OpenLight Dependency on Tower PH18DA

This risk goes to the heart of OpenLight’s value proposition, because the company is selling not just photonic IP, but a promised path from design file to shipped chips on a very specific manufacturing stack at Tower. OpenLight’s integrated laser advantage depends on Tower’s PH18DA process, and the company also promises first article performance and 8 to 12 week delivery, so any line interruption, capacity squeeze, or reprioritization at Tower can hit both revenue timing and customer trust at the same time.

• PH18DA is not a generic silicon photonics node. It is the process that lets OpenLight put indium phosphide lasers and other active optical parts directly onto silicon photonics wafers, which is the main technical reason customers use OpenLight instead of a standard passive silicon photonics foundry.
  1 sacra 3 openlightphotonics
• Tower is both partner and potential bottleneck. OpenLight relies on Tower for production, while Tower also markets silicon photonics services to other customers and continues expanding work on the same platform with companies like Innolight and Salience Labs. That means OpenLight does not fully control fab priority on the process it depends on most.
  1 sacra 3 openlightphotonics 5 towersemi 6 towersemi
• Competitors give customers a fallback, but not the same product shape. GlobalFoundries Fotonix offers high volume silicon photonics manufacturing and stresses supply confidence on 300mm lines, while Intel ships its own laser integrated photonics at scale inside a closed product stack. OpenLight sits between those models, with differentiated technology but less manufacturing control.
  7 gf 8 intel

The next phase of competition will be won by the photonics platform that can pair integrated lasers with dependable, repeatable volume supply. If OpenLight broadens foundry access or deepens process ownership around PH18DA, it strengthens its pitch to AI optics customers who care as much about delivery certainty as device performance.