• Technical Conference:  30 March – 03 April 2025
  • Exhibition: 01 – 03 April 2025
  • Moscone Center, San Francisco, California, USA

Moore's Law: A Photonics Perspective for the Next Decade

Tuesday, 26 March, 14:00 – 16:00

San Diego Convention Center

Room 6D

Di Che, Nokia Bell Labs, USA
Paul Gunning, BT, UK
Kaida Kaeval, Tallinn University of Technology, Estonia
Emerson Moura, Cisco Systems, Brazil

Gordon Moore observed that the cost of a silicon transistor got cheaper the smaller you could make it. Moore’s Law states that the number of transistors on a silicon integrated circuit doubles every 18 to 24 months. Across six decades, the silicon semiconductor industry has been able to pack more and more transistors onto an integrated circuit.

Functional groups of transistors form CPU/GPU/NPU/TPU silicon chips of great capability. But will all this potentiality be stranded on these chips, like remote islands cut off from the outside world? Where can photonics help?

Moore's Law is entering the Angstrom node era, so what are the prospects for electronics and photonics over the next decade? What exciting technologies will have been implemented and deployed when we re-convene for our retrospective workshop at OFC 2034? Our invited experts will provide their best insights on how the continued advancement of Moore's Law will directly impact:


  • Does Moore’s law apply to photonics - and if so, what are the limits?
  • Opportunities from Quantum technologies.


  • Aggregate bisection bandwidths of silicon switching ASICs approach hundreds of Terabit/s;
  • Migration from copper to optical backplanes within chassis;
  • Evolution of compact pluggable transceivers, co-packaged optics and other advanced technologies;
  • Advancements in A/D & D/A and DSP logic ASICs;
  • High-bandwidth, low-latency, low-jitter, error-free interconnects to support the evolution of AI and photonic AI ‘logic chips’.


  • Datacom & non-datacom;
  • Business opportunities & technology outlook;
  • Refreshment cycles e.g. shorter versus longer operational lifetimes; system dependability.


  • The need for lower energy per bit (fJ/bit?…aJ/bit?) whether computed, switched, transmitted, or stored;
  • Space reduction and power consumption;
  • Perhaps we don't need to evolve Moore's Law; the focus should be on better Network efficiency and better existing resource utilization.

Andreas Bechtolsheim; Arista, USA
Keeping up with Moore's Law

Amit Nagra, Intel, USA
Keeping up with and Enabling Moore’s Law: Role of Photonics I/O

Katharine Schmidtke; Eribel Systems, USA
Moore’s Law Redefined for AI/HPC Systems

Rebecca Schaevitz; Lightmatter, USA
Breaking Down the Interconnect Bottleneck - A Third Dimension

Vladimir Stojanovic ; Ayar Labs, USA
In-Package Optical I/O: Bridging the Gap Between Moore's Law and Amdahl's Law in Modern Compute Systems

Anna Tauke-Pedretti; DARPA, USA
The Path for Scaling Photonic Integrated Circuits