WSW, NY, October 20th, 2026, FinanceWire
Photonic computing is emerging as a potential answer to AI's insatiable demand for compute, and one Israeli company is pursuing it as part of Nasdaq-listed CollPlant.
In 2026, governments around the world announced very large funding commitments for a very specific bet: that the next generation of computers will not be built on silicon alone. In May, the U.S. Department of Commerce announced letters of intent for $2.013 billion in proposed CHIPS Act incentives for nine companies, including planned funding of $1 billion to IBM and $375 million to GlobalFoundries to stand up domestic quantum foundries. In March, the U.K. committed up to £2 billion to quantum innovation, including procurement of large-scale quantum computers for deployment by the early 2030s. In August, Canada signed a CAD $195 million agreement with Xanadu to build photonic quantum manufacturing. And in October, the Pentagon announced roughly $350 million for defense quantum computing.
AI Is Challenging Computing Infrastructure
Those commitments span national security, research and industrial competitiveness. Alongside that quantum push, artificial intelligence is showing an appetite for computation that may grow faster than the hardware that feeds it. Today's GPUs remain the workhorses that train and run AI models, but pushing conventional chips harder can bring greater heat and power demands, while feeding them with power, memory and data is becoming a larger challenge. The International Energy Agency estimated in 2025 that a typical AI-focused data center consumes as much electricity as 100,000 households. The pressure to keep adding capacity can collide with what power grids can support. At the same time, for memory-bound AI workloads, shuttling data between processor and memory can itself become a bottleneck. That pressure creates a strong incentive to find new ways to execute demanding workloads.
Governments are backing new computing architectures, and quantum developers are pursuing major technical milestones. Superconducting quantum systems such as IBM's require near-absolute-zero temperatures. Alongside GPUs and quantum machines sits a third idea: compute with light.
Light Does the Math
LightSolver's photonic approach uses photons, the particles of light, to perform computation. Instead of forcing a problem through millions of electronic switches, LightSolver's architecture maps the problem onto lasers and lets the physics of light do the work. The lasers interact inside a closed optical loop, settle into a physical equilibrium, and the answer is read off the light itself. Because light moves fast and many laser states can interact at once, the architecture is designed for massively parallel computation with lower power consumption on targeted workloads. The company describes a room-temperature, rack-sized platform designed for integration alongside conventional processors.
That is the idea behind LightSolver, an Israeli deep-tech subsidiary of CollPlant Biotechnologies Ltd. (NASDAQ: CLGN) whose Laser Processing Unit, or LPU, is built around interacting lasers. It works by translating a problem into laser parameters, letting the light settle into a physical steady state, and reading the answer back out in a single snapshot. The company describes the LPU as delivering the equivalent of 1,000 TOPS with 200 terabytes per second of memory bandwidth at about 100 watts, for complex analog computations on sparse, structured matrices. Its investor presentation describes up to 10,000-fold GPU acceleration on certain workloads, based on measured results and projections from current systems, emulation and scaling models.
Real Customers Are Testing It
The technology is progressing through customer development and validation programs. Boeing has moved LightSolver from a proof of concept to a multi-year engineering program, for which its September investor presentation reports projected corrosion-modeling acceleration of up to 100 times, based on emulation and proof-of-concept studies. A joint project with HLRS, a European supercomputing center, published research in the ACM Digital Library projecting approximately 40-fold to more than 80,000-fold acceleration for selected sparse linear systems, using an emulator of the LPU architecture. LightSolver says it is also working with a top-tier global financial institution, defense companies, and oil and gas and energy firms.
After announcing the acquisition agreement in August, CollPlant announced its completion on September 3, making LightSolver its subsidiary. The transaction used equity and milestone-based warrants. CollPlant continues to operate in regenerative medicine alongside its photonic computing business. LightSolver targets commercial availability of the LPU in 2028. CollPlant's September investor presentation cites Hyperion Research's 2024 projection of an approximately $89 billion high-performance computing market by 2034.
What the 2026 spending wave signals is that the next computing platform is now a matter of national priority, not just a research curiosity. LightSolver's September roadmap targets 100,000-variable proof-of-concept and MVP systems in 2027, a standard one-million-variable LPU in 2028, and enterprise clustering beyond, with timelines and specifications subject to change. Whether light becomes the workhorse of the AI era is still an open question, but the money and the customers are now moving in that direction. It is a company worth watching.
Recent News Highlights from CollPlant (NASDAQ: CLGN)
CollPlant to Host LightSolver Investor Webinar on September 17, 2026
CollPlant Completes Acquisition of LightSolver, Entering the Photonic Computing Sector
CollPlant is Acquiring LightSolver, Enters Supercomputing and Photonic Computing Markets
Important Disclaimers and Disclosures: The author, Wall Street Wire, is a content and media technology platform that connects the market with under-the-radar companies. The platform operates a network of industry-focused media channels spanning finance, biopharma, cyber, AI, and additional sectors, delivering insights on both broader market developments and emerging or overlooked companies. Wall Street Wire is not a broker-dealer or investment adviser. References to market size estimates, valuations, price targets, or other third-party data are provided strictly for informational purposes. Wall Street Wire receives cash compensation from CollPlant Biotechnologies Ltd. (the "Issuer") for coverage and awareness services, which are provided on an ongoing subscription basis. The content above is a form of paid advertising and promotion and is for informational purposes only and does not constitute financial or investment advice. This article may contain forward-looking statements about the Issuer's products, plans, or prospects that are subject to risks and uncertainties; actual results may differ materially, and readers should review the Issuer's public filings on SEC EDGAR (sec.gov/edgar) for full risk factors. Market size figures, research estimates, or other third-party data referenced in this article are quoted from publicly available sources believed to be reliable; however, we do not independently verify or endorse them, and additional figures or estimates may exist. Full compensation details, information about the operator of Wall Street Wire, and the complete set of disclaimers and disclosures applicable to this content are available at: wallstwire.ai/disclosures. This article should not be considered an official communication of the Issuer.
ContactWall Street Wire
[email protected]