research at Wing ItOpen Telos ↗
← Talos
HardwareIn development

Talos 2

A programmable NPU for 180 nm.

TALOS 2−1 0 +1INT8 → ACCUMULATE → INT32GF180 / 180 nmArchitecture illustrationIn development
Process target
GF180 · 180 nm
Weights
Ternary: −1, 0, +1
Activations
Signed INT8
Accumulation
Signed INT32
Development status
Architecture, reference, and RTL work

Talos 2

Talos 2 is a programmable embedded inference accelerator targeting the GF180 180 nm manufacturing process. It is the second generation in the Talos neural chip family, moving toward tiled execution of trained dense layers using external model and working memory.

The current design uses ternary weights and INT8 activations, with a shared add/subtract/accumulate engine and INT32 partial sums. The intended host interface is SPI.

Fabrication is planned for December 2026.

Current work

The work includes a numerical contract, a software reference, and RTL components for arithmetic, buffering, scaling, and partial sums. Complete-engine integration, physical fit, timing, and end-to-end performance remain to be validated.

Spiking state updates are a possible later extension. Version 1 focuses on programmable dense inference.

Architecture

Host SPI→Sequencer→External memory↔Input buffer→Ternary arithmetic→INT32 partial sums

Schematic of the planned data path, based on the implementation plan. Status: in development.