The Singapore company teaching heavy machinery to predict its own failures - before the line ever stops.
Most factories still run on a simple, expensive rule: fix the machine after it breaks. A pump seizes, a conveyor stalls, a ship idles in port, and only then does someone reach for a wrench. Groundup.ai, a deeptech company working out of a workshop in Kaki Bukit, Singapore, was built on the argument that this order of operations is backwards - and that the machines themselves have been trying to warn us all along.
The company's product clips a small, magnetic-based sensor onto a piece of industrial equipment. From there it listens: to vibration, to temperature, to sound. Those three streams feed an AI platform the company calls GINA, which learns what a healthy machine looks and sounds like and flags the subtle drift - a new tremor, a hotter bearing, an unfamiliar hum - that precedes a failure by days or weeks. Groundup.ai brands the whole approach "Cognitive Maintenance," and its tagline states the thesis plainly: the best fix is the one you never have to make.
That framing has attracted real money and real contracts. The company raised a roughly US$1.8M seed round led by Wavemaker Partners, followed by a US$4.25M Series A in April 2025 led by Singapore's Tin Men Capital. In early 2026 it announced the largest deal in its history - a multi-site Cognitive Maintenance contract worth more than US$10M. For a company of around 34 people, those are outsized numbers, and they point to a market that is far larger and far less glamorous than most of what gets funded in AI.
We're not just optimising maintenance; we're transforming the way industrial operations run.
What makes Groundup.ai worth a closer look is not that it detects anomalies - plenty of tools do that. It is where the company draws the line between an alert and an answer, and how deliberately it has engineered away the friction that usually kills industrial-tech deployments. Below, a field guide to what the company does, who buys it, and where it sits in a crowded market.
The pipeline is deliberately short - the harder the deployment, the fewer factories say yes. Groundup.ai's design collapses the path from a bolted-on sensor to a decision an engineer can act on.
The sensor snaps onto a machine magnetically - no drilling, no downtime, no compromise to structural integrity.
It continuously records vibration, temperature and sound, even in harsh environments, on extended battery life.
The AI platform detects anomalies against learned baselines and its asset library, then diagnoses likely causes.
Agentic AI surfaces root cause and recommended remedies, so teams fix the problem instead of chasing it.
A failing machine rarely goes quiet. It changes how it moves, how hot it runs, and how it sounds. Groundup.ai's sensor treats each as an independent early-warning channel - and reads them together for a fuller picture of machine health.
Detects imbalance, misalignment and bearing wear through changes in a machine's mechanical tremor.
Flags friction, overload and cooling issues as components begin to run hotter than their baseline.
Uses machine acoustics to catch internal wear and developing faults the human ear would miss.
Groundup.ai sells hardware and software as one system. The sensors are the eyes and ears; GINA is the brain; Cognitive Maintenance is the promise that ties them together. Around that core sits an asset library that speeds deployment and a training academy that teaches engineers to work alongside the AI rather than around it.
Agentic AI agents, proprietary sensors and deep machine understanding that deliver root-cause insight and remedies - billed as the world's first agentic AI for maintenance.
The AI platform and interface that analyzes real-time signals, detects anomalies and recommends actions to maintenance teams.
Magnetic-base sensors capturing vibration, temperature and sound - non-intrusive, retrofit-ready and built for harsh environments.
Condition-based monitoring software for real-time machine health, anomaly detection and downtime prevention.
A growing library of asset and fault profiles that accelerates deployment and sharpens diagnostic accuracy across machine types.
Training on human-AI synergy - teaching engineers to interpret live signals and make Cognitive Maintenance decisions.
Asset-heavy operators - multinationals and government-linked entities across Southeast Asia and the Middle East - in manufacturing, maritime, defense, oil & gas, food & beverage, materials and infrastructure. Public references include a Coca-Cola use case.
Unplanned downtime is the most expensive event on a factory floor or a ship: a stopped line, an idled asset, a missed shipment. Traditional maintenance is either reactive (fix on failure) or wasteful (fix on a fixed schedule). Groundup.ai targets the gap between the two.
Two things: friction and depth. The magnetic sensor retrofits without downtime, and the agentic AI moves past raw alerts to name a root cause and a fix - positioning itself as a co-pilot for engineers, not a replacement.
This isn't just a win for Groundup.ai; it's a massive signal for the future of Industry 5.0.
Groundup.ai runs a B2B model that blends hardware and subscription software: customers buy the sensors and pay for ongoing access to the GINA platform and Cognitive Maintenance service, usually as multi-site enterprise deployments. In Singapore, the Cognitive Maintenance solution is listed as a pre-approved option under the Productivity Solutions Grant (PSG), lowering the barrier for local adopters.
The company sits in the industrial predictive-maintenance and condition-monitoring market, competing with players such as Augury, Samotics, Waites, Petasense and Siemens' Senseye, as well as the incumbent it most often displaces: traditional reactive and scheduled maintenance. Its wedge is a combination of easy retrofit hardware and an AI layer that emphasizes actionable diagnosis over dashboards of raw data.
| Round | Amount | Date | Lead & Investors |
|---|---|---|---|
| Seed | US$1.8M | 2021-2022 | Wavemaker Partners (lead), SEEDS Capital, angels |
| Series A | US$4.25M | Apr 2025 | Tin Men Capital (lead), Wavemaker Ventures, SEEDS Capital, HIVEN |
| Total Disclosed | ~US$6.05M | - | Across seed and Series A |
The Series A is earmarked to expand product innovation and scale market presence across Asia and the Middle East, with planned entry into Australia, Europe and North Asia.
Serial entrepreneur Leon Lim starts Groundup.ai to bring AI to industrial operations.
Alex Wong joins to lead operations; the magnetic sensor and predictive software mature.
Wavemaker Partners leads, with SEEDS Capital and angel investors.
Tin Men Capital leads a round to scale across Asia and the Middle East.
Launches agentic Cognitive Maintenance and closes its largest-ever multi-site deal.
Serial tech entrepreneur with two prior exits in Asia-Pacific. Sets the company's Industry 5.0 vision and led both funding rounds.
Operations and chemical-engineering veteran who frames the mission simply: turn data into decisions.
Founding-team sales engineer helping translate the technology into field deployments for industrial clients.
Product demos, dashboard walkthroughs and interviews from the Groundup.ai channels and training academy.
It prevents unplanned downtime for industrial machinery by pairing proprietary magnetic IoT sensors - capturing vibration, temperature and sound - with an agentic AI platform (GINA) that detects faults early and recommends fixes. The company calls this Cognitive Maintenance.
CEO Leon Lim, a serial entrepreneur with two prior exits, founded the company. It is headquartered at 10 Kaki Bukit Ave 4, Singapore. Alex Wong serves as COO and co-founder.
Around US$6M disclosed - a ~US$1.8M seed round led by Wavemaker Partners and a US$4.25M Series A in April 2025 led by Tin Men Capital.
Asset-heavy sectors: manufacturing, maritime, defense, oil & gas, food & beverage, materials, energy & utilities and infrastructure - primarily large enterprises and government-linked entities across Asia and the Middle East.
Its sensors mount magnetically for fast, non-intrusive retrofits, and its agentic AI goes beyond alerts to provide root-cause diagnoses and recommended actions - positioning AI as a co-pilot for maintenance engineers.