Ask any anaerobic digestion (AD) operator what keeps them up at night, and the answer probably won’t be “data.” But maybe it should be.

In an industry where margins are tight, subsidies uncertain, and feedstock variability the norm—not the exception—the ability to predict biogas production is the difference between thriving and merely surviving. And yet, prediction is still the weakest link in biogas production.

Biogas is produced through one of the most complex biological systems used in industrial settings. It’s multi-phase, multi-input, and intensely sensitive to variables like temperature, retention time, nutrient balance, and microbial health. Operators are expected to run these systems smoothly with limited tools, limited time, and limited team capacity.

Compare this to the petrochemical sector: decisions are driven by real-time models, integrated data layers, and decades of structured analytics. In biogas, we still rely far too heavily on gut feel and “what worked last time.” That’s not a strategy—it’s a risk.

Paradoxically, biogas plants collect plenty of data—SCADA systems, lab reports, feedstock contracts, maintenance logs, market prices. But turning that raw data into something actionable is another story. Most plants don’t have a dedicated analyst, let alone a data science team.

This leaves operators vulnerable. Highly risk-averse decision-making is understandable when your data isn’t helping you see ahead. But the cost is real: overfeeding, underfeeding, unplanned downtime, or missed opportunities in energy trading. At scale, this leads to suboptimal asset performance and over-reliance on subsidies to stay afloat.

At BiofuelAI, a new spinout from the University of Surrey, we believe prediction is no longer a nice-to-have—it’s the next frontier for operational excellence in biogas.
We’re using machine learning and artificial intelligence to build models that learn from your plant’s unique operating conditions, feedstocks, and performance data. These models can predict biogas production, identify early signs of instability, and even suggest feedstock blends or process adjustments before issues arise.

Think of it as an intelligent layer sitting on top of your plant’s existing systems—always learning, always improving, and most importantly, making complexity manageable.

Our goal is simple: empower operators to make better decisions with less effort. Whether it’s a farmer co-digesting manures and food waste, or a utility-scale plant bidding into the energy market, BiofuelAI aims to deliver a suite of predictive tools that link inputs, process, and revenue in real time.

Because when you can see what’s coming, you can prepare, adapt—and thrive.