DEV Community

Cover image for UK Startup siliXon Raises $1.5M to Build AI That Generates PCB Designs from Text Prompts
AtlasPCBEngineering
AtlasPCBEngineering

Posted on • Originally published at atlaspcb.com

UK Startup siliXon Raises $1.5M to Build AI That Generates PCB Designs from Text Prompts

From Text Prompt to Circuit Board: siliXon's Vision

UK-based startup siliXon has raised $1.5 million in a seed round led by German early-stage investor System.One, with participation from Antler, to build AI tools that generate complete printed circuit board designs from text prompts. The company aims to make hardware design as accessible as writing a software specification.

How Text-to-PCB Works

siliXon's approach treats circuit board design as a generation problem—similar to how large language models generate text or image models generate pictures from descriptions:

Input example: "Design a Bluetooth Low Energy sensor board with BME280 temperature/humidity sensor, LSM6DSO accelerometer, nRF52840 MCU, CR2032 coin cell power, 25mm circular form factor, I2C debug header"

Output: Complete schematic, component selection, PCB layout, BOM, and Gerber files ready for fabrication.

The system is trained on databases of existing PCB designs, component datasheets, and manufacturing design rules. For well-characterized patterns (IoT sensor nodes, motor controllers, LED drivers, USB hubs), the generation can produce reasonable first-pass designs.

The European Manufacturing Angle

siliXon explicitly frames its mission beyond just design automation. The company wants to "help Europe reclaim its technology supply chain" by:

  1. Lowering barriers to entry: If designing a PCB requires only describing what you want, more European companies can develop custom hardware without outsourcing design
  2. Enabling local manufacturing: Simpler, standard-geometry designs generated by AI are well-suited to European PCB fabricators
  3. Reducing prototype cycles: When iteration is cheap and fast, companies keep manufacturing closer to R&D

Where This Fits in the AI EDA Landscape

The AI-powered EDA market has rapidly segmented into three tiers:

Tier Company Approach Best For
Generative siliXon Text-to-complete-design Beginners, rapid prototyping
Autonomous Quilter Schematic-to-layout Professional engineers, complex boards
Augmented Siemens Fuse, Cadence Cerebrus AI copilot in existing tools Enterprise design teams

siliXon targets the most ambitious tier—generating designs without requiring the user to create a schematic first.

Technical Challenges Ahead

Despite the impressive demo potential, text-to-PCB faces real engineering challenges:

Design rule complexity: A simple text prompt doesn't capture the hundreds of constraints needed for a manufacturable board (impedance requirements, thermal considerations, EMC compliance, testability).

Safety validation: For anything beyond hobby electronics, generated designs need simulation verification (SI/PI analysis, thermal modeling) that current AI tools don't fully automate.

Manufacturing awareness: Without knowing the specific fabricator's capabilities (minimum trace width, layer count limits, material availability), AI may generate designs that are theoretically correct but practically unfabricatable.

Liability: When an AI-generated design fails in the field, the responsibility chain is unclear—this limits adoption in safety-critical markets.

Market Opportunity

The PCB design services market was valued at approximately $4.8 billion in 2025. siliXon's bet is that text-to-PCB can capture the long tail of simple designs that currently go to freelance designers.

If the technology works as promised, it could:

  • Reduce simple PCB design costs from $500–$5,000 to near-zero for tool subscribers
  • Compress prototype timelines from 2–4 weeks to hours
  • Enable non-specialists (mechanical engineers, software developers) to create functional hardware prototypes

The Bigger Picture: EDA's AI Moment

EDA tool revenue for PCB design reached $4.2 billion in Q1 2026, marking 20 consecutive quarters of growth. The acceleration isn't from incremental improvements—it's driven by AI-native features that command premium pricing because they demonstrably reduce design time.

Whether text-to-PCB becomes as transformative as text-to-code remains to be seen. But the investment is flowing, the demos are improving, and hardware engineers should be paying attention.


Originally published at AtlasPCB Engineering Blog

For engineers looking to turn AI-generated or traditional designs into production-quality PCBs, check out AtlasPCB's manufacturing capabilities — free DFM review included with every order.

Top comments (0)