Keysight’s EDA tools are helping AttoTude accelerate advanced AI interconnect ASIC development, cutting design cycles by more than half while supporting first-pass silicon success.
Keysight Technologies’ electronic design automation (EDA) software is helping AttoTude accelerate the development of advanced ASICs for AI and hyperscale data-center interconnects, reducing IC design cycles by more than 50% and cutting the path from concept to tape-out to less than six weeks.
The development highlights the growing importance of simulation accuracy and workflow integration as semiconductor designers push operating frequencies from conventional RF into sub-terahertz and terahertz domains. AttoTude is developing ASICs based on its guided-wave, ASICs-over-Dielectric interconnect technology, targeting per-lane data rates of 200G, 400G and 800G for next-generation AI infrastructure.
At these speeds, interconnect structures increasingly behave as electromagnetic waveguides, making accurate modelling essential. Design errors can result in expensive silicon respins, particularly when engineers are simultaneously developing complex subsystems across RF, sub-THz and THz frequency ranges.
AttoTude has expanded its use of Keysight’s Advanced Design System (ADS) alongside the company’s engineering data-management software to create a more integrated IC development environment. The workflow combines electromagnetic simulation, circuit design, layout data and measurement correlation while providing engineers with shared, version-controlled design information.
A centralised data environment is particularly important for parallel engineering teams, helping maintain consistency between layouts, electromagnetic models and simulation results. The approach also enables system-level scenario planning, allowing designers to evaluate performance trade-offs before committing designs to silicon.
According to AttoTude, the combination has enabled its designs to meet specifications on first silicon across advanced RF, sub-THz and THz tape-outs. Achieving first-pass success is becoming increasingly valuable as AI infrastructure drives demand for faster, higher-bandwidth semiconductor technologies.
The results also demonstrate how EDA platforms are evolving beyond individual simulation tools into connected design environments. For increasingly complex high-frequency ICs, faster development depends not only on more powerful simulation engines but also on maintaining traceability between design decisions, models, layouts and measured hardware performance.
As AI systems demand higher data throughput and greater interconnect efficiency, technologies capable of shortening design cycles while improving silicon predictability could become a significant advantage for semiconductor developers working at the boundaries of high-frequency electronics.
