AI chip startup TYLSemi has raised $43 million in funding to develop open-standard AI chiplets that could make custom AI chips faster, cheaper, and easier to build. The company wants to give businesses an alternative to proprietary chip technologies that currently dominate the AI infrastructure market.
TYLSemi Is Betting on Open AI Infrastructure
The funding round marks an important milestone for the San Jose-based semiconductor startup.
TYLSemi is building modular chip components known as chiplets. Companies can combine these building blocks to create custom AI accelerators for their own workloads and applications.
Unlike traditional AI chips, these chiplets are based on open industry standards rather than proprietary communication technologies.
As a result, businesses gain more flexibility when designing their AI hardware.
What Are AI Chiplets?
Chiplets work like LEGO blocks for semiconductors.
Instead of designing an entire AI processor from scratch, companies can combine smaller specialized chips into a single package.
This approach offers several advantages:
- Faster development cycles
- Lower engineering costs
- Easier upgrades
- Better scalability
- More design flexibility
The model has become increasingly popular as AI systems grow larger and more complex.
Challenging Proprietary AI Ecosystems
Today, companies such as Broadcom and Marvell provide proprietary technologies that help AI chips communicate at high speed.
However, businesses that want access to those technologies often need to work directly with those suppliers.
TYLSemi wants to change that.
The company plans to offer open-standard interconnect technologies that customers can mix with solutions from multiple vendors.
This strategy could reduce vendor lock-in and encourage more competition across the AI hardware industry.
Experienced Founders Lead the Startup
TYLSemi was founded by semiconductor veterans Mohit Gupta and Sunil Bhardwaj.
Both founders previously worked at AlphaWave, the connectivity company acquired by Qualcomm in 2025.
Their experience also includes leadership roles at companies such as SiFive, Cadence Design Systems, and Rambus.
Investors Back the Open Chip Vision
The $43 million funding round was led by Matter Venture Partners.
Other investors include:
- Viola Ventures
- GHOVC
- Egis Technology
- Strategic semiconductor investors
The round was reportedly oversubscribed, highlighting strong investor interest in AI infrastructure startups.
Demand for Custom AI Chips Continues to Grow
The AI boom has created massive demand for custom silicon.
Companies such as Meta, Microsoft, Google, and Amazon are increasingly developing their own AI chips to reduce costs and improve performance.
However, building custom semiconductors remains expensive and time-consuming.
TYLSemi says its chiplet platform could reduce AI chip development costs and timelines by as much as 50%.
Why This Matters
The future of AI may depend as much on hardware as software.
While companies continue competing to build better AI models, the race to develop faster and more efficient AI infrastructure is accelerating.
Open chip standards could play a major role in that future.
If TYLSemi succeeds, businesses may gain a new way to build AI hardware without relying on a single supplier or proprietary ecosystem.
By
