Why Chip Makers Are Pulling Design Tools In-House
The latest wave of design automation acquisitions shows semiconductor companies buying specialized simulation and modeling software so they can control more of the chip design flow, shorten project cycles, and handle the rising complexity of AI and data center systems within tightly integrated environments. As semiconductor design tools grow more central to competitive advantage, leaders are knitting together device, system, and embedded software models that used to sit in separate applications and teams. Instead of pushing files between isolated point tools, engineers can now carry a single model through chip design simulation, multiphysics signoff, photonics simulation software, and embedded software development. That consolidation is especially important for AI accelerators, network switches, and high-speed optical links, where electrical, optical, thermal, and behavioral effects interact. Keysight, Renesas, and Synopsys are each building their own flavor of this connected environment through targeted design automation acquisitions and tightly fused tool portfolios.
Keysight Adds VPIphotonics for End-to-End Photonics Simulation
Keysight’s purchase of VPIphotonics expands its photonic design automation stack from device physics to full system behavior for data center and AI infrastructure designs. RSoft already covers device-level effects in waveguides, gratings, modulators, and laser sources, while Photonic Designer targets circuit-level work on photonic integrated circuits. With the VPIphotonics Design Suite, Keysight can now offer system-level photonics simulation that joins these layers into one continuous flow. A concrete example is VPI Optical Link in Keysight ADS, which simulates the complete electrical–optical–electrical (E-O-E) transceiver path in a single analysis, including bit error rate. That avoids shuttling designs between separate electrical and optical tools and keeps chip design simulation aligned with high-speed digital test instruments. For designers of co-packaged optics and silicon photonics, this kind of integrated photonics simulation software helps expose link problems earlier and compress iteration cycles.
Renesas Uses Pictorus to Bring Embedded Software into the Loop
Renesas is pushing beyond silicon into embedded software development by acquiring Pictorus, a browser-based platform for modeling, simulating, and generating control software. Pictorus lets engineers sketch system behavior as block diagrams in a web interface, simulate those diagrams, and turn them into executable embedded software, generated in memory-safe Rust while still interoperating with C, C++ and Python code. According to eeNews Europe, the Renesas Pictorus acquisition adds cloud behavioral modeling and Rust code generation to the Renesas 365 system-design and lifecycle platform built around Altium tools. Instead of separate steps for component selection, system modeling, firmware coding, and device management, Renesas 365 aims to connect them with a single digital thread. For software-defined automotive, robotics, and industrial systems, that means earlier validation of timing, memory use, and device choices, and faster, more repeatable application development across a fleet of boards and microcontrollers.
Synopsys Multiphysics Fusion Targets AI-Scale Chip Complexity
While Keysight and Renesas are buying companies, Synopsys is knitting together existing assets into its Multiphysics Fusion portfolio, aimed squarely at AI and high-performance computing designs. The offering combines Synopsys EDA tools with Ansys signoff engines so that timing signoff, design closure, multi-die analysis, and analog and photonic design workflows share a multiphysics model. One flow links PrimeTime timing analysis with RedHawk-SC, RedHawk-SC Electrothermal, StarRC extraction, and HFSS-IC so that IR drop, thermal, and stress effects are reflected directly in chip timing. Another uses PrimeClosure and RedHawk-SC to bring power integrity into late-stage optimizations, cutting design iterations on power, performance, and area. GPU-accelerated flows based on NVIDIA CUDA-X libraries support the scale of multi-die packages and co-packaged optics. For design teams, it means fewer blind spots between signoff tools and a higher chance of first-pass success on large, expensive AI accelerators.
A New Playbook for Semiconductor Design Automation
Taken together, these moves point to a design automation landscape where major chip makers and EDA vendors treat simulation platforms as strategic infrastructure, not interchangeable tools. Keysight is tightening its grip on photonics simulation software from device to full E-O-E link behavior; Renesas is bringing embedded software development into its Renesas 365 lifecycle platform through model-based design; and Synopsys is fusing signoff and multiphysics to keep pace with multi-die and co-packaged optical systems. The common thread is consolidation: fewer manual handoffs, more shared models, and broader coverage of physics and behavior within one environment. For AI and data center infrastructure chips, this should reduce time-to-market and late surprises, but it may also raise the barrier to entry for smaller tool vendors that cannot match such deeply integrated, cloud-connected semiconductor design tools.
