Specifications
Architecture
Analog Wave Processor (AWP) — weighting, delay and accumulation performed directly on broadband waveforms in the analog domain
Platform
Lotus 200G/lane family
Copper Redriver
TS5802
Optical MZM Driver
TS8802 linear Mach-Zehnder modulator driver
Optical TIA
TS9802 linear transimpedance amplifier
Link Intelligence
TSLink link training and diagnostics via OIF CMIS-LT protocol
Per-Lane Rate
200G/lane PAM4; roadmap to 1.6T and 3.2T modules
Applications
ACC active copper cable, LPO, NPO and CPO co-packaged optics
Equalization
Wave-domain FFE for precursor and postcursor ISI correction; equalization beyond Nyquist
Signal Response
Linear amplitude and phase response across wide bandwidths
Reflection Control
Reflection mitigation via weighted and delayed waveform processing
Power / Latency
DSP-free architecture — lower power and latency than conventional DSP-based retimers
Standards
OIF CMIS-LT; multi-vendor LPO/NPO interoperability demos
Event
Demonstrated at ECOC 2026, Malaga, Spain (September 21-23, 2026)
Vendor
TeraSignal, Irvine, California
Use Cases
AI scale-up fabrics, 51.2T/102.4T switch platforms, GPU cluster interconnect
Overview
TeraSignal's Lotus 200G/lane platform targets the next generation of AI scale-up and scale-out interconnect, where 200G per lane is the building block for 1.6T and 3.2T optical and copper links. The platform pairs the TS5802 copper redriver, the TS8802 linear MZM driver and the TS9802 linear TIA with a shared intelligent connectivity architecture.
At its core is AWP (Analog Wave Processor), a wave-domain signal-processing engine that performs weighting, delay and accumulation directly on broadband electrical waveforms rather than digitized samples. Combined with TSLink link intelligence over the OIF CMIS-LT protocol, it lets linear interconnects correct channel impairments while preserving the low power and low latency that make LPO, NPO and CPO attractive for AI data centers.
Key Benefits
DSP-free linear connectivity cuts power and latency versus retimed DSP modules. Wave-domain FFE and reflection mitigation extend DSP-free link reach. TSLink closed-loop diagnostics deliver link visibility and margin. One architecture spans copper ACC and optical LPO, NPO and CPO at 200G/lane.
Applications
1.6T and 3.2T AI scale-up fabric, GPU cluster interconnect, 51.2T/102.4T switch platforms, LPO/NPO/CPO optical modules, and active copper cables for AI data centers.
Request a Quote — TERASIGNAL LOTUS 200G/LANE AWP PLATFORM — INTELLIGENT LINEAR INTERCONNECTS FOR 1.6T/3.2T AI
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →