Tunisian startup NextAV is betting big on a tricky challenge: using low-resolution satellite imagery to detect methane and oil leaks across vast territories — then enhancing it with AI to make it accurate, scalable, and cost-effective. Selected as one of the five startups by Terna for the Innovation Zone Tunisia, NextAV is a young and lean team of seven based in Sousse.It is now closing its first funding round and aiming to scale into Europe and the GCC region.
We spoke with CEO and co-founder Hichem Mokni, who gave us an inside look at how their AI works, why regulations are finally catching up to the tech — and what keeps him cautious about the startup hype.
Hichem, your tech is impressive, but sounds almost too good to be true. You’re turning free, low-res satellite images into data good enough to detect methane leaks? Really?
Hichem Mokni: (smiles) That’s exactly the reaction we get. But yes — it works. Public satellite data is accessible to everyone, but the resolution is too poor for industrial monitoring. So, we trained an AI model on thousands of image pairs — both low-res and high-res versions — to teach it how to artificially enhance new low-res images.
The result? We can scan huge areas, every five days, and deliver reliable environmental analytics — at a fraction of the cost of commercial satellite imagery.
So instead of paying for expensive high-res images, your clients can rely on you — and you only pay for premium imagery if there’s something suspicious?
Exactly. We provide what we call high-scale monitoring. Our AI-enhanced images are precise enough to flag anomalies: methane leaks, oil spills, heat signatures, etc. When unusual is detected, we can trigger follow-up inspections — either with commercial satellite data or drone operators on-site.
Scanning an entire country weekly with commercial satellites would be absurdly expensive. Even major oil and gas players typically do it once per year. We’re making it possible to do it 12–50x more often.
Methane is clearly your focus — because of regulations or because it’s the most valuable use case?
Both, actually. We didn’t start with methane. We started with super-resolution as a core technology. Methane just turned out to be the perfect use case: high impact, underserved, and increasingly regulated.
The new EU regulation, set to enter into force between 2025 and 2030, changes everything. It doesn’t just affect emissions in Europe — it also demands accountability from upstream producers abroad. If Italy is importing gas from Algeria, for example, those emissions need to be measured and reported. Otherwise, the product can’t enter the EU market.
Sounds like perfect timing for a Tunisian startup bridging North Africa and Europe.
Exactly. Being in Tunisia helps us be cost-effective while also strategically placed. We know the territory, the infrastructure, and the language. We can offer a compliance tool for European buyers and a monitoring solution for producers in North Africa — and do it much cheaper than companies based in the US or EU.
What’s under the hood of your AI model? Is it a black box or can you explain it?
It’s a deep learning model, trained with terabytes of satellite data, using thousands of matched image pairs. During training, we use both commercial and public datasets. Once the model is trained, it can take only the low-res image and generate a super-resolved version with high accuracy.
When we enter a new region, we run an initialization phase: we buy high-res images once, fine-tune the model for that geography, and from then on, we monitor using only low-res data.
So you’re not competing with commercial satellite providers — you’re using them strategically?
Right — we’re not replacing them. We reduce the cost and frequency barrier. We scan regularly with public data, and only use commercial sources when we need to validate or localize an anomaly. It’s complementary.
Even in mining — our first paying customers are in that sector — we follow the same logic: scan regularly with AI, then confirm with drones or higher-res data.
You’re raising your first funding round — how’s that going?
We’re closing a $400K pre-seed round, and we’re in the final phase of signing. It’s been intense, but we’ve stayed lean and focused. The round will help us grow our sales capacity in Europe, consolidate our presence in the GCC region, and launch a generic version of our super-resolution tech — not just for methane.
Let’s talk growth. You’re seven people. Is that enough?
For now, yes. We’re engineers — we build before we talk. But we know we need to grow. That’s why programs like Innovation Zone Tunisia are helpful: they open doors to Europe, give us our first Italian reference, and introduce us to the right partners.
Do you think you’ve waited too long to invest in communication?
(laughs) Possibly. But we’ve been cautious — maybe too cautious — about not overselling. We see startups that spend more time posting on LinkedIn than building the product. We didn’t want to become one of those.
Final question — what would you say to someone skeptical of AI-based solutions like yours?
We get that. It sounds like science fiction. But we’ve shown real, measurable results. Our clients aren’t buying “AI hype” — they’re buying operational visibility and regulatory compliance.
We didn’t invent satellite data. We just found a way to make it work harder, for more people, more often.