Hardware engineers at scale design chips through remote VNC sessions into aging Linux desktops, using closed-source tools from a triopoly of Cadence, Synopsys, and Siemens EDA — workflows where Git is not even assumed. That's the daily reality Matt Boisvert describes in a February 2026 opinion piece, and it's his launching point for arguing that AI may be the first force capable of meaningfully loosening that grip.
Boisvert attributes EDA entrenchment to a specific combination of pressures absent in most software domains: enormous upfront hardware development costs, export control restrictions on both chip designs and the tools themselves, and decades of supply chain concentration that have given the big three vendors near-total control over advanced process nodes. For historical context on how geopolitics shaped this landscape, he cites Chris Miller's "Chip War: The Fight for the World's Most Critical Technology."
A growing open-source hardware movement exists, but Boisvert frames it as still limited. He points to RISC-V as an open instruction set architecture reducing ARM and x86 licensing costs, Tiny Tapeout as an initiative lowering the barrier to chip fabrication for educators and hobbyists, and SiliconCompiler as an effort to bring modern software infrastructure practices to hardware workflows. He draws an analogy to a hypothetical where Meta and Google never open-sourced React or Angular — while acknowledging the comparison breaks down given EDA's regulatory complexity and manufacturing supply chain constraints.
The article's central argument — captured in its deliberately ironic title — is that proprietary EDA solutions are unlikely to disappear at the leading edge of advanced process nodes, but that AI is actively eroding the switching costs and knowledge barriers that have kept them dominant. This cuts both ways: AI accelerates proprietary capabilities as major EDA vendors race to integrate it into their toolchains, while simultaneously signaling that underlying design intelligence may become more democratizable, gradually undermining the moat. AI also lowers the practical cost of migrating to open-source flows by helping engineers navigate undocumented toolchains, generate automation scripts, and integrate OSS tools with specific business needs.
As a bellwether for AI's broader impact on large-systems engineering, Boisvert highlights a post by LLVM creator and Modular co-founder Chris Lattner titled "The Claude C Compiler," which examines what AI's ability to approximate complex engineering systems reveals about the pace of change ahead. For the agent ecosystem, that matters: as AI reduces the friction of understanding and working within highly specialized, poorly documented toolchains, it opens a category of agentic application — <a href="/news/2026-03-15-blueprint-vibe-create-hardware-design-ai">hardware design assistance</a>, automated verification, OSS flow migration — that has been largely inaccessible to generalist AI tools until now. Boisvert stops short of predicting the imminent collapse of Cadence or Synopsys, but frames the current moment as an inflection point in how semiconductor design expertise is created and distributed.