The chip design cycle is a bottleneck measured in months and millions. Xcelsa Labs, a 2026 spinout from New York University, is betting that an AI system can compress a critical part of that process from weeks to hours [Design-Reuse, November 2026]. Its first public validation is a collaboration with MIPS, the processor architecture firm, aimed at accelerating verified, workload-optimized custom silicon [Edge AI and Vision Alliance, September 2026]. The startup’s seed round, led by Sierra Ventures and joined by NYU’s own venture fund, suggests investors see a wedge.
The Academic Wedge
Xcelsa’s bet sits at the intersection of AI, mathematical optimization, and formal verification. The company describes its offering as "foundational intelligence for chip design," a system that can propose and prove optimized chip layouts [LinkedIn company page]. For companies designing custom silicon for AI, automotive, or edge applications, the promise is a shorter, more reliable path from an application's workload to a production-ready design [Xcelsa Labs]. The core product is an AI system called Apex, focused on chip-design optimization [Xcelsa Labs]. The founders call it a "rapid chip discovery engine" with formal reasoning built in [Chinmay Shringi LinkedIn]. It is a deeply technical sell, aimed at a market where a 10% efficiency gain can justify a nine-figure investment.
The Professor-Founders
The company is built on academic credibility. All three co-founders are NYU professors, a structure that provides research depth but raises questions about commercial tempo. The team is led by CEO Srikanth Jagabathula, a Stern School professor whose public research has focused on analytics and AI model performance, including a study showing AI models exceeding CFA Level III exam standards [NYU Stern, September 2025]. CTO Siddharth Garg and Chief Scientific Officer Chinmay Hegde bring the direct technical expertise in AI and chip design [Xcelsa team page]. The early technical team includes engineers with backgrounds at Qualcomm and in machine learning for 3D geometry, suggesting a focus on the physical design challenges [Jayanand Asok Kumar LinkedIn][Anushrut Jignasu LinkedIn].
| Role | Name | Background |
|---|---|---|
| Co-founder & CEO | Srikanth Jagabathula | Professor, NYU Stern School of Business; analytics & AI research. |
| Co-founder & CTO | Siddharth Garg | Professor, NYU; expertise in AI and chip design. |
| Co-founder & CSO | Chinmay Hegde | Professor, NYU; chief scientific officer. |
| Early Engineer | Jayanand Asok Kumar | Previously at Qualcomm, WiFi/WLAN PHY layer hardware. |
| Early Researcher | Anushrut Jignasu | PhD candidate at Iowa State, machine learning for 3D geometry. |
Traction and the Partnership Play
Public traction is early but pointed. The MIPS collaboration, announced in late 2026, is the most concrete signal. The announcement claims the partnership can deliver "double-digit efficiency gains" and speed up the process "10x faster,with formal verification built in" [Siddharth Garg LinkedIn, 2026]. It is framed as a collaboration, not a disclosed customer contract, but it places Xcelsa’s technology in the workflow of a known player in the semiconductor IP ecosystem. The company is hiring, with a software engineer role listed in San Francisco through late 2026, indicating a build-out phase [Callings.ai]. Third-party data suggests a team size around 10 employees [RocketReach, 2026].
The Capital and the Credibility
Money followed the academic pedigree. Xcelsa closed a seed round in September 2026. The amount is undisclosed, but the investor list carries specific weight [CB Insights, September 2026].
- Sierra Ventures. A traditional early-stage VC firm providing institutional backing.
- NYU Innovation Venture Fund. The university’s own capital, a vote of confidence in the spinout’s potential.
- Osage University Partners. A firm specializing in commercializing university research.
- Doug O’Laughlin. The founder of semiconductor analysis firm SemiAnalysis, whose personal investment acts as a domain-specific signal of technical credibility [X, 2026].
This blend of institutional venture, university support, and industry-specific angel backing is a textbook seed round for a deep-tech academic spinout. It funds the attempt to translate a research wedge into a commercial tool.
Where the Wheels Could Come Off
The risks here are not subtle. The semiconductor design market is dominated by entrenched EDA giants like Cadence and Synopsys, with decades of software integration and customer trust. Xcelsa is not proposing to replace them, but to sell an AI optimization layer that must plug into existing, highly complex workflows. Adoption requires convincing conservative, risk-averse engineering teams to trust an AI’s proposed design with formal verification. The company’s early narrative is heavy on research concepts and light on named, paying customers or disclosed deployment timelines. Furthermore, a founding team of full-time professors must navigate the dual demands of academic tenure and the relentless pace of a venture-backed startup. The bet is that their technical moat is deep enough to justify the slower burn.
The Next Twelve Months
The coming year will be about moving from collaboration to contract. The MIPS partnership provides a platform; the next milestone will be converting that technical validation into a revenue-generating engagement, or announcing a design win with an end-customer building custom silicon. Another seed extension or a Series A will likely be necessary to scale the engineering team and commercial efforts, with valuation anchored to proof that the "weeks to hours" compression claim holds outside the lab. The investor syndicate, particularly Doug O’Laughlin’s involvement, connects the company directly to the semiconductor analysis community, a channel that could be as valuable as the capital.
The seed round from Sierra Ventures, NYU, and a SemiAnalysis founder backs a simple, expensive premise: that the AI flooding software can also optimize the physical substrate it runs on. For chip architects drowning in design-space complexity, is an AI co-pilot that promises double-digit gains worth the risk of a new toolchain? Xcelsa Labs is building the answer.
Sources
- [CB Insights, September 2026] Xcelsa Labs - Crunchbase Company Profile & Funding | https://www.cbinsights.com/company/xcelsa-labs
- [Design-Reuse, November 2026] MIPS and Xcelsa Collaborate to Accelerate Verified Workload-Optimized Custom Silicon | https://www.design-reuse.com/news/202531205-mips-and-xcelsa-collaborate-to-accelerate-verified-workload-optimized-custom-silicon/
- [Edge AI and Vision Alliance, September 2026] MIPS and Xcelsa Collaborate to Accelerate Verified, Workload-Optimized Custom Silicon - Edge AI and Vision Alliance | https://www.edge-ai-vision.com/2026/09/mips-and-xcelsa-collaborate-to-accelerate-verified-workload-optimized-custom-silicon/
- [LinkedIn company page] Xcelsa Labs | https://www.linkedin.com/company/xcelsa-labs
- [Xcelsa Labs] Xcelsa Labs | Vastly Better Compute for the Age of AI | https://xcelsa.ai/
- [Xcelsa team page] Team | Xcelsa Labs | https://xcelsa.ai/team/
- [NYU Stern, September 2025] Research Highlights | Srikanth Jagabathula | AI Passes Wall Street's Toughest Exam | https://www.stern.nyu.edu/experience-stern/faculty-research/ai-passes-wall-streets-toughest-exam
- [Siddharth Garg LinkedIn, 2026] LinkedIn post on MIPS partnership | https://www.linkedin.com/posts/srikanth-jagabathula_physicalai-semiconductors-chipdesign-activity-7490438722350419968-47dg
- [Jayanand Asok Kumar LinkedIn] Jayanand Asok Kumar - WiFi/WLAN PHY Layer HW... | https://www.linkedin.com/in/jayanand-asok-kumar-0453706/
- [Anushrut Jignasu LinkedIn] Anushrut Jignasu LinkedIn profile | https://www.linkedin.com/in/anushrut-jignasu
- [Callings.ai] Software Engineer at Xcelsa Labs San Francisco | https://www.callings.ai/jobs/590b7e98-1925-4fd4-bc51-4123cac3f3a/software-engineer-at-xcelsa-labs-san-francisco
- [RocketReach, 2026] Xcelsa Labs company data | https://www.gritt.io/company/xcelsa-labs/
- [X, 2026] Doug O'Laughlin profile | https://x.com/dougolaughlin