CollPlant Biotechnologies Ltd Ordinary Shares Status update

Webinar will begin with Eran Rotem, Deputy CEO and Chief Financial Officer, who will provide a brief overview of the transaction. Then Dr. Ruti Ben-Shlomi, LightSolver Co-Founder and CEO, will take you through the technology and the business. We will then address questions that were submitted in advance. If you would like to submit a question during today's webinar, please use the chat button in Zoom. Before we begin, please note the disclaimer on this slide. Today's remarks include forward-looking statements, including statements about LightSolver's technology and product roadmap, commercialization plans, customer activity, market opportunities, expected performance, and future financial objectives. These statements reflect current expectations and assumptions and are subject to risks and uncertainties that could cause actual results to differ materially. Please refer to CollPlant's SEC filings for a discussion of these risks and for additional information regarding the transaction.

We do not undertake to update forward-looking statements, except as required by law. Now, I will hand over to Eran. Eran, please begin. Good morning, everyone.

The recent acquisition of LightSolver makes a pivotal moment for CollPlant, representing our strategic entry into the high-performance computing, HPC sector, a significant new growth engine for the company. Our rationale is centered on three key pillars. First, the opportunity. We are entering a massive market projected to reach nearly $90 billion by 2034, with applications across critical high-value industries. Second, the technology. This is not a concept. LightSolver has a proven working system that have already achieved real-world validation with a global leader like Boeing. Third, the path to scale. This is where CollPlant creates decisive value. We provide a robust public company platform, the access to capital, and experienced leadership to accelerate LightSolver's journey to commercialization. For our shareholders, this creates a direct path to participate in what we believe will become a disruptive new computing platform.

Crucially, we also structured the deal itself to ensure that this participation is tied directly to execution and value creation. This leads me to the transaction structure, which was designed to be as strategic as the acquisition itself. Our core principle was to align all parties toward one goal, creating tangible value through execution. The deal is performance-driven. While the upfront consideration was equity-based, the most significant component of the potential ownership is tied directly to the performance through milestone-based warrants. This structure is highly beneficial for our existing shareholders. It provides immediate participation in LightSolver's potential while ensuring that future dilution is linked to the achievement of significant technology and commercial milestones. In short, as LightSolver proves its value, it earns its equity. To power this next stage, CollPlant has also invested $5 million into LightSolver to fuel its operational and development needs.

Finally, as we advance this new strategic direction, we continue to actively evaluate alternatives for our biotech assets to maximize value for our shareholders from all parts of the business. With that, I will turn it over to Ruti, who will take you deeper into the LightSolver technology itself.

Ruti? Thank you, Eran, and thanks everyone for joining us today.

I'm Ruti Ben-Shlomi, Co-Founder and CEO of LightSolver. My background is in quantum physics. For nearly 20 years, I worked with cold atoms and trapped ion systems, two of the leading physical platforms being developed for quantum computing today. I also spent a few years at Intel as a process engineer, working at the cutting edge of semiconductor manufacturing and helping build advanced silicon transistors. I had the opportunity to experience two very different frontiers of computing, silicon and quantum. I saw their exceptional capabilities, but also their fundamental challenges. Computing is increasingly becoming an energy problem. Billions are being invested in better chip, quantum computers, nuclear power, even data centers in space. Today, the world is investing enormous resources in one question. How do we get dramatically more computing power without consuming dramatically more energy?

Our question was, what if we could compute with light itself? Together with Chen, Chief Scientist and Co-Founder, we invented a fundamentally different computing paradigm in which the core computation happens entirely in the all-optical compute loop. Instead of using billions of electronic transistors, it uses the physics of interacting lasers to perform computation directly with light. Because these lasers interact simultaneously, the system can explore real physical effect. That is the basic idea behind LightSolver, and that's how our journey began. I'm happy to introduce you the LPU, Laser Processing Unit. The LPU gives us massive parallelism, up to 1 million variables per core with optical dynamics operating on a few nanosecond timescales. Because the computational state remains inside the optical system, we minimize the constant movement of data between compute and memory with an equivalent internal bandwidth of approximately 200 terabytes per second.

The result is the potential to compute targeted workload up to 10,000 times faster while operating at room temperature in a rack-scale system and consuming less than 1 kilowatt. In simple terms, we use light to achieve more computation, faster, with dramatically less energy and infrastructure. Now let's switch gears and let me show you how it works. Let's start with an example of a computational problem, which is called image generation. We know how to translate the mathematical problem into a set of parameters that our optical system can work with. These parameters are then embedded into the laser array. In fact, we are programming the physical system to represent the problem we want it to solve. The computation is performed directly by the physics of the interacting lasers.

We start with a small amount of random photons that run back and forth inside a cavity, which means essentially two mirrors. The photons are being amplified while simultaneously interacting with one another, and the system evolves until it reaches a stable state, what we call the steady state solution. Once the system reached that steady state, the camera takes a snapshot of the laser value. Those values are then translated back into a digital solution that can be used by the customer's existing software environment. The loop is very simple. We translate the problem into light. We let the physics perform the computation and translate the result back into digital form. From the user's perspective, you don't need to know anything about lasers. You work through the software interface, and we handle the translation into and out of the optical system.

This is what excited us from the beginning. Rather than calculating every step, we set up the physical condition and let the system settle naturally into the answer. We started with a scientific breakthrough back in 2022. We demonstrated for the first time that interacting laser could solve your computational problem using only the physics of light. One year later, in 2023, we showed that the approach could scale to larger and more complex problem. Since then, we have moved from those early scientific demonstration to a working prototype in the lab, and from there, toward the next generation of the LPU. The progression is important. This is no longer just a physics experiment. We are taking a scientific breakthrough and turning it into a product that can ultimately be deployed commercially. Where are we focusing first?

Today, we are focused on three broad areas: optimization and large scale simulation. We are also exploring capabilities in AI and physical AI. These are the areas where we currently see the strongest combination of technical fit, customer need, and commercial opportunity. Very different real-world problems can ultimately reduce to similar types of computation that we can map onto the LPU. In optimization, that can mean finding the best solution among a very large number of possibilities and constraint in areas such as finance, manufacturing, and logistics. In simulation, the range of application is very broad, designing an aircraft, forecasting the weather, modeling heat flow in a semiconductor. Physical AI is a particularly exciting area for us. As robots, autonomous vehicles, and drones became more capable, they need to model, predict, and respond the physical world around them, creating a very demanding set of compute problems.

These workloads address a very large market. The overall HPC market is projected to reach approximately $89 billion by 2034. Within that, even if you take only differential equation workloads, which are the equation used to model physical system, it represent more than 60% of the HPC workloads. This translates into serviceable market of approximately $53 billion by 2034. Importantly, these numbers don't even include AI. AI and physical AI represent an additional opportunity on top of what you see here. We have already run POCs across many of the industries we just discussed, working on real customers' problems where conventional computing has become a bottleneck. What you see here are just a few examples. Across these projects, we've demonstrated significant improvement in speed, accuracy, and solution quality across very different types of workloads. Now we are starting to run that technical validation into paid customer programs.

Several customers are already moving from POC into paid pilots and NREs, and we expect more to follow. This is an important step for us. These paid programs allow us to go deeper with our customers, work on real-world application, and ultimately move those relationship towards broader commercial deployments. We cannot share the names of most of our customers we work with, but Boeing is one we can talk about publicly. Boeing asked us to work on corrosion because it is a major economic and operational problem in aerospace and defense. Corrosion-related costs run into tens of billions of ILS, and industry studies suggest that a meaningful share on the order of tens of percent can potentially be avoided through better corrosion management. It affects maintenance, aircraft availability, material selection, structural weight, fuel consumption, and ultimately, CO2 emission. The challenge is that high-fidelity corrosion simulation is extremely computationally intensive.