Loading...
News Article

TYLsemi launches chiplet platform

News

Backed by $43 million in early-stage funding, TYLsemi aims to accelerate custom AI chip development through a chiplet platform that combines production-ready dies with advanced packaging and supply-chain integration.

TYLsemi has emerged from stealth with $43 million in early-stage funding to commercialize a chiplet-based platform designed to simplify and accelerate the development of custom AI accelerators.

The funding round was led by Matter Venture Partners, with participation from Viola Ventures, GHOVC, Egisten and strategic investors from the semiconductor industry.

The startup is targeting the growing demand for custom AI silicon by offering production-ready, standards-based chiplets that can be integrated into advanced packages, reducing the cost and complexity of designing application-specific processors.

The company claims its approach can cut custom silicon development costs by more than 50% while reducing development time by more than half.

TYLsemi's platform leverages 2.5D and 3D packaging technologies to integrate heterogeneous compute, memory, power and connectivity chiplets into a single high-performance package.

Built around the UCIe die-to-die interconnect standard, the company says its offering enables customers to reuse validated building blocks rather than developing every component from scratch.

Its portfolio includes TYL.IO connectivity chiplets, TYL.Power integrated power delivery chiplets, the forthcoming TYL.Mem memory connectivity solution, and TYL.Forge, an end-to-end platform covering chiplet integration, advanced packaging, foundry management and production readiness.

Looking ahead, TYLsemi's roadmap also includes support for co-packaged silicon photonics, targeting future AI infrastructure that requires high-bandwidth optical interconnects.

The company expects to begin providing samples of its TYL.IO and TYL.Power chiplets to qualified customers next year through manufacturing partner TSMC, as it works to make custom AI silicon development more accessible beyond hyperscale cloud providers.