A brand-new startup from Palo Alto named Architect Labs has officially come out of stealth with $24 million in seed funding. They’re tackling a problem the chip industry has wrestled with for years, making custom chip design faster, cheaper, and way less toiling.
Architect Labs is building an AI-powered platform that automates a lot of the manual work in chip development. Think turning high-level design files into complete chip blueprints and checking that the designs actually work before anyone spends millions on manufacturing.
The seed round was led by Kindred Ventures, with TQ Ventures, Race Capital, and Together Fund also joining in. The investor list includes some serious names in AI and computing, like Google DeepMind chief scientist Jeff Dean, executives from OpenAI and Nvidia, Perplexity AI CEO Aravind Srinivas, and Lukasz Kaiser, one of the co-inventors of the Transformer.
Custom Chips Deriving a New Era
Designing a custom chip today can take around two years and cost multiple millions of dollars. That’s a huge hurdle for companies that want hardware optimized for their AI models, robotics systems, or cloud services.
Architect Labs wants to make chip design as accessible as TSMC made chip manufacturing.
The goal is to let software companies, AI labs, and robotics startups create custom silicon without needing a massive internal hardware team.
How the AI Platform Works
The platform starts with an RTL design which is a high-level blueprint that describes what the processor should do, how many circuits it includes, and how data moves between them. Engineers usually write this in languages like Verilog, which is complex and time-consuming.
Architect’s AI system is built to:
- Generate chip designs from high-level specs
- Verify designs by simulating real-world conditions like temperature and load
- Use formal verification to mathematically check for errors across all possible circuit states
Fewer manual steps, fewer mistakes, and a much faster path from concept to a manufacturable chip blueprint.
Who Will Use This?
Architect Labs isn’t just going after traditional chipmakers. They plan to sell their software to:
- AI model developers who need accelerators built for their specific models
- Robotics startups that want custom hardware for sensors and control systems
- Neocloud operators looking to optimize their infrastructure
The goal is to co-design chips that match each customer’s workload instead of relying on generic off-the-shelf hardware.
Designing custom AI accelerators
Architect Labs is stepping into a market dominated by Broadcom and Marvell, two companies that already design custom AI accelerators for cloud giants like Amazon and Google.
This isn’t a small target. Both have deep experience in custom silicon and massive customer relationships.
By making the design process automated, faster, and more accessible, they’re trying to unlock a new wave of AI-first semiconductor design.
What the Funding Will Do
The $24 million will go toward:
- Scaling up computing infrastructure for training and running AI models
- Deepening research into chip design automation and verification
- Supporting early co-design partnerships with customers
Kindred Ventures founder Steve Jang will join the board as the company moves forward.
