So, using the handy-dandy AI in my new Rabbit r1, I recorded and transcribed today's AGM. It's garbage, but that seems fitting. I'll include it below.

https://x.com/InThisThinRain/status/1804369360636489745

So long and thanks for all the fish

As we move into the informal part of the meeting, we are pleased to have doctor Suresh Venkateshan present the future of poet. The theme of his address is pure play AI Datacom hardware. His address will be followed by a question and answer session. If you have a question or comment, please submit it through the Loomi meeting platform by clicking on the messaging icon. I and he has a mechanism for had everything to The businesses we're pursuing our future opportunities, and what makes us tick. In 2024, Poet has been able to add approximately $29,000,000 to its balance sheet. Significantly improving its financial stability inclusive of private placements, ATM shell, where shares and win conversions. It will port the invigorate your map. The instances map, one to the limits. Running to product test and explicitly dedicated to the growing demands placed by AI on hardware infrastructure While we've made meaningful progress on our financial measures, What we're most pleased about is the continued product development and customer engagement across our business. First, But first is let's discuss the markets we serve. Generative AI may be the largest technology transformation since the cloud. Which itself is still in the early stages. And perhaps since the Internet, AI is what we at Ford helped the market driver for photonics 3 point o A wave of photonics growth propelled by the insatiable data requirements of generative AI. Generative AI is typically an application built on what are called foundational models. And it is in the creation of these foundational models where the hardware pinches Foundational models require compute, that are needed. Transfer data gels. Are partnering for Take a and the to think that that bill made is the data communications fabric that is required to move vast quantities of data with low latency across the compute servers. This is the market that POT is at play. As we go through 2024 and into 20 25, business catalyst. Propelling our vectors of growth. From high speed 800 gig You can see in this mock up of a typical NVIDIA DGX box, A DGX box is an AI server. Which has in it multiple GPUs, memories, and other compute elements. And each DGX box represents a vast number of optical OSFP transceivers. And these are used for server to server transfer. Within an AI cluster. All the other links inside of the DGS box are still copper today. But soon, these will also be converted into optical links, and there are numerous companies vying to penetrate space with distinct solutions. Both target solutions both within the AI server and for server to server connectivity. The primary solution for optical connectivity for AI clusters is pluggable transceivers. This is like data center networks except that the data rate and the unit volume requirements for AI clusters is much higher. PoEt optical engines are highly integrated solutions that address higher speeds of 1.60 and 3.2 t soon, It's product line that addresses this market and offers a cost effective and highly scalable solution We are a participate in more remarkable opportunity Not only are they vast and constantly expanding, but they also manifest unique attributes where we tackle complex challenges requiring feature rich solutions that preclude any chance of commoditization giving rise to high barriers of entry that keep potential competitors at bay. Soon, You'll learn about a notable AI industry award that we won highlighting Poets' architectural advantages in AI, and providing more third party validation and exposure. With that bot backdrop, what we're doing now sets the foundation for rap growth in 20 events. With renewal to 2020 2017. Any successful business requires execution around 3 vectors. Key customers, key partners, and key products and development. We announced our engagement with Luxshare last year One of the premier module suppliers in the world, Both these engagements are around the 800 gig transceiver products, with an extension to 1.60 as the market prepares for that transition. Companies like Foxconn and Luxshare are rushing to address the gaps in supply chain, for opt optical transceivers for the AI market. With Foxconn, we're collaborating on developing 800 gig and 1.60 optical transceiver modules that address the demands of the AI industry Both Qualcomm and Luxor already established supplies in the higher hyperscale data center market and now they are gearing up to address the AI demand. Particularly with Foxconn, we're working in parallel on both conventional transceivers and solutions for linear pluggable optics called LPO. Which issues the DSP and saves significant tons of profit from We believe LPO will be the future of optical transmission once the kinks are worked out. Ties to the hyperscalers. Multilane has a desire to expand their product lines to encompass term cooled optical have established an interproduction capability for modules. Multilane's choice to work with Poet was influenced by the low CapEx costs for module manufacturing when using our highly integrated optical engines. Multilane has some unique capability that complement our technology and products. We're collaborating with them to design pluggable optical transceivers using our optical engines That will help our customers to accelerate their design cycles even further. The objective is to design cost optimized 800 gig modules that all offer superior power and performance. This collaboration helps reduce ultimate outlay by nearly 5000000 annual. Engines with DML lasers. Several providers necessitated a change Interposer architecture and process and transition to developing EML based solutions in q 4 of last year. We were able to generate first time right designs result to be success for any of that day for our civilization and we'll see for 2020 Yeah. The OLED display to now produce 400 gig and 800 gig transmit solutions based on EMLs has been truly transformational in terms of customer interest. EML lasers comprise 80 plus percent of the optical transceiver usage worldwide. And we've allotted about the world's 1 and only integration platform for EML lasers. Okay. To 100 100 gigs 400 gig. 400 gig. And 100 gigaSV chiplets. One of the many advantages of the Poet platform is its speed of development. If the building blocks are done well, we can generate rapid transitions in its products. We've also expanded our supply of EML lasers to include Mitsubishi Electric, and Yonjei in addition to Lamentum. With regards to 1.16, 2 things need to happen for AI to deploy it. 1st, 200 gig per lane on the electrical side has to be available for deployment. Large players like IBM VIA and Broadcom have already announced products. With 102 gig per lane and the expectation is that it will be available for deployment late in 2025 and into 2026. The second thing is the availability of 1 since this opt is T and G. Let us call point customers and bring 1.60 transceivers to the market in 2025. The first time since I've been at Putt we talk to customers about 1.60, and they tell us we're early This has never happened before. We've always been playing catch up. So our transition over the past 6 months to the EML platform and the significant leapfrog in capability is really positioning us at the forefront of optical transceiver adoption. But we're not stopping there. OSFPXD is an industry standard that is already defined It can carry 16 lanes of 200 gig, which gets us to 3.2 t. The challenge is to fit all the high speed optical and electrical components inside this package. Poet already has a solution today for 16 lanes of 25 gig that fits inside a q QSFPTT. That we're building for ADVA. That's a package that's even smaller than an OSFPXT. One of the benefits for our AirPodsert technology is its scalability. We can scale as hard lines from 16 lanes of 25 gig to 6 lanes of 200 gig. Having a roll There are of course, that come with higher speed, which our engineers are working on now. We're also collaborating with some industry leaders to come up with innovative solutions for 3.2t with a goal to showcase our solution at the OFC next year. Likewise, we're transcending our remote light sources The one that provides immense technical superiority stickiness in the market and with customers, doesn't only come with cost. There needs to be a plus one feature that results in stickiness and that is what's striving to do. To capture a disproportionate share of the remote light source market as that market develops and matures. We're developing brand new architectures and concepts around the interposer that can build multiple lanes of lasers on the interposer with frequency spacings down to 50 gigahertz. In an uncooled application. Which is unheard of, So these are things that we're working on now and and more to follow. We will begin sampling our 100 gig per lane 800 gig transmitter solution in July, to accompany our already developed receiver solutions to multiple customers. With 3 committed to module design across multiple form factors in addition to our multilay collaboration, In parallel, we continue also to debit order. 1st generation of the remote light source product, the Starlight. By the end of this year. We also need to constantly look at our business model as the demands on our product are evolving with successive generations. While the demand for optical engines versus modules was muted at lower speeds, The transition to EML architectures has completely transformed this demand pick We now see significant demand coming up for our optical engines. At 800 gig in various form factors, d r two zero, f r a, f r 4, l and its supplies. Light. So we're we are now adopting a dual model. Optical engines to large suppliers and optical modules and niche applications. This prevent direct conflict with customer orders. And also enables Poet to proliferate its solutions outside the mainstream. Given the shift in demand, especially with the likes of Foxconn, Luxshare, and the 3rd to be named supplier, We are working to consolidate our subsidiaries in China to allow more operational control and to be able to recognize revenue sooner. And then there is the issue of coping with the China West split, especially as it relates to AI and high speed transmission. While no one explicitly says no China, it is implicit. Inolite, Eoptilink, and several module companies in China have established operations in Thailand over the past year. Customers are dictating component choices based on geography. While it's not clear where this is headed, it would be prudent to project that we will require optical engine manufacturing outside super photonics and outside China. We're working on this issue with renewed urgency. We believe in the not to this future since we need to adopt a bifurcated manufacturing and sales strategy. Within China, we will be setting up a sales and marketing company with external investment. Dedicated to winning and growing in China. Leveraging Chinese manufacturing and Chinese components, Whereas outside China, we will use our internal sales and marketing engine establish manufacturing potentially in the Singapore Malaysia corridor. In both cases, our collaboration with multilane, provides us a source of optical module manufacturing. In closing, I'm grateful to our collective teams who have delivered on behalf of our shareholders and customers. These results represent a lot of invention collaboration, discipline, execution, and reimagination across both. There's nothing so limiting as viewing opportunities through the same lens as yesterday's challenges. We at Port strive to innovate with out of the box solutions to challenges versus incremental improvements which do not sustain in the long term. We will continue to focus on hiring and retaining versatile and Thank you for your I'm k. So I think moving to the q and a now. And Suresh, the first, question is when do you expect to be able to announce details of orders from celestial AI and will ship to celestial in 2024? I mentioned 4. Should be We will press the please at the end of 2024. Okay. And next, this question is actually for Tom Micah. What percentage of shares are owned in the USA Canada, and elsewhere? Say Europe. It's, roughly about a 40 60 split in terms of it was actual shareholding. The number of the, for shares would be, upper and care hold to make be there. Shareholding is actually less per per holder What is interesting is is that we get a lot more trading volume on the Nasdaq than we do on the Venture Exchange in recent Yes. Okay. And back to Suresh. For this one, Suresh, how has the relationship with Luxshare evolved? So you know, we're building a in wallet air, in LAPRACIST, all good, and, film with come, you know, increase in the church's ways and, shows were getting designed into multiple sockets, not just a single one. So that's also good. So I think owning a chipset which is transmit and receive, is something that we're now starting to do. And and help you, we'll we'll look maybe placing the 2 directors that have resigned. An important cover for it. Know, it's always important to have customers that already have significant engagements in significant sales. Spot on among among our entire customer base, rep the largest module maker, and therefore, they are a premier supplier. What would one degree of separation from the end users, primarily the hyperscalers. You know, I think, with especially with our EML platforms now, at think the customers we engage with have that presence, which then gives us more exposure directly to some really large companies. And so it's a very, very important project for us, and it's obviously critical execution play put set on the mentorship, you know, a credit flow that kits at Voxcom. Okay. Great. Thanks, Rash. That is the last question. If you have any final remarks, 1.60 this year, I think we'll be again one of a handful at 3.2 t next year. Right? So we we are showcasing the scalability of our platform and partnering with some of the most influential component manufacturers in the world. That's just gonna elevate our profile considerably over coming year. So I'm really excited about that, and I think you all should be as well. I think, you know, we we we look forward to the developments over the coming year. Good. Thank you, Suresh.