TYLsemi, a San Jose-based startup, announced on July 14, 2026, that it has raised $43 million in an oversubscribed early-stage funding round. The company, which has emerged from stealth, aims to accelerate the development of AI infrastructure by providing standardized, production-ready “chiplets”—modular pieces of silicon that can be combined to form custom AI chips. The funding round was led by Matter Venture Partners, with participation from Viola Ventures, GHOVC, and Egis Technology. Additionally, TYLsemi secured strategic investments from unnamed companies across the global semiconductor and AI infrastructure ecosystem.
The Shift Toward Modular AI Design
The move toward chiplet-based architectures is driven by the physical limitations of current monolithic chip designs. As AI systems require ever-increasing amounts of compute, memory, and power, the industry is hitting bottlenecks in chip size and connectivity. According to TYLsemi CEO and co-founder Mohit Gupta, the AI accelerator market is projected to reach $604 billion by 2033, with custom silicon XPUs (accelerator processing units) being the fastest-growing segment. “At that scale, chiplet-based design is no longer optional,” Gupta said. Currently, companies such as Meta Platforms have worked with firms like Broadcom to develop custom semiconductors. However, access to high-speed communication technology for these chips is often restricted to proprietary ecosystems. TYLsemi intends to challenge this model by offering open, standards-based chiplets, allowing customers to mix and match components from various providers. “I feel progress happens with standardization,” Gupta told Reuters. “Whenever you do proprietary lock-in, it’s a short-term game. Yes, you can squeeze customers given your position and whatnot, but it’s not healthy for the market.”
A Portfolio of Standardized Components
TYLsemi is the first company to deliver a portfolio spanning IO, power delivery, and memory paired with custom silicon design services. By providing these pre-validated building blocks, the company estimates it can reduce the development time and cost of custom AI silicon by up to 50% from architecture to high-volume manufacturing. The company’s initial product roadmap includes:
| Product | Function |
|---|---|
| TYL.IO | Connectivity chiplets supporting PCIe, ESUN, and UALink, with a roadmap for co-packaged optics. |
| TYL.Power | An integrated voltage regulator (IVR) chiplet for optimized system-level power efficiency. |
| TYL.Mem | A planned product family focused on memory connectivity for advanced AI systems. |
| TYL.Forge | A full-stack platform for customer-defined XPU and compute designs. |
Leadership and Industry Context
TYLsemi was co-founded by Mohit Gupta and Sunil Bhardwaj. The pair previously led engineering and operations teams at several prominent semiconductor firms, including Alphawave (acquired by Qualcomm), SiFive, Cadence Design Systems, and Rambus. The company was established less than a year after Qualcomm’s acquisition of Alphawave. Industry experts note that the ecosystem is currently struggling to keep pace with the shift toward modular design. Jim Handy, General Director at Objective Analysis, stated that the company’s ability to deliver pre-validated, standards-based silicon represents a significant opportunity to accelerate AI infrastructure deployment. Wen Hsieh, Founding Managing Partner at Matter Venture Partners, noted that the startup’s foundational technologies are intended to make the design process “faster, less risky, and more accessible,” potentially opening the door for mid-tier infrastructure players and emerging AI companies that previously lacked the resources for full custom silicon development. TYLsemi reports that it is already engaging lead customers for its TYL.Forge platform. Samples of the TYL.IO and TYL.Power chiplets are expected to be available to qualified customers in 2027 through a partnership with TSMC.
Find more reporting in our Technology section.
Keep reading
Discover more from Archyworldys
Subscribe to get the latest posts sent to your email.
