The software company that quietly designs almost every advanced chip you've ever used - and most you haven't.
Open a phone. Boot a laptop. Train a model in a data center humming on the far side of the country. At some point before any of that silicon existed, an engineer sat in front of a screen running Synopsys software, asking a machine to arrange a few billion transistors into something that works. Synopsys does not make chips. It does not make phones. Its logo appears on nothing you own. And yet it sits inside the design of nearly everything you own.
This is the strange position Synopsys occupies: a $7-billion company that most people have never heard of, holding one half of a near-duopoly over the tools that make modern computing physically possible. Electronic design automation - EDA, to the people who live in it - is the unglamorous layer beneath the glamour. Synopsys is the layer beneath the layer.
In the early 1980s, designing a chip meant something close to drafting it by hand. Engineers laid out logic gate by gate. This worked, in the way that walking to another city works - fine until the distance becomes absurd. Transistor counts were climbing past the point where any team of humans could keep up. The hard problem was not building the chip. It was designing it before the design itself collapsed under its own complexity.
The bet that became Synopsys was simple and, at the time, slightly heretical: what if you described what you wanted the circuit to do, in something closer to code, and let software figure out the gates? That idea - logic synthesis - moved chip design up a level of abstraction. It is the difference between specifying a building and stacking the bricks yourself.
Aart de Geus founded the company on December 18, 1986, with two colleagues from General Electric's advanced computer-aided engineering group, David Gregory and Bill Krieger. They called it Optimal Solutions and set it up in Research Triangle Park, North Carolina - which is a lovely place to do research and a less obvious place to sell to Silicon Valley. In 1987 they moved west and renamed the company Synopsys, a compression of "Synthesis and Optimization Systems." Engineers, it turns out, will name a billion-dollar company after a function call.
De Geus then did the thing founders rarely manage: he stayed. He led Synopsys for roughly 37 years, through more than 100 acquisitions and the slow transformation of computer-aided design into a full electronic-design-automation empire. In January 2024 he handed the CEO title to Sassine Ghazi, a long-time insider who had been chief operating officer, and stepped into the executive chairman seat. De Geus also plays a respectable blues guitar, which is the kind of detail that makes a 37-year tenure marginally less intimidating.
A timeline in which "renamed after a pun" and "$35 billion acquisition" are both true sentences.
If you want to understand Synopsys, picture the assembly line of a chip and notice that Synopsys sells a tool for nearly every station on it. Describe the logic, synthesize it, place and route it, verify it works, emulate it before any silicon exists, and drop in ready-made building blocks so you don't reinvent a USB or PCIe interface from scratch. Then, increasingly, let AI drive the whole thing.
RTL-to-GDSII implementation. The digital design workhorse - 500+ tapeouts from 40nm down to 3nm.
The first autonomous AI for chip design. Reinforcement learning hunts a vast solution space for better power, performance, and area.
An AI layer across the whole flow - design, verification, test, analog, 3D - plus a generative Copilot that writes and debugs RTL.
Functional simulation and exhaustive formal verification - the tools that find the bug before it costs a fortune in re-spins.
Silicon-proven interface, security, and processor IP (PCIe, USB, DDR, Ethernet, ARC) - reusable Lego for SoCs.
Hardware-assisted emulation and FPGA prototyping - run the software before the chip physically exists.
Six products, one quietly radical idea: the chip should mostly design itself, and you should supervise.
It is easy to claim importance. It is harder to claim that the foundries, the chip designers, and the hyperscalers all route their most advanced work through your software. Synopsys can. Its flows are certified on TSMC's leading process nodes; in 2026 the two companies extended that to agentic-AI run assistance on TSMC's A14. Samsung and Intel foundries enable Synopsys IP. NVIDIA, AMD, Qualcomm, and the custom-silicon teams at the big cloud providers are the kind of names that show up when a near-duopoly is doing its job.
Figures are approximate industry estimates shown for scale, not audited shares. The point survives the rounding: a handful of firms tool nearly the entire chip world.
A bar chart of a market so concentrated that "the competition" fits on one hand.
Then there's the 2025 move that changed the shape of the company: the roughly $35-billion acquisition of Ansys, the simulation house behind Fluent, HFSS, Mechanical, and LS-DYNA. Synopsys could already help you design the chip. Now it can help you simulate the heat, the airflow, the electromagnetics, and the structure of the entire system the chip lives in. Design and simulation, under one roof.
Synopsys frames its purpose around enabling the next generation of chips and systems for the AI era - what it now calls engineering the future, from silicon to systems. Strip the slogan and the substance is real: the AI boom is fundamentally a hardware boom, and the hardware boom runs on design tools. The accelerators that train large models were themselves designed in EDA software. Those same models are now folded back into the tools to design the next generation of accelerators. The snake is eating its tail, productively.
The original problem - chips outpacing the humans who design them - never went away. It got worse. Multi-die systems, chiplets stitched together in a single package, 3D stacks, process nodes measured in single-digit nanometers: every advance adds complexity that no human team can manage unaided. Which is precisely the gap Synopsys was founded to fill, now widened to a canyon and increasingly bridged by AI.
So return to the opening scene. A chip is being born, right now, in someone's screen. Forty years ago that engineer would have been drawing gates by hand and losing the race to complexity. Today they describe what they want, an AI explores billions of possibilities, and a working design comes back - faster, cooler, smaller than a human would have found. The logo on the software is still purple, still unnoticed, still attached to almost everything. Synopsys didn't make the chip in your pocket. It made the thing that made it possible to design at all.
Watch & learn: the founders on the show "The Software Behind Silicon" (Acquired), plus product demos and DSO.ai walkthroughs on the official Synopsys YouTube channel.