluding remarks, I want to briefly highlight one other real-world application and opportunity that we are actively engaged in development of with a prospective customer. Today, the current landscape within defense tactical communications is rapidly driving the evolution of next-generation solutions. One real-time example of this is hybrid UAV networks. In these architectures, 60 gigahertz communications links are combined with other transport layers such as satellite backhaul or selective fiber to deliver high-capacity, low-latency connectivity while still maintaining the intended stealth and anti-jamming benefits of mmWave technology. As visualized on this slide, this hybrid approach allows operators to optimize a complete end-to-end network for range, resilience and mission flexibility in contested or GPS-denied environments. I wanted to briefly share this example of a hybrid UAV network on today's call for 2 reasons. First, it reflects the type of advanced system design work that we are actively collaborating with partners to solve and work toward bringing to market. And second, I believe it further illustrates the expanding role 60 gigahertz can play beyond traditional point-to-point or fixed wireless use cases. In closing, our near-term focus remains firmly on the areas within the company's control. We continue to take steps to strengthen the resilience of our supply chain and reduce single supplier bottlenecks while also reducing to secure additional distribution partners that will expand the market reach of our 60 gigahertz mmWave portfolio. Our highest priority is driving commercial traction, converting existing customer engagements, proof of concepts and technology demonstrations into design wins, purchase orders and production revenue that can support renewed top line growth. While near-term visibility, particularly within fixed wireless access remains limited due to external market dynamics and irregular order patterns, the breadth of our current engagements and the growing recognition of 60 gigahertz differentiated capabilities makes us optimistic for the future. With that, I'll turn the call over to Jim for a detailed review of the financial results for the quarter.

Thank you, Ron. Turning to the results for the quarter. Total net revenue for the second quarter was $1.3 million compared with $1 million for the prior quarter and $2.2 million for the second quarter of 2025. Product revenue in the second quarter was $1.2 million compared with $0.7 million in the prior quarter and $2.2 million in the second quarter of 2025. The sequential increase in product revenue for the second quarter of 2026 was attributable to higher shipments of mmWave products. Sales of our legacy memory and related products were less than $25,000 in each of the first 2 quarters of 2026. We recorded no sales of legacy memory and related products in the second quarter of 2025. Gross margin was 63.7% in the second quarter compared with 61.5% in the prior quarter and compared with 48.3% in the year ago quarter. The sequential increase for the second quarter of 2026 was primarily attributable to mmWave product mix, including increased sales of inventory written down in prior periods as partially offset by a decrease in nonrecurring engineering services or NRE revenue. The year-over-year increase was primarily attributable to an increase in NRE revenue and mmWave product mix, combined with an increase in sales of inventory written down in prior periods as partially offset by a decrease in sales of memory and related products. GAAP operating expenses for the second quarter of 2026 were $3.1 million, consistent with the prior quarter and compared with $2.9 million in the second quarter of 2025. Non-GAAP operating expenses, which excludes stock-based compensation, were $2.9 million in the second quarter, consistent with the prior quarter and compared with $2.7 million in the second quarter of 2025. Our recent non-GAAP operating expenses level of approximately $2.9 million per quarter continues to reflect the benefits realized from previously implemented cost reductions and our ongoing cost containment initiatives. GAAP net loss for the second quarter of 2026 was $2.2 million or a loss of $0.16 per share compared with a net loss of $2.5 million or a loss of $0.22 per share in the prior quarter and compared with a net loss of $1.8 million or a loss of $0.31 per share in the same quarter a year ago. Non-GAAP net loss, which excludes stock-based compensation and changes in fair value of warrant liabilities for the second quarter of 2026 was $2.1 million or a loss of $0.15 per share. This compared with a non-GAAP net loss of $2.3 million or a loss of $0.20 per share in the prior quarter and a net loss of $1.7 million or a loss of $0.28 per share in the same quarter a year ago. The weighted average number of basic and diluted shares outstanding for purposes of calculating both GAAP and non-GAAP EPS for the second quarter of 2026 was approximately 14.1 million shares. Adjusted EBITDA, which we define as GAAP net income or losses reported, excluding stock-based compensation, change in fair value of warrant liabilities, interest expense, depreciation and amortization and the provision for income taxes was negative $2 million in the second quarter of 2026 compared with negative $2.3 million in the prior quarter and negative $1.6 million in the second quarter of 2025. With regard to the balance sheet, as of June 30, 2026, the company had approximately $3.3 million of cash compared with $2.7 million as of March 31, 2026. The net increase of approximately $0.6 million in the company's cash balance at quarter end reflected approximately $2.4 million of net proceeds from sales under the company's at-the-market offering program during the second quarter of 2026. As of today's call, the company has approximately 15.1 million shares of common stock and exchangeable shares outstanding. As previously disclosed, the company has been exploring potential strategic alternatives, including a merger sale of assets or other similar transactions as well as various potential sources of additional capital. On June 30, 2026, the company entered into a committed equity facility with ROTH Principal Investments. This facility allows but does not obligate the company to issue and sell up to $25 million of shares of its common stock. In addition to providing access to additional working capital for general corporate purposes, the facility is expected to support continued product development targeting expanded opportunities in drone, defense and tactical communications markets. Independent of the newly entered equity facility, we don't have any updates to share today in the company's broader strategic review process. As Ron previously mentioned, overall visibility into future near-term demand remains challenging due to the irregular order patterns from our fixed wireless access customers. Due to insufficient visibility, we will not be providing quarterly guidance on today's call. This concludes our prepared remarks, and we thank you for your time this afternoon. Operator, please commence the Q&A session.

[Operator Instructions] Your first question is coming from Jon Hickman from Ladenburg.

This new communications strategy, so who did you say your partner was again that you're working with to design for the, like, the factory floor and UAVs and all that?

Are you referring to the final slide?

No, the second -- no, it's the third to the last one, showed all the robots and stuff.

Yes. Well, that -- again, it's obviously still confidential, Jon. But the concept there is -- I think what's really insightful about that situation is that -- and you're seeing this a lot in the U.S. as well as China, with humanoid robots, which basically have full freedom, the models are complicated. So it turns out the data flow between the network and these humanoid robots is substantial. That's where our chips come into play. And if you look at that, I love that slide because actually, that slide just shows you the benefit of our technology over traditional wireless technology, which is that we have a very narrow beam, very, very physically limited area of operation versus traditional wireless, which has a very wide field of operation, and that's why we win in those circumstances. So we really see that market as an important emerging market because of that capability of our technology. So I hope that slide got the message across. I think it did. I mean that's really the concept behind it.

You said you'd have chips available for people to experiment or see if they work and by the end of the year in the fourth quarter?

No, our chips -- no, no, what we meant to say was our chips are available now. The only concern about -- we -- overall, we have run into an issue. We've seen this with other companies, whereby the availability of third-party memory like DDR4, DDR5 is limiting sales and making orders kind of lumpy. So that's what I was referring to. As far as the silicon for that application, we are in full production and it's available today.

So what were you talking -- when you said you would have chips available for in the fourth quarter.

Yes, we have a new module product based on our existing silicon. And the new module, we did a press release yesterday, and that is called the PRM2145. And that is -- I think that's -- so that's a new module with some -- you'll have to go to our website to see the specific technical details. But basically, it's another part of our module family based on our existing silicon. Of course, the significant change to that particular product is, again, a new antenna, which we designed in-house ourselves. So we've got the ability to change the antenna to provide different RF characteristics.

Okay. And then you've been working with some partner, I think, in Israel on the friendly fire thing for a while now, more than a year. So are you generating revenues there?

Well, we definitely have generated revenue there, substantial revenue, 7-figure revenue from 2 perspectives. One perspective is from an engineering perspective, and we actually have shipped volume. Generally, I would say what we're seeing is originally, that business and still is based on what we call in country. But actually, we're starting to see a shift over to the drone application as well. So just you can imagine the need for friendly fire in circumstances where you've got a drone and infantry on the ground. And the drone wants to make 100% sure that it's not targeting ground personnel. So it's really an evolution of that model from kind of ground-based communications to aerial-based communications. So we're starting to see that transition. And we really do anticipate more volume from that customer over the next few quarters.

So this issue of drones being jammed or their communications interrupted to make them ineffective. So theoretically, every drone should have millimeter microwave (sic) [ mmWave ] on it so that it can't have...

Yes, theoretically. And frankly speaking, that -- again, that is a core value proposition. So from our perspective, we've always had this capability. What's changed over the last year, Jon, is the fact that traditional wireless technology that was used to control drones now is getting jammed by the enemy. So we frankly see our -- really our technology potentially going -- and we have to keep in mind, this is really more for military and security applications where the user wants to avoid being jammed by the enemy. Obviously, if you're just flying a drone at a wedding and you want to take some pictures, I mean, not really necessarily a requirement for anti-jamming. But frankly, for military and security applications, we're seeing very, very strong interest there, and you're going to see a lot of announcements from us over the next few weeks and months. And as a matter of fact, that was the announcement we made yesterday was a specific new module for the drone application. So we see it as a very, very important part of the drone rollout over the coming years.

So can you -- so it's kind of a defensive technology, you're preventing communications interruptions on the drone. Can you turn it around and make it an offensive like use it to more precisely interrupt other people's drones? Or is that not?

No. I wouldn't -- it's really more command and control. I mean really the 2 primary applications are the friendly fire, which I described, whereby the drone can determine whether the ground personnel or ground devices are friend or foe. But of course, the other application is command and control of the drone. So your definition of defensive or offensive is kind of up to you. But yes, we don't see -- we do have the ability eventually to also detect drones, but we're not doing that today. In other words, it's kind of a radar applications. But really, the focus for us right now with our existing technology is command and control for the drone application. And with the concept that we -- it's very -- I would say, very, very difficult to jam our link. And I hope our slides today really communicated that concept.

Well, how are you getting yourself noticed, like, the military...

Yes, it's a very good question. Like -- so basically, we -- really traditional, like, we do in the States, in America, we've got some terrific partners that are helping us in that regard, and that's still under confidentiality. We're actually -- to be frank, we've been accepted by a NATO consortium that's actually promoting technology in NATO-based countries. And that's actually been a very, very strong channel for us to get recognized by drone manufacturers. It's probably our primary channel right now for engagements. I would say, just off the top of my head, we are for the last 2 to 3 months, engaging with new drone customers, and I mean design wins on a weekly basis. So the exposure is very, very good through NATO, for example. We do attend trade shows and have traditional marketing channels, but the NATO alliance has been very helpful for us. So there's been a variety of methods that we're undertaking to really get ourselves. And obviously, as the word gets out that our technology can be used for jam-resistant communications, that really spreads the word as well, and that helps our communication. So we really do find that we're getting more and more inquiries on a weekly basis.

So just one more question there. So some of these drone manufacturers are making thousands a week or a month. If somebody wants that kind of volume, can you do that?

Yes. I mean today, I would say our capacity is about 30,000 to 50,000 units a month. But obviously, we've kind of achieved that in the past. I would say we would have no problem getting to 100,000 units a month. So we -- our primary silicon supplier is TSMC, which for them, that's kind of peanuts. And I think with our other suppliers, it's just a matter of time before we can get to those volumes. So we're not worried about achieving those volumes from an operational perspective.

Your next question is coming from Kevin Liu from K. Liu & Company.

Just a couple of follow-ons on the defense opportunities that you have. Can you talk about the pipeline? And more specifically, how soon do you think you can kind of get the second and third production customer? Is that a couple of quarters out? Is that longer term? And then for the hybrid UAV network that you referenced, I was curious if there are any specific LEO partners that you can name there?

For the first part of your question, Kevin, I mean, actually, it turns out there's been a very -- there's kind of a very interesting development, which is that obviously, we're targeting drone manufacturers, but we are seeing actually traction in a related business, which is actually like fixed wireless communications, but secure fixed wireless communications. So it turns out that some of the customers who are actually providing drone communications also would like us to provide communications on the ground. And we could see those shipments starting really in Q4. So that's very exciting. I think the drone specifically will be Q1, Q2 next year. But we -- I actually am hoping that we -- like the whole concept of secure communications is just a very, very important concept. And again, I mean, we tried to make it clear in the slides today, but our ability to avoid jamming is just very fundamental. It just basically boils down to very, very narrow beam and very directional, directional and dynamic, like all those things that add up to make the whole communications channel very, very difficult to detect and to jam. So I would say there's -- we could start shipping in lower volumes even later this year. And so that's really our target. But I'm thrilled to say that the momentum certainly on defense manufacturers overall is turning out to be very, very positive for us. Now the second part of your question is one of the -- again, without getting specific, one of the opportunities we're specifically working on, and I find this very, very interesting. In a certain battlefield environment, and I can't say where, there is a model whereby a LEO satellite provider can provide a link to a control drone, call it, and that control drone can manage, kind of, client drones. And so what that means is the drone operator can be really anywhere. And you can be anywhere in the globe and really see a real-time first-person video depiction of those client drones. And so you can do the math on how many LEO operators there are in the world. But like certainly, we're -- that's one of the applications we're working on is that specific scenario where the actual pilot of the drone is nowhere near the battlefield and using the LEO environment to actually control the client drone. So that's a really interesting application for us. So we're working on that kind of in real time. So that is probably as much as I could say at this time. But certainly, suffice it to say that, that's an application that people are very, very intrigued about.

Yes. Just one quick follow-on on that point is with all these new things that you guys are working on, how quickly can you actually get these things to the battlefield? Are these all products that are contemplated for kind of the current conflicts that are ongoing? Or are these more so kind of next-generation technologies where it could take a year or more for them to really get into production?

Kevin, I think it's going to be phased. And frankly, again, I mean, going back to my earlier comments to Jon, I mean, we -- I can see us starting to ship literally later this year on ground-based applications. So just to give you an example, like there's -- some of our customers are providing radar and they need communications, like, data links to their radar units and the radar units need to be secure. So you can see a situation where we could be shipping. There's very, very little development that has to take place because it's really like a point-to-point applications for fixed wireless access as it turned out. It just turns out it's secure. There may be a few more bells and whistles, but I can see us starting later this year. I think from -- look, I mean, if we're talking about a very, very sophisticated kind of swarm, I think swarm drones in particular, is a complex problem. We're probably a year away from that. But there's no doubt in my mind, we'll start shipping later this year or early Q1 and then just ramp it up from there.

I appreciate that. Maybe switching gears a little bit on to more of your traditional fixed wireless business. I certainly understand some of the challenges your customers are going through and why they're holding off on purchases. I'm curious what you're hearing from them in terms of where their inventory levels are with your chips. And when they do start to reorder again, are you expecting kind of a big ramp-up as they restock pretty fully? Or do you think it's kind of a more gradual uptick?

I'm going to share some insight with you, Kevin, that even over the last few days, which is that even though there's a lot of buzz on the military side of things, on the fixed wireless side, we've had 3 major customer opportunities arise very, very recently, including -- believe this was very, very good news for us is one of the largest OEMs in India now it feels like they're going to -- they'd like to start shipping our technology in India, which is obviously a tremendous market, Middle Eastern customer as well as a large OEM. So -- and the message from all of those potential customers is that they're seeing traction. I think with regard to our own customers, we -- I'm going to let Jim kind of answer in terms of specifics, but we actually are starting to see reorders from some customers. So we're starting to see that turnaround. We're hoping -- we know for sure in Q4, we're going to start to see that turnaround. In terms of what we can share globally, Jim can speak to that more specifically. But we really -- I think the good news and certainly what I've seen even over the last week is we're getting real momentum from new customers, and I'm talking very large opportunities in India and the Middle East as well as a new OEM partner. And we're hoping by Q4, we really overcome this memory situation. But by the way, one of the key points, I think, that our customers are trying to do on the memory side of things is kind of switching to DDR3. I mean, so one of the problems is the AI customers broadly are using DDR4, DDR5, which is very, very high speed. But luckily, in most cases, our customers can get away with DDR3. So we're starting to see that transition, which we are hoping by Q3 will alleviate some of this pressure.

Great. One last one, perhaps for Jim. Just it doesn't sound like it, but curious with the ROTH equity facility that you guys secured after quarter end. Did you guys tap into that at all? Or is that still fully available to you?

No, we have not tapped into the ROTH facility or used our ATM facility since I think the -- yes, since June 30, I think we had some trades under the ATM maybe settled on July 1. No, there's been no activity under either facility. We'll have comments, obviously, in the 10-Q regarding our runway and liquidity, et cetera, going forward. Obviously, we've got some market conditions working against us on the fixed wireless side and while a tremendous amount of activity on the drone defense side. And I kind of joke with folks. I know there's a lot of activity because a lot of NDAs are crossing my desk when Ron has come back from trade shows, which is always a good sign to me. I've had quite a number of them. So there's some good -- to your comment earlier about the pipeline. But we'll be looking at that going forward. And as has been the case and kind of my mantra in past quarters, first and foremost, if we can't get our products out the door as quickly as we'd like based on customer orders, we remain focused on nonrecurring engineering services, NRE transactions. I believe Ron and team are working on one right now in the defense space. And generally, those contracts have no incremental cost to us. So it's funded R&D to me that drops to the bottom line. So we've got at least one of those that's hopefully as high as mid-6 figures that we're working as well with a number of the programs out there on technology, Ron has been very active in pursuing grants from the Canadian government. Obviously, each country wants to have its independent technology and resources. And while Peraso is a U.S. public company, the majority of Ron and the majority of the team and all the engineers working on our mmWave technology sit in Canada. So we've been pretty aggressive about applying for those. Time will tell. We've done the data submissions on the administrative side, Ron and led by Alex, our CTO; and Brad, our COO and our VP of Biz Dev have done the technical details, provided the specs, et cetera. So we're optimistic about some of those coming in. Anything I missed there, Ron, on that front?

No, that was... Yes. No, perfect. Terrific, Jim.

I'll stop there. I don't know if that answered your question.

Definitely. Look forward to seeing some of these defense announcements in the near future.

I show that there are no further questions in the queue at this time. That will conclude today's conference call. Thank you for your participation. You may now disconnect.