Longhorn Silicon

Chips, designed
at Texas.

First student silicon team at UT Austin, taping out an AI inference accelerator on TSMC 16nm.

Check us out

Silicon is the new frontier. We're training people who'll shape it.

Flagship
Introducing Lambda.
Longhorn Accelerator for Matrix-Based Dataflow & Attention. Powering AI models up to 200M parameters
Tapeout
Real silicon. Real packages.
Tapeouts via TSMC University FinFET Program with imec.
FRONTEND
SystemVerilog · UVM
Backend
Cadence · Synopsys
Nodes
TSMC 16nm
Vision
Longhorn Silicon Lambda M.2 module with silver heat spreader and LPDDR DRAM

Roadmap. From charter to silicon.

  • 01
    Spring 2026

    Charter & Tooling

    Secure industry-grade EDA stack and establish versioned RTL and verification infrastructure.

  • 02
    Fall 2026

    Architecture Lock

    Define target workloads, finalize microarchitecture, and lock the system-level specification.

  • 03
    Spring 2027

    RTL Freeze

    Feature-complete accelerator RTL with full design verification signoff.

  • 04
    TBD

    Tapeout

    GDSII submission with LVS and DRC clean on TSMC 16nm.

  • 05
    Post-silicon

    Bring-up

    Silicon validation, board-level bring-up, and first on-chip inference.

Partners & platforms. Standing on tall shoulders.

CadenceChipAgentsBronco AILaboratory for Computer ArchitectureSigmantic AI

Interested?

We're looking for intrinsically motivated, intelligent students, plus industry partners and faculty willing to support.

info@longhornsilicon.com