At Yuji Labs, we’re building a hardware-and-software powered industrial intelligence platform for core sectors. Currently we are focused on Green Hydrogen and renewable sector.
Industrial equipment works tirelessly but can’t signal distress. In green-hydrogen plants, electrolyzers are high-cost, sensitive, and prone to silent failures; existing systems merely log data without diagnosing root causes.
We change that by fusing first-principles physics models, IIoT connectivity, and AI analytics into a unified stack. Our physics-driven(PhysicsOps) digital twins accurately mirror real equipment behavior, predict in-service degradation, and calculate live LCOH, while AI layers (AIOps) detect anomalies and forecast failures.
The outcome is actionable insights that boost uptime, improve efficiency, and cut OPEX by at least 2% annually.
In the long term, we’ll extend our modular, protocol-agnostic platform across power generation, oil & gas, and other heavy-industry sectors.
Startup Description
Both our founders worked together for over five years first at Mitsubishi’s Bangalore office, where Deeparnak was my (Bhavik's) manager on cleantech projects. We collaborated end-to-end on engineering design and global deployment.
Later, we reunited at deep-tech startup Newtrace in its pre-seed phase. Deeparnak led R&D while I (Bhavik) drove engineering, techno-commercial execution, and scaled the product from lab to industrial deployment. We’re also co-inventors on multiple patents.
Working side-by-side in both a large corporate environment and a fast-moving startup built deep mutual trust and explored a major gap in existing industrial-intelligence systems, which inspired us to found Yuji Labs.
Product Screenshots
Business Model
We operate a B2B model with three revenue streams as below :
1. IIoT hardware sales (one-time, ~40% gross margin)
2. Physics-and-AI analytics subscriptions (SaaS, ~65% gross margin)
3. Custom model licensing for bespoke enterprise deployments
After gathering MVP feedback, we plan to offer a profit-sharing tier , tying a portion of our fees to the OPEX savings we deliver, aligning incentives with our customers’ success.
Today we’re focused on green hydrogen and renewables, with expansion into power, oil & gas, and manufacturing on the roadmap. IRENA projects ~80 GW of electrolyzer capacity needed annually by 2030 (19 GW/year in India per NITI Aayog).
A 2% OPEX reduction which translates to 50 Cr./year for 1 GW plant on that base represents a multi-hundred-million-dollar value pool; capturing just 5–10% of that translates to $20–30 M ARR by Year 5.
Customers
based in
Middle East & North Africa, India
Business
Modal Type
SaaS + Physical Hardware
Type of
Sales
B2B
Competitors
Our main competitors are Litmus Automation, Xeeed.io, and Inductive Automation.
Most existing platforms focus on data aggregation and visualization, with a heavy reliance on cloud infrastructure. They treat AI as a bolt-on feature, rather than integrating it with deep domain understanding. In contrast, we build context aware, physics driven intelligence at the edge which is critical for real-time, high stakes industrial decision-making.
While many physics-based models stay stuck in academic research, we bring them into production , mimicking real system behavior to generate insights that are both accurate and actionable. We don’t just reference research , we make it usable.
Teams
Bhavik Modi
Co-founder, CEO
Deeparnak Bhowmick
Co-founder, CTO
Tell us a bit about how founding team knows
each other.
We’ve worked together for over five years first at Mitsubishi’s Bangalore office, where Deeparnak was my (Bhavik's) manager on cleantech projects. We collaborated end-to-end on engineering design and global deployment.
Later, we reunited at deep-tech startup Newtrace in its pre-seed phase. Deeparnak led R&D while I drove engineering, techno-commercial execution, and scaled the product from lab to industrial deployment. We’re also co-inventors on multiple patents.
Why did you decide to start this
company?
After a decade of fixing problems others missed, we wanted a system grounded in real physics, not just boilerplate AI , letting machines "talk" and predict issues before they become failures.
Are all the founders full-time on this
project?
Yes
Looking for an
opportunity to invest in companies
like Yuji Labs?
