The Wait is Over: Lumilinks is the 2025 Snowflake Startup Challenge Winner

Eight months. Over one thousand submissions from more than one hundred countries. Ten semi-finalists. Three finalists. Seven heart-pounding minutes as the judges deliberated. And finally, one winner: we are thrilled to announce that Lumilinks is the 2025 Snowflake Startup Challenge Winner!

The judges zeroed in on the potential of several aspects of Lumilinks’ product and business strategy, including its focus on solving business users’ problems and finding impact with more conventional businesses, like utilities and logistics companies, that some would assume to be AI skeptics and less tech-forward with their investments.

“The vision of Snowflake has always been about democratizing access to data; Snowflake is the place where data does more,” said Denise Persson, CMO of Snowflake and a Startup Challenge judge. “And what Lumilinks does is bring that access to data and AI to organizations that would normally not be able to do all this themselves.”

Illuminating the missing links in your data

Based in London, Lumilinks builds “market-ready” AI applications on Snowflake, including its flagship FleetSense AI app that helps commercial fleets keep vehicles on the road through predictive analytics. Instead of just identifying problems, however, Lumilinks develops apps to address a specific business issue and produce tangible value. 

“Data and applications isn’t about the technical audience — it’s about the business user,” said Gary Cole, CEO and Co-founder of Lumilinks. “Our goal is to find a business user with a problem and solve it” while delivering realized business impact and measurable ROI.

It’s a lesson that the company learned early on, and one its management team built into Lumilinks’ three-part strategy:

• Create tangible business value from data

• Automate delivery of that value using Snowflake and its full stack of tools

• Scale those automations into Snowflake Marketplace so they can reach a wider audience without additional overhead on Lumilinks’ side

FleetSense AI, a Snowflake Native App, is a direct result of this strategy. Developed through work with several large utility and logistics providers in the UK, FleetSense AI addresses a common problem that affects any company with a commercial fleet: vehicle offroad time. For these organizations, every day a vehicle is not on the road affects their bottom line.

To address this ongoing issue, “we have built ML and AI algorithms that predict failure in parts in vehicles, that run natively in Snowflake and use telematics data and historic repair data to help business users keep those vehicles on the road with preventative maintenance,” said Seb Bennett, Data Science Lead at Lumilinks. Under the hood, Lumilinks uses Snowflake Cortex, Document AI, Secure Data Sharing, Snowpark Container Services and UDFs — and, of course, the Snowflake Native App Framework. 

On average, FleetSense AI customers see a 30% reduction in vehicle offroad time, “which is an impact in the seven figures” according to Bennett.

Importantly, FleetSense AI brings together fleet managers, vehicle operators, procurement teams, finance teams and C-level execs, each of whom need different outcomes from the same data — something traditional fleet management systems can’t easily deliver. It also means that Lumilinks needs to build applications that have high impact, are easy for business users to adopt and make it clear how the customer’s AI efforts are paying off. 

“Everyone wants to show that they’re investing in AI, and sometimes that’s really hard to do,” said Cole. “Every Lumilinks project includes a value report showing the customer “this is what you invested, this is what we believe you can achieve, and this is what you’re achieving, and we keep that up to date.”

Scalability and flexibility are also part of the recipe. With Snowflake as the foundation for FleetSense AI, Lumilinks was able to tap into the same data and develop two related AI apps within six months: Smart Insights, which gives C-level and exec-level users live reporting and BI dashboards natively in Streamlit, and Smart Repair, to help engineers and service/maintenance managers book repairs with the right agents at the right price. 

The same adaptability and willingness to take on a new challenge is behind Lumilinks’ next offering, an entity and data quality app “that’s going to be our first self-service app in the Snowflake Marketplace,” said Cole. “That’s another multi-billion pound industry that we’ve been able to enter because we’ve got clients, we’re solving the problem and then we’re working with them to say ‘How do we scale this?’”

Recognizing the runners-up: DeepTempo and Winning Variant

DeepTempo and Winning Variant made strong showings in the finale, and each will be considered for up to $250,000 in potential investment from Snowflake. 

DeepTempo is focused on applying AI to bolster defenses against evolving cyber threats. Its Snowflake Native App, Tempo, uses DeepTempo’s Log Language Models (LogLMs) to address customer security pain points and reduce costs by reducing the volume of raw logs that are sent downstream to security information and event management (SIEMs). Tempo helps customers identify security incidents and analyze their scope and severity; the LogLMs help improve accuracy and ability to adapt to new environments as they are exposed to new data distributions. The app was pretrained using enormous quantities of security logs and is particularly focused on the pattern of events, including relative and absolute time.

Winning Variant offers a Snowflake-native experimentation platform as a Snowflake Native App, allowing customers to run innovative experiments directly inside the AI Data Cloud. Experimenting with new features, pipelines and product designs is key to maintaining agility and being able to keep up with changing customer preferences and market trends. With Winning Variant’s Native Split Testing app, teams can deploy and manage experiments using data available in Snowflake in real time, without having to deal with access to a third-party platform, exfiltrating sensitive conversion data or building complex data pipelines to get the data they need.

Join the Powered by Snowflake Startup Program

If you’re a startup building on Snowflake, be sure to check out the Powered by Snowflake Startup Program that provides startups with resources to build and grow data-intensive applications and products. It’s also never too soon to start thinking about next year: Complete this form to get notified when the 2026 Snowflake Startup Challenge registration opens.