It has taken several years for CPO to develop and as usual things likely speeded up as it became clear, faster and lower power processors were needed as AI was launched with the availability of CharGPT three years ago. TSMC has been sampling 200G MRRs using CPO technology for Nvidia and perhaps Broadcom so it is unlikely that LWLG polymers are being integrated for use in the first generation of the MRR processors. Nokia is strongly aligned with Nvidia so we should follow both as far as press releases. It makes sense to release a simpler version of a vastly new technology to make absolutely certain it meets or exceeds expectations. It doesn't mean that LWLG polymers aren't being used right now in building the next generation Rubin processors coming out in 2027 and it could be an updated version of 200G, a 400G version (even though LWLG has not officially announced that version is ready) or a different series of both to meet different levels of speed. I asked AI Overview about the possibilities and here was the answer,
AI Overview Based on current industry reports and technical roadmaps, TSMC is indeed finalizing Co-Packaged Optics (CPO) using Micro Ring Resonators (MRR) to achieve 200G+ per lane for next-generation AI processors, likely aiming for the Rubin/Vera Rubin generation (2026-2027). Regarding the integration of Lightwave Logic (LWLG) organic polymers into this specific design: LWLG Status in Current Design: While LWLG has demonstrated 200G/lane modulators and has strong potential for CPO due to low power and high speed, there is no official confirmation that LWLG’s material is designed into the initial TSMC/Nvidia Rubin 200G MRR architecture. LWLG operates on a co-development model with various partners, and while their polymers are relevant, they are not currently listed in the immediate, finalized supply chain of the first-wave Rubin CPO. Possibility of Integration: It is highly possible that LWLG material is being tested for this design, but a more realistic timeline for the adoption of organic polymer modulators in high-volume, mission-critical GPU products like Rubin would be as a follow-up or in the "next generation" of the platform. The Next Gen (Rubin Ultra): Industry reports indicate that Nvidia's roadmap includes the Vera Rubin Ultra NVL576 in 2H27, which requires further innovations in bandwidth and power. As CPO moves from its initial adoption (2025-2026) to more widespread, higher-bandwidth implementation, materials that offer better performance (like polymers) have a higher likelihood of adoption. Summary: LWLG polymers are a strong candidate for future CPO upgrades, but it is uncertain if they are in the initial design. They are more likely to appear in later, upgraded iterations of the Rubin/Rubin Ultra architecture rather than the absolute first-generation, finalized design.
It has taken several years for CPO to develop and as usual things likely speeded up as it became clear, faster and lower power processors were needed as AI was launched with the availability of CharGPT three years ago. TSMC has been sampling 200G MRRs using CPO technology for Nvidia and perhaps Broadcom so it is unlikely that LWLG polymers are being integrated for use in the first generation of the MRR processors. Nokia is strongly aligned with Nvidia so we should follow both as far as press releases. It makes sense to release a simpler version of a vastly new technology to make absolutely certain it meets or exceeds expectations. It doesn't mean that LWLG polymers aren't being used right now in building the next generation Rubin processors coming out in 2027 and it could be an updated version of 200G, a 400G version (even though LWLG has not officially announced that version is ready) or a different series of both to meet different levels of speed. I asked AI Overview about the possibilities and here was the answer,
AI Overview Based on current industry reports and technical roadmaps, TSMC is indeed finalizing Co-Packaged Optics (CPO) using Micro Ring Resonators (MRR) to achieve 200G+ per lane for next-generation AI processors, likely aiming for the Rubin/Vera Rubin generation (2026-2027). Regarding the integration of Lightwave Logic (LWLG) organic polymers into this specific design: LWLG Status in Current Design: While LWLG has demonstrated 200G/lane modulators and has strong potential for CPO due to low power and high speed, there is no official confirmation that LWLG’s material is designed into the initial TSMC/Nvidia Rubin 200G MRR architecture. LWLG operates on a co-development model with various partners, and while their polymers are relevant, they are not currently listed in the immediate, finalized supply chain of the first-wave Rubin CPO. Possibility of Integration: It is highly possible that LWLG material is being tested for this design, but a more realistic timeline for the adoption of organic polymer modulators in high-volume, mission-critical GPU products like Rubin would be as a follow-up or in the "next generation" of the platform. The Next Gen (Rubin Ultra): Industry reports indicate that Nvidia's roadmap includes the Vera Rubin Ultra NVL576 in 2H27, which requires further innovations in bandwidth and power. As CPO moves from its initial adoption (2025-2026) to more widespread, higher-bandwidth implementation, materials that offer better performance (like polymers) have a higher likelihood of adoption. Summary: LWLG polymers are a strong candidate for future CPO upgrades, but it is uncertain if they are in the initial design. They are more likely to appear in later, upgraded iterations of the Rubin/Rubin Ultra architecture rather than the absolute first-generation, finalized design.
Here is a press release from Nvidia from 3 weeks ago showing that Rubin will be offering new versions of usage for AI.
Market Data 
Markets 
AI
