Alphawave Semi delivers UCIe IP on TSMC 2nm with 36G die-to-die speeds

Alphawave Semi pioneers next-generation connectivity with its UCIe IP subsystem on TSMC´s 2nm process, reaching 36G die-to-die data rates for advanced Artificial Intelligence and HPC solutions.

Alphawave Semi has announced a significant accomplishment: the successful tape out of one of the industry´s first UCIe IP subsystems fabricated on TSMC´s advanced N2 (2 nanometer) process. The new design supports data rates of 36 gigabits per second for die-to-die communications, a capability poised to dramatically enhance chiplet architectures and the scalability required for upcoming computational workloads.

Critical to this achievement is the complete integration with TSMC´s Chip-on-Wafer-on-Substrate (CoWoS) advanced packaging technology. This integration enables higher bandwidth density and superior scalability, crucial for constructing next-generation systems based on chiplet methodologies. As chip designers adopt disaggregated system-on-chip (SoC) frameworks to meet demanding bandwidth and performance needs, especially in Artificial Intelligence and high-performance computing (HPC) realms, seamless die-to-die links become a cornerstone of infrastructure readiness.

This development builds directly on Alphawave Semi´s recently introduced Artificial Intelligence platform. By achieving the tape out of UCIe connectivity on 2 nanometer nanosheet technology, the company positions itself at the forefront of the emerging open chiplet ecosystem. This step not only underlines support for hyperscale infrastructure but also signals readiness to facilitate scale-up for disaggregated SoCs in a rapidly evolving computing landscape.

77

Impact Score

Memory makers move to hourly contracts as artificial intelligence demand drives volatility

Major memory suppliers are shifting to hourly pricing contracts as artificial intelligence driven demand sends DRAM prices fluctuating by the hour, reshaping leverage between large cloud buyers and smaller firms. Smaller enterprises are being squeezed by rapid cost swings while hyperscalers, automakers, and top smartphone brands secure priority access and better terms.

Bridging the operational artificial intelligence gap in enterprises

Enterprises are moving artificial intelligence from pilots to production, but many struggle without strong integration, governance, and operational foundations. New survey data from senior IT leaders links enterprise-wide integration platforms to broader, more autonomous artificial intelligence deployments.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.