Device simulation that shows its work.
Greater Today is building an AI-native TCAD platform: a browser workbench for 2D simulation of GaN HEMTs and silicon MOSFETs, with a readable input deck, results you can compare, target search, and an engineering assistant that only reports numbers the solver actually produced.
- 2D drift-diffusionon the GT Engine
- GaN + silicontemplates, plus your own decks
- AIassistant that works through tools
Electron density and drain current solved by the GT Engine for the example deck that ships with the platform, at Vd = 1 V and five gate biases, blended between them. The moving particles are decorative.
Built on open, proven foundations
GT Engine, Bayesian design search, LLM tool calling, Python, NumPy, Any modern browserThe problem
Device simulation is essential. Starting one shouldn't take weeks.
Every new transistor, from a GaN power switch to a silicon MOSFET, is explored in simulation before it is built. But for a university lab or a young hardware team, the tools get in the way of the question.
- Heavy setupWorkstation installs, licence servers and dense input languages come before the first useful curve.
- Slow iterationEach what-if means editing files, re-running and comparing plots by hand.
- Hard-to-check answersResults lose the inputs that produced them, and limits of the models are easy to miss.
Our approach
Established physics, a modern workbench, and AI where it helps.
Physics stays in charge
A drift-diffusion solver computes every number. AI helps set up runs, search designs and explain results — it never replaces the solver.
Everything in one place
Deck, structure, runs, comparisons, design search and downloadable files live in one browser workspace, saved as named projects.
Every claim checkable
Each run keeps its inputs and material snapshot. Partial results stay marked, and model limits are published.
Start from a deck
A few readable lines describe a whole device.
Variables become form fields and search ranges. Materials can be choices. Sweeps live next to the geometry. This is a real example deck that ships with the platform.
- Validated as you type, with jump-to-error
- Variables with units, limits and labels
- Material choices the search can explore
- Import and export as plain
.decktext
title "NMOS with selectable gate dielectric"# Target-search demo: which dielectric and gate voltage give a required drain current?set DIEL=SiO2 options=SiO2,Si3N4,Al2O3,HfO2 label="Gate dielectric"set tox=0.005 min=0.002 max=0.02 unit=um label="Dielectric thickness"set VG=1.0 min=0 max=3 unit=V label="Gate bias" region name=diel material=$DIEL x=0.4,0.9 y=-tox,0region name=si material=Si x=0,1.3 y=0,0.6 electrode name=gate type=gate wf=4.1 x=0.4,0.9 y=-tox,-toxelectrode name=source type=ohmic x=0,0.3 y=0,0electrode name=drain type=ohmic x=1.0,1.3 y=0,0electrode name=body type=ohmic x=0,1.3 y=0.6,0.6 doping uniform region=si acceptor=3e17doping gaussian donor=1e20 x=0,0.4 y=0,0 char=0.0332 lateral=0.7doping gaussian donor=1e20 x=0.9,1.3 y=0,0 char=0.0332 lateral=0.7 models ith=0.1solve op vg=VG vd=1solve idvg vd=0.05,1 vg=-0.5:2:0.1
The workflow
From a few lines of deck to a decision.
One workspace carries a device from definition to a design choice, keeping every run, input and output file.
In the workbench
Built for the engineer at the bench.
Real screenshots of the private build, captured during browser verification with test projects.
Use cases
Questions engineers bring to it.
From GaN power transistors to teaching labs, each starts from an example that ships with the platform.
Explore a HEMT
Gate length, Al fraction and work function against Vth, gm and Ron.
Pick a dielectric
Search SiO₂ or HfO₂ and gate bias for a target drain current.
A browser device lab
Students write decks and see band diagrams, maps and I–V curves.
Narrow the design space
Shortlist the few variants worth careful simulation and measurement.
The assistant
AI is not the physics engine. It makes the physics engine easier to use.
The assistant is an AI model working through the same tools you do. It sets up runs and design searches from a plain-language request, and reports back with run IDs you can open.
- Every number comes from a run.Figures it reports must come from a simulation result in the conversation, never an estimate.
- It states model limits.When a simplification matters — no self-heating, classical 2DEG — it says so.
Engineer
Assistant · tool calls
Reports the best gate length found, its gm and the run IDs — and notes that self-heating isn't modelled.
Illustrative. The tool and metric names are the real ones.
Physics
Established physics, stated plainly.
2D drift-diffusion on the GT Engine, with the model limits published rather than hidden.
AlGaN/GaN HEMT
Polarisation-induced 2DEG, transfer and output sweeps, off-state breakdown estimate.
Silicon NMOS
Threshold, subthreshold swing, on/off current and transconductance.
Custom 2D deck
Rectangular regions, boundary contacts, doping profiles and DC sweeps, defined in text.
How we make claims
Trust before scale.
Numbers come from runs
Results, the assistant's answers and downloadable files all trace back to a simulation you can reopen.
Partial stays partial
If a sweep fails to converge, completed points are kept and the run is marked partial — it can never become an optimiser's best design.
No unearned comparisons
We don't claim accuracy or speed advantages over established TCAD tools before published benchmarks against measured devices.
"AI is not the physics engine. It makes the physics engine easier to use."The principle we build around
Built in India
For a growing semiconductor ecosystem.
India is investing in its own semiconductor industry, through programmes such as the India Semiconductor Mission. New fabs, design teams and university programmes all need engineers who can reason about devices in simulation.
We're building tools that make that easier to learn and faster to use — starting with GaN and silicon, and with university labs as our first users.
Does it replace Sentaurus or Silvaco?
No. It hasn't been benchmarked against established TCAD tools or measured devices, and it doesn't claim equivalent capability. It is for fast exploration; confirm important results with established tools.
Which devices can I simulate?
AlGaN/GaN HEMTs and silicon NMOS from built-in templates, plus your own 2D devices made of rectangular regions, written as a deck.
Do I have to use the AI assistant?
No. Everything works without it. When you use it, your messages and the results it works with are sent to our AI provider to generate replies.
How do I get access?
Email us with who you are, what you simulate and the first question you'd want answered. We reply personally.
Working on GaN or silicon devices?
Create a free account and run your first simulation in minutes. 30 simulations a month on the Free plan.