Drones in commercial settings are certainly no longer a novelty; they've become part of how serious sites already run.
Which is why the interesting question has changed. It's no longer whether you can fly a drone, because almost anyone can learn that now. The real question is whether you can run a fleet of drones across live sites, safely and reliably, every single day, and turn what they capture into something a business can actually use. That's a far harder problem, and it's the one we've spent 14 years on at Sphere.
The flight is the visible part. The system around it is the product.
Most of the industry will sell you one piece of this and leave you to stitch the rest together. We took the opposite approach and built the whole thing ourselves, here in Australia, so that the hardware, the software, the people and the approvals all come from one company that's accountable for the lot.
It comes down to four integrated layers that have to work as one:
Put those four together under one roof and you get something you can't buy off a shelf: a single point of accountability for the whole programme, instead of a stack of independent vendors pointing at each other when something goes wrong.
Here's what that looks like in practice, running at scale, all drawn from our own deployments.
We've processed more than 113,000 flights through our platform, and we currently have over 500 drones under management for clients across mining, utilities and government.
In the last two years of managed hub operations alone, we've put 17 units in the field across more than 3,300 cumulative operating days, flown close to 4,000 flights under direct control from the operations centre, and covered nearly 15,000 kilometres.
Fleet-wide uptime sits at 99.5%, the figure we commit to in contract and the one we actually deliver. And in the entire history of the hub fleet, there have been zero serious aviation safety occurrences and zero notifications to the national transport safety regulator.
The number I keep coming back to, though, is a response time. At one of our mining sites an unplanned inspection request came in at 1:17pm; the drone was airborne eight minutes later, and the data was in the customer's hands by 1:52pm, with no on-site mobilisation, no drive-out and no crew sent across the pit.
Thirty-five minutes from request to data delivered. That's the value the system actually gives you, and it isn't a flight, it's the ability to answer a question in minutes when the old way took days.
Flying beyond visual line of sight, over distance, at night and near people is where most operators stop, because it's hard, heavily regulated, and impossible to fake your way through the audit.
We're approved and trusted by CASA to run broad-area BVLOS operations, and we're one of only a handful of operators in the country cleared to self-assess our own areas of operation, which last year took us from a 291-day approval queue to being operational in days. That's the difference between a one-off flight and an ongoing programme, like the broad-area BVLOS operations we run for Yancoal.
On top of the regulator's approval, we hold BARS Gold, the highest accreditation in the Flight Safety Foundation's program and one of the most rigorous aviation safety standards in the world. Gold takes three consecutive years of independent, comprehensive audits to earn, which means it can't be bought or fast-tracked. It's a track record, checked by someone other than us.
And none of those flights are left to run themselves. There's always a pilot in our operations centre holding the command link, ready to take over the moment something looks off. Full autonomy might be the end goal for some, but it's that human oversight and the systems around it that makes the semi-autonomy safe enough to trust on a live mine site.
A single inspection can capture more than 8,000 images across 240 kilometres of pipeline in one morning, and a single drone will generate five to eight terabytes of imagery in a year. Once you multiply that across a whole fleet, the hard part stops being the flying and becomes the distance between flying the drone and actually knowing what needs fixing. That distance is where the value now sits.
It's also where we're putting our newest work. We're using our own CuroAI and computer vision models, trained on real site imagery rather than a generic dataset from another country, to close that gap automatically.
One model already in use spots leaks along dewatering pipelines and returns geolocated, confidence-scored findings inside the same workflow that requested the inspection, so the answer comes back without a review team having to scale up to match the data. It gets sharper with every inspection cycle, and there's a lot more coming on this front soon. The short version is that as inspection becomes cheap and frequent, the data is what becomes valuable, not the flight that captured it.
There's a deliberate reason all of this sits in-house. We design and build the hardware in Australia, write the software that runs on it, and fly every mission with our own remote pilots based in Sydney.
That isn't about waving the flag; it's about control. When something goes wrong in the field, we can trace it back to the assembly step, the build record and the technician who signed it off. When a customer asks where their data lives, we can point to Australian infrastructure that we run and control ourselves. And when the roadmap needs to move, we move it, rather than waiting on a vendor in another time zone to agree.
Our hubs are built by hand on one factory floor in our Sydney HQ, by a team that covers the full span of the work in one place: fabricators, assemblers, embedded engineers, flight-ops crew and production managers, all under the same roof. That's what Australian sovereign capability actually looks like in 2026.
The drone in the sky is the smallest and most visible part of what we do, and it's also the part everyone fixates on. The work that genuinely matters is happening somewhere else: in the operations centre, in the compliance system, in the data platform, and in the years of audited approvals that let any of it leave the ground in the first place.
Anyone can put a drone in the air. The system around the flight is the product, and we're only just getting started.
