Tattvam AI Raises $1.7M to Automate Chip Design

Tattvam AI was co-founded by CEO Bragadeesh Suresh Babu, an IIT Madras alumnus who previously worked at startups CoMind and Fractile, and Lannan Jiang, a chip developer from a research lab at ETH Zurich. Babu reportedly turned down a position on Google's TPU team to launch the venture. The company is targeting the physical design stage of chip development, a critical bottleneck that converts a chip's architecture into a manufacturable layout. This phase is a notoriously complex and iterative process that can require thousands of engineering hours, remaining largely manual even as other areas of design have seen efficiency gains. Tattvam AI's strategy is not to replace Electronic Design Automation (EDA) incumbents like Cadence and Synopsys, but to build an AI abstraction layer on top of their existing toolsets. The objective is to compress chip design cycles from a multi-year timeline down to just weeks, enabling companies to iterate faster and better align chip capabilities with rapidly evolving AI models. The cost of this process is a major driver for automation, with the design and development of a single 2nm chip estimated to cost as much as $725 million. The largest portions of that expense are software development and the extensive verification required to find bugs before committing to manufacturing. The pre-seed round included participation from notable semiconductor angel investor Stan Boland. Boland has a significant track record in the industry, having previously led the sales of chip company Element 14 to Broadcom and Icera to Nvidia.