← systems
Frontier · Photonic

Silicon-photonic MAC tile

Matrix-multiply done in light at 1550 nm — sub-picojoule per MAC. The same descent, one paradigm out.

baseline: open silicon photonics — GDSFactory / SiEPIC components, AIM Photonics PDK
carrier 1550 nmenergy sub-pJ / MACPDK AIM Photonics (open MPW)core MZI mesh

The frontier exemplar, and the point where the classification flips. On the ring, the compute (the MCU) was settled reuse and the frontier sat in the sensor. Here the compute itself is on the frontier — there is no photonic matrix core to buy off a reel — in the substrate where a multiply-accumulate costs the fewest joules. The passives, the laser, the readout, and the drive electronics around it are still settled reuse. Whether that frontier is worth it depends on the goal.

5 of 8 settled or cyclical — 3 on the frontier. Each part placed by cost and differentiation — the landscape, not a verdict. The true cost of building without knowing the market and the benefit, surfaced — so you choose by your goal.

Exploded

Decompose the product into its stack. Each layer is colour-coded: settled, cyclical, or frontier.

1550 nm laser source (off-chip) Settled
Grating couplers — fiber I/O Settled
MZ modulators — input encode Frontier
MZI mesh + phase shifters — the matrix core Frontier
Ge photodetectors — readout Settled
CMOS TIA / DAC / ADC — drive Cyclical
Si / SiN waveguide substrate Settled

Bill of materials

Settled ×4 Cyclical ×1 Frontier ×3

Every part is a real, sourced component from the open baseline — no invented part numbers.

MZI mesh (Clements/Reck) SiEPIC / GDSFactory components Frontier · higher cost and risk — differentiation, SOTA, or science
Matrix-multiply core — triangular/rectangular Mach-Zehnder mesh — the programmable linear-algebra fabric

This is the computer. There is no off-the-shelf photonic matrix core — the mesh topology, the loss budget, and the calibration are the design. The multiply-accumulate primitive realised in the substrate where it costs the least energy.

https://gdsfactory.github.io/gdsfactory/ →
Thermo-optic phase shifter PDK standard cell Frontier · higher cost and risk — differentiation, SOTA, or science
Weight tuning — heater-tuned π phase shift per MZI arm

Thermal phase shifters burn static power continuously — the scaling constraint. A frontier option: non-volatile phase shifters (phase-change / MEMS / BTO) that hold weights at zero idle power, at the cost of a less-proven device.

Mach-Zehnder modulator AIM Photonics PDK Frontier · higher cost and risk — differentiation, SOTA, or science
Input encode — electro-optic intensity modulator @ 1550 nm

Compact slow-light EO modulators (TeMPO-class) cut the area and drive energy of the encoder — co-designed with the CMOS driver below.

https://www.aimphotonics.com/pdk →
Grating coupler AIM / SiEPIC standard Settled · mature — lowest cost and risk
Optical I/O — fiber-to-chip coupler, ~1550 nm band

Mature PDK passive. Reuse as-is.

https://github.com/SiEPIC →
Germanium photodetector AIM Photonics PDK Settled · mature — lowest cost and risk
Readout — on-chip Ge PD, 15 Gbps tested; squared modulus gives the nonlinearity

Standard, qualified PDK component. Reuse.

https://www.aimphotonics.com/pdk →
1550 nm DFB laser commodity telecom Settled · mature — lowest cost and risk
Light source — off-chip continuous-wave carrier

Telecom-volume commodity. Reuse.

CMOS TIA + DAC/ADC standard mixed-signal Cyclical · refreshes on a clock
Electronic drive — transimpedance readout + modulator drivers + converters

Standard CMOS that refreshes on a node clock — reuse the current generation; the co-design opportunity is the interface, not the process.

Si / SiN waveguides PDK passives Settled · mature — lowest cost and risk
Substrate — low-loss routing on SOI / silicon nitride

Settled PDK passive layer. Reuse.

Where openWafer can help

At a new paradigm the cost-and-differentiation question dominates the whole compute path. The MZI mesh, the phase-shifter technology, and the modulator are the matrix-multiply primitive realised in light, in the substrate where it dissipates the fewest joules per operation; the passives, laser, photodetector, and CMOS drive are settled reuse. This is the case where the frontier is most of the system — a fit when the goal is differentiation or advancing the science, a poor fit when the goal is to ship a known product cheaply. One frontier move that makes photonic compute scale is replacing power-hungry thermo-optic tuning with non-volatile phase shifters that hold weights at zero idle power. Same descent as the ring, one paradigm out: the goal decides whether the frontier is worth it.

Open sources — decomposed and linked, not hosted