Tattvam AI Raises £1.3M Pre-Seed to Build a Reasoning Model That Designs Semiconductor Chips From First Principles

February 25, 2026

Tattvam AI, the London-based semiconductor design automation company, raised £1.3 million in a pre-seed round in February 2026, led by Seedcamp, with participation from EWOR, Entropy Industrial Ventures, Concept Ventures, and angel investor Stan Boland. The funding will be used to expand the engineering team, accelerate research, bring the chip design product to market, and strengthen partnerships with global semiconductor design teams. Tattvam AI was founded in 2025 by Bragadeesh Suresh Babu (CEO) and Lannan Jiang.

The semiconductor chip is the most complex manufactured object in human history. A modern processor contains billions of transistors, connected by layers of interconnects designed to precise nanometre specifications, with every architectural decision producing cascading trade-offs between performance, power consumption, heat generation, manufacturing yield, and cost. Designing such a chip is a multi-year exercise involving hundreds or thousands of engineers working through a series of abstraction layers — from system architecture to logic design, physical implementation, verification, and sign-off — each of which requires specialised expertise and iterative refinement. The time and cost of chip design has become one of the most significant constraints on innovation in the semiconductor industry: a major processor design can cost hundreds of millions of dollars and take three to five years from initial specification to tape-out. For companies seeking purpose-built chips optimised for specific applications — AI acceleration, edge inference, automotive, industrial — this barrier is prohibitive.

Existing AI tools, including large language models applied to engineering tasks, have made inroads into software development but have struggled with hardware design. The reason is structural: chip design requires not just pattern recognition across a corpus of existing designs but genuine reasoning about the physical and logical constraints of circuits — understanding why a particular architectural choice creates a power budget problem downstream, or why a routing decision at the physical layer constrains performance at the logic layer. Tattvam AI is building a reasoning model specifically for this task: a system that represents circuits from first principles, understands the constraints and trade-offs at each level of the design hierarchy, and can navigate the interdependencies between them in the way a world-class chip architect would — but at a speed and consistency that human teams cannot achieve. Stan Boland, the angel investor in the round and a serial semiconductor entrepreneur, is well positioned to assess the commercial significance: his background includes founding multiple chip design companies and building products across multiple generations of mobile and networking silicon.

Seedcamp’s participation in the round is significant: the fund has a strong track record in deep-tech infrastructure companies and its involvement validates both the technical approach and the team’s ability to execute. EWOR’s participation adds the structured founder development support that is characteristic of the accelerator’s approach. Entropy Industrial Ventures brings domain expertise in industrial deep tech. The ambition Suresh Babu has articulated — reducing chip design timelines from years to weeks — is the kind of step-change that historically only materialises when a new computational paradigm (EDA tools in the 1980s, synthesis tools in the 1990s, place-and-route automation in the 2000s) transforms a previously manual process. Tattvam AI is placing a bet that AI reasoning models are the next such paradigm.

Sources