Skip to content

AI-native engineering

Design chips you can trust.

The Design workspace showing the sample LNA's schematic with transistor Q1 selected, and its recorded result: gain 28.6 dB at 10 GHz, simulated.

Daedyl is an AI-native engineering environment for semiconductor design. Design, simulate, optimize, validate, and prove your work with Daedyln, an AI engineering agent that works alongside you throughout the engineering process.

  • AI-native engineering
  • Human-controlled
  • Evidence-backed

One engineering environment

Design. Simulate. Optimize. Prove.

Move from engineering intent to validated results in one AI-native environment, without losing context between tools, models, decisions, or evidence.

Explore all capabilities
  • Design across abstraction levels

    From requirements and architecture to schematics, RTL, layout and implementation with Daedyln working in context.

  • See how the design behaves

    Run circuit, RF, electromagnetic, thermal and device analysis without leaving the engineering environment.

    -30-1501530251020Frequency (GHz)dB
    • S21 28.6 dB
    • S11 -12.2 dB
    • S22 -12.1 dB
    • at 10.000 GHz
  • Know what supports every result

    Connect requirements, decisions, artifacts, simulations, results and approvals into an auditable engineering record.

    1. REQ-001Gain ≥ 25 dB
    2. DEC-017Cascoded LNA
    3. SIM-042AC simulation
    4. RES-071Gain 28.6 dB
    5. APP-012Design review

From intent to evidence

One continuous engineering workflow.

Daedyl maintains engineering context as work moves from an idea to implementation, simulation, optimization, and validation. Daedyln works across that entire lifecycle rather than operating as a separate chatbot.

  1. 01

    Define

    Requirements

  2. 02

    Design

    Schematic / RTL

  3. 03

    Build

    Layout / Implementation

  4. 04

    Simulate

    Circuit / EM / Thermal

  5. 05

    Optimize

    Explore trade-offs

  6. 06

    Validate

    Evidence / Approvals

In 3D

See the chip you designed, layer by layer.

Daedyl turns a project's own routed layout into a 3D model: every metal layer, via and standard cell, taken from the recorded GDS and placement. Pull the stack apart, isolate a layer, follow a net from the code to the silicon, or render it for a design review.

  • Interactive in the app

    Drag, zoom and explode the stack, switch layers on and off, and click any cell to see what it is and where it came from in the RTL.

  • Assurance on the geometry

    Colour the chip by what matters: radiation-hardened copies and voters, timing paths, power or temperature.

3D render of a chip's routed layout with its metal layers pulled apart: copper power straps over a dense mesh of signal wires and vias, above the cells on the silicon die.Exploded metal stack · Blender, from the GDS
The same chip rendered with each layer in its own colour: metal straps, signal routing and cells.
Daedyl's demonstration design: a radiation-hardened lander controller (fictional) on SkyWater SKY130, rendered from its recorded layout. 699 cells after triple modular redundancy on every register; its layout passes DRC in KLayout and Magic and matches its schematic (LVS). Not a flight design.
Heat map of the die: temperature from 25.2958 °C at the cool lower-left corner to 25.2981 °C at the hot spot in the upper right, where the busiest cells sit.
The same temperature field painted onto the chip's 3D model in Blender: the cells toward the upper right warmest, those at the lower left coolest.
Recorded thermal solve of the demonstration design at the typical corner. Stated assumptions: 300 µm silicon on a 25 µm die attach (2 W/m·K) to a 25 °C sink.

Thermal

Know where the heat goes before the chip exists.

Daedyl takes the power each placed cell draws from the timing analysis, spreads it over the die exactly where the layout put it, and solves the temperature with a finite-element solver. The map lands on the layout and on the 3D model, and every run checks its own energy balance.

0.64 mW

power of the placed cells (OpenSTA)

+0.30 K

peak rise above the sink

164 × 164 µm

die, binned 16 × 16

0.002 %

heat out vs power in

A larger or faster design shows its hot spots the same way, so cooling and placement can be changed while it is still a layout.

Daedyln

An AI engineering agent that works where you work.

Daedyln is built into Daedyl and works directly with the engineering context around it. It can reason over requirements, designs, simulations, results, evidence, and project history while keeping engineers in control of consequential decisions.

  • Select a transistor

    Ask Daedyln about Q1

    Understand its operating point, dependencies, results and evidence.

  • Select a timing path

    Explain this violation

    Inspect the path, constraints, implementation state and evidence.

  • Select a requirement

    Show supporting evidence

    Trace the requirement through design decisions, simulations, results and approvals.

  • Select a simulation

    Compare against an earlier run

    Understand what changed, what improved, what regressed and which parameters caused it.

Daedyln does not sit beside the engineering environment. It operates inside it.

DAEDYLSearch projects, files, tools, or ask Daedyln…
Daedyln answering about transistor Q1 on the sample project, with the records its answer rests on. Example IDs; values from a recorded simulation.

Built AI-native

AI isn't a feature bolted onto Daedyl.

Daedyl was designed around collaboration between engineers and AI from the beginning.

  • Context-aware

    Understands the project, selected object, requirements, artifacts, results and history.

  • Tool-capable

    Helps prepare and execute engineering workflows through Daedyl's integrated capabilities.

  • Evidence-aware

    Answers and recommendations can be connected to the evidence that supports them.

  • Human-controlled

    Assists, analyzes, explains and proposes. Engineers retain authority over approvals and consequential decisions.

Security

Control stays with you.

How Daedyl runs today. Ask us about your own deployment requirements.

  • On infrastructure you control

    Daedyl runs on hardware the team owns, on a private network, not on a shared public service.

  • One sign-in, checked every time

    Single sign-on through Daedyl's own identity provider; every request is checked against the project and the person's role.

  • Every action on the record

    Changes, runs, approvals and Daedyln's own actions are written to an append-only audit history.

  • Approvals by people

    Consequential approvals need a person with the right role; any AI approval must be granted explicitly and is labelled as the AI's.

Engineering platform

One interface. Deep engineering underneath.

Daedyl brings specialized design, simulation, verification, optimization and analysis capabilities into one environment.

  • Digital Design

    RTL · Synthesis · Physical Design · Timing · Verification

    Sample project, recorded simulation.

  • Circuit & RF

    Circuit Simulation · RF Analysis · Noise · Electromagnetics

  • Multiphysics

    Thermal · Fields · Device Physics · Coupled Analysis

  • Process & Materials

    Process Modeling · Materials · Device Modeling

  • PCB & Systems

    PCB Design · Architecture · System Modeling

Engineering evidence

Every number should have a history.

Daedyl connects requirements, design decisions, artifacts, simulations, results, and human approvals so engineers can understand not only what a result says, but how it came to exist.

Explore evidence
  1. REQ-001Requirement

    Gain ≥ 25 dB

  2. DEC-017Design decision

    Cascoded LNA architecture

  3. SIM-042Simulation

    AC and noise analysis

  4. RES-071Result

    Gain 28.6 dB at 10 GHz

  5. APP-012Approval

    Design review

Example on the sample project. The result is its recorded simulation; the record IDs are illustrative.

What Daedyl can do today.

An AI-native environment for engineering work.

  • RF & Analog

    Circuit simulation, S-parameters, noise, electromagnetic and thermal analysis.

  • Digital IC

    RTL, synthesis, physical design, timing, DRC/LVS and GDS workflows.

  • Design Exploration

    Parameter sweeps, optimization, sensitivity analysis and candidate comparison.

  • Evidence & Provenance

    Trace requirements through decisions, artifacts, simulations, results and approvals.

  • Daedyln

    Context-aware AI engineering assistance throughout the project lifecycle.

  • Compute

    Execute engineering workloads through managed or controlled compute infrastructure.

Built for engineering organizations where mistakes are expensive.

  • 01

    Defense & Aerospace

    Engineering workflows where traceability, controlled execution, evidence and human accountability matter.

  • 02

    Semiconductor Companies

    Move from requirements through design, implementation, simulation, optimization and verification in one environment.

  • 03

    Research Labs

    Connect models, experiments, simulations, literature, results and engineering decisions without losing provenance.

Questions

Common questions about Daedyl.

Daedyl

Engineering software should show its work.

Design. Simulate. Optimize. Prove.

Build hardware with an AI-native engineering environment where Daedyln works alongside your team and every important result can be traced back to the engineering that produced it.

The Design workspace with the sample LNA's schematic, and a result card: gain 28.6 dB at 10 GHz, simulated.

Request access

Daedyl is opening to a small number of teams first. A way to ask for access will be here soon. Nothing is collected on this page.