Synopsys and NVIDIA Revolutionize Engineering with AI Workflows

The hum of a massive convention hall is a familiar sound, but this week at the DAC Chips to Systems Conference in San Francisco, it’s underpinned by a different kind of energy. It’s the quiet thrill of watching a machine think. In a corner of Synopsys’ sprawling booth, a screen displays lines of code and complex simulations ticking along at a pace no human team could match. This isn’t just another demo; it’s a glimpse into what Synopsys and NVIDIA are calling the “agentic” future of engineering. The announcement, dense with technical benchmarks, points to a seismic shift: the move from AI-assisted tools to fully autonomous, long-running AI agents that manage entire engineering workflows, from silicon chip verification to system thermal analysis. For an industry perpetually racing against the physics of miniaturization and the economics of time-to-market, this isn’t an incremental upgrade. It’s a potential paradigm shift in how we build the fundamental technology of our world.

Let’s unpack the core of what was unveiled. The headline-grabber is a fully autonomous design verification agent. Chip verification is a notorious bottleneck, a grueling process where engineers spend weeks, even months, hunting for bugs and ensuring a design meets its specifications. It’s iterative, labor-intensive, and consumes enormous computational resources for ever-diminishing returns. Synopsys’ new agent, powered by its AgentEngineer technology and built on NVIDIA’s AI infrastructure—including the Nemotron 3 Ultra model and OpenShell runtime—acts as a master orchestrator. It takes high-level goals, deconstructs them, and autonomously manages a closed-loop workflow from test plan generation all the way to coverage closure and debug. The claimed results are staggering: compressing weeks of work into hours, achieving up to 50 times faster time-to-validated RTL (Register Transfer Level, a design abstraction), and even improving coverage by an additional 20 percent. This moves the engineer from being in the trenches of debugging to a supervisory role, setting intent and evaluating outcomes.

The implications here are profound. As Ravi Subramanian, Chief Product Management Officer at Synopsys, noted, AI is fundamentally reshaping engineering. This agentic approach represents a maturation of that transformation. It’s not about replacing a single task but re-architecting the entire process flow. Similarly, in the realm of thermal management for electronics—a critical challenge for everything from data center servers to electric vehicles—Synopsys demonstrated a fully autonomous CAE (Computer-Aided Engineering) workflow. Using technology from Ansys Icepak, now part of Synopsys, this agent can autonomously handle the tedious setup, pre-processing, and post-processing of complex thermal simulations. What was once a manual, time-consuming series of steps becomes a streamlined, automated insight engine. Tim Costa from NVIDIA framed it perfectly, stating the future is “agentic, where AI agents reason, plan, execute complex workflows and verify their own work.” This self-verification loop is key; it’s what turns a clever script into a trustworthy engineering partner.

Beyond these flagship autonomous agents, the collaboration is supercharging traditional tools through massive GPU acceleration. Synopsys now boasts over twenty GPU-accelerated EDA and multiphysics products. The performance leaps are dramatic. PrimeSim SPICE simulations, the gold standard for circuit analysis, now run up to 18 times faster on NVIDIA GPUs. For materials science, crucial for developing next-generation semiconductors, the QuantumATK platform sees speedups of 50x to 200x for quantum chemistry simulations using NVIDIA’s cuEST libraries and the new Blackwell GPU architecture. These aren’t just marginal gains; they are orders of magnitude that change what’s computationally feasible. Engineers can run more simulations, explore more design corners, and iterate faster, leading to more robust and innovative products. It democratizes high-fidelity simulation, moving it from an occasional, resource-intensive check to a continuous part of the design process.

So, what does this mean for the engineers, the companies, and ultimately, the pace of innovation? The narrative shifts from pure productivity—doing the same work faster—to capability expansion. An engineering team equipped with these agentic workflows can tackle problems with a new level of depth and complexity. They can set more ambitious verification goals, explore more radical thermal solutions, or simulate novel materials that were previously out of reach due to time constraints. The human role evolves toward strategic intent, creative problem-framing, and high-level validation. The grunt work of execution and iterative optimization is handled by the AI. This could significantly compress development cycles for everything from AI accelerators and smartphones to automotive systems and medical devices, accelerating the entire tech ecosystem’s innovation flywheel.

However, this brave new world of autonomous engineering isn’t without its questions and challenges. Trust in AI-generated outcomes is paramount, especially in safety-critical applications. The “black box” nature of some AI models necessitates robust verification of the verifiers themselves. Furthermore, the industry must navigate the skills shift, ensuring that the next generation of engineers is trained to work with these powerful AI agents, leveraging them as force multipliers rather than being displaced by them. The ethical and economic ramifications of such a productivity leap will also need careful consideration, as it reshapes R&D economics and competitive landscapes. As these technologies move from demonstration to general availability planned for the second half of 2026, their real-world adoption and impact will be the ultimate test. For now, standing in that buzzing convention hall, the message is clear: the era of autonomous, agentic engineering is no longer a speculative future. It’s being built, one simulated transistor and one thermal analysis at a time, and it promises to redefine the art of the possible in technology creation.

• Fully autonomous design verification agent
• Compression of weeks of work into hours
• Up to 50 times faster time-to-validated RTL
• Improvement in coverage by 20 percent
• Streamlined CAE workflows for thermal management
• Massive GPU acceleration across products