Investors

Anvil Sim is AI-native engineering simulation software.

The first wedge is chip design simulation for AI chips and semiconductor packages. The larger opportunity is a multiphysics platform for thermal, structural, fluid, electromagnetic, and coupled engineering work.

Read investment case →Watch flagship demo Discuss pilot fit

why now

Simulation is becoming an AI workflow.

Existing engineering tools prove the market. Anvil is building the next interface: AI-assisted setup, solver evidence, benchmark comparison, and reviewable artifacts.

wedgeAI chip simulation

stackRust + Swift + Python

modelAI-first R&D

trustBenchmarks + demos

business opportunity / market sizing

A real market with expensive problems.

Engineering simulation is a serious software category because the pain is concrete: faster design loops, fewer failed prototypes, better thermal decisions, safer structures, and more confidence before hardware spend. Anvil is aimed at that budget, not a thin AI assistant category.

Large proven category

Simulation software already commands real engineering budgets because bad physical decisions are expensive.

Demand is expanding

AI chips, robotics, electrification, cooling, and advanced manufacturing create more design loops that need serious simulation.

AI changes the workflow

The opportunity is not a chat box on top of CAE. It is software that helps engineers set up, run, debug, compare, and explain simulations.

competitive landscape

Competing in a proven category, not inventing demand.

The named landscape includes established simulation incumbents and AI-native challengers. Mentioning them is market context, not an affiliation claim.

AnsysAbaqusCOMSOLSimScalenTopNeural ConceptPhysicsXLuminary

why anvil can win

Why Anvil can move differently.

Anvil is not trying to clone legacy CAE. It is built around a cheaper, research-heavy operating model, a modern native stack, and a narrow path to customer value.

AI-first operational model

ADVANTAGE

Anvil is built around agentic R&D, deep research, benchmarks, and fast iteration from day one. That keeps the build loop cheap and unusually broad for deep-tech simulation.

Not an inference-cost trap

ADVANTAGE

Anvil can support local models and customer-provided API keys. Product usage does not require Anvil to burn hosted inference cost on every customer action.

Apple Silicon wedge

ADVANTAGE

SwiftUI, Unified Memory, Metal, and MLX create a native macOS path for engineering simulation and local AI workflows as more technical work moves onto Apple hardware.

Modern technical stack

ADVANTAGE

Rust solvers, Swift UI, and Python scripting give Anvil a quality-first base without inheriting decades of legacy desktop simulation debt.

Benchmark discipline

ADVANTAGE

The public benchmark wall keeps references, QoIs, status, artifacts, and next steps visible so product claims grow with evidence.

Narrow wedge, broad platform

ADVANTAGE

The short-term focus is chip design simulation for AI chips and semiconductor packages; the long-term product is multiphysics engineering simulation software.

why me

Mechanical engineering, software engineering, and AI systems.

Mechanical engineering and mathematics training at UC San Diego, class of 2023.
Senior software engineer with three years of full-time industry experience.
AI engineer building agentic systems, reading current research, and turning research into shipped software.
Pre-agentic-era projects included autonomous vehicles, a JWST replica, and a database engine.
Thousands of repo-backed commits across the current Anvil build sprint, with public demos and benchmark rows shipped.

near-term focus

Chip design simulation is the wedge.

The first commercial direction is an end-to-end simulation workflow for chip designers, AI chips, and semiconductor packages.

AI chip and semiconductor package thermal workflows
Cooling airflow, heat sink, and enclosure evidence
Structural stress, warpage, and modal checks
Electromagnetic, PDN, and coupled chip-package review

Open chip design tab →

Flagship visual demo

Boomerang first: trajectory and diagnostics.

A 3D flight-path replay leads into return-window diagnostics and the evidence packet behind the demo.

EVIDENCE

Return-window evidence path

The rendered trajectory makes the motion legible. The evidence packet links the demo to the configured v4 return-window solver gate and checked-in provenance.

evidence packet: v4 return-window manifest, available in reviewer packet

Visual: flight-path replay

Solver gate: v4 return window

Provenance: .anvilprov manifest

Validation: flight-validation path

Evidence stack

Evidence first, claims second.

Boomerang flight-path replay

EVIDENCE

Flagship visual opener: 3D trajectory replay plus return-window diagnostics

Rendered motion is paired with the configured v4 return-window solver gate.

Investor simulation evidence reel

EVIDENCE

Long-form walkthrough: FEA fields, modal animation, native capture, FRF/NVH

100s 1080p60, 6,000 frames, receipt-backed source MP4s and evidence packets

FEA + FRF evidence reel

EVIDENCE

Static FEA, modal FEA, native receipt, forced-response FRF

5 clips, 6 MP4s, 38 .anvilprov hashes, ffprobe checked

FRF cantilever benchmark

EVIDENCE

8,150 Hz Anvil peak vs 8,155 Hz analytic reference

10 verified .anvilprov artifacts, Bode magnitude + phase lag

D24 native modal overlay

EVIDENCE

AnvilNative CGWindowID capture with Blevins rows visible in-video

1920x1080, 480 frames, window_id 198012, max Blevins error 1.9%; decision readiness expansion is planned

Captioned overview

EVIDENCE

Robot-cell bracket FEA, drone-arm modal, Tet10 cantilever modal

32s 1080p60 overview with burned-in truth captions

Benchmark ledger

Reference passes and expansion paths.

Domain	Benchmark row	Status	Validation focus
Boomerang flight path	Boomerang return-window gate	SCENARIO EVIDENCE	Current public evidence is scenario-local return-window evidence; measured-flight and wind-tunnel correlation are the next expansion path.
Structural dynamics	FRF cantilever harmonic response	REFERENCE PASS	Current public evidence is cantilever harmonic response against analytic and CalculiX references.
Static FEA	Static cantilever Hex8-EAS stress	REFERENCE PASS	Current public evidence is integration-point stress evidence with mesh refinement.
NAFEMS structural	NAFEMS linear-elastic core rows	REFERENCE ROWS	Current public evidence separates passing linear-elastic rows from modal and shell expansion rows.
Modal dynamics	NAFEMS FV4 offset-mass cantilever	VALIDATION TRACK	Mode-shape, model-form, and dense-solver comparison are the current validation focus.

Full public status now lives on the benchmark wall: reference, QoI, validation focus, and latest evidence.

Open benchmark wall →

Validation plan

Current evidence and expansion path.

Topic	Current truth
Certification	Certification workflows require dedicated evidence packages.
Solver evidence	Strongest today: linear static, modal, thermal, and FRF rows with visible validation steps for broader claims.
Nonlinear	Plasticity, large deformation, and full transient dynamics sit on the roadmap.
Competitive position	Current public position is design-stage screening with reproducible evidence and a growing benchmark wall.