FlightHub 2 Cloud, On-Prem or All-in-One: Picking a Dock Backend

A client in Canada spends about four hours on a set of aggregate measurements. In a recent demo, Jared Westrem and Adam Chartrand showed them a drone dock doing the same job remotely in under an hour. The dock gets the attention, but in this episode of Tap Into Data, the hosts spend most of their time on what buyers tend to forget: the software that runs the dock and where its data lives.

The dock needs a brain

Jared starts with something new dock buyers often miss. A regular drone comes with a handheld controller, usually Android-based, where you plan and fly missions. A dock has no screen. Your controller becomes a web interface you open in a browser, and with DJI, that interface is FlightHub 2.

So once you choose a dock, you also have to decide where FlightHub runs. DJI offers three answers, and the right one usually depends on how strict your IT team is.

Three ways to run FlightHub 2

FlightHub 2 online (DJI cloud): This is where most people start. It's DJI's hosted version, running on AWS in a Virginia data center and managed by DJI. It's the fastest way to get going. For larger companies, though, the fact that the data passes through something DJI owns and manages can still raise flags, even though the servers are in the US.

On-prem: DJI released this around the time Dock 3 came out. You get the same capabilities, but you host FlightHub in your own AWS environment and apply your own security requirements. The catch, Jared says, is a steep learning curve. Not every company buying a dock has the engineers to stand up and maintain its own AWS instance.

All-in-one (AIO): This is DJI's middle option: a Lenovo desktop server that ships with the software already installed. You own the hardware and keep it isolated on your own network, without building anything yourself. The limit is how many devices it supports. Jared puts it at fewer than 20 docks. That covers most organizations. For the operators he works with that run dozens of docks, some close to a hundred, the full on-prem deployment is the way to go.

The usual path, he says, is to start on the DJI cloud, see what automation can do, and then decide whether to move to the all-in-one or a managed on-prem setup. For some teams that can use DJI's hosting, staying on the native cloud may simply be the most cost-effective choice.

The real gain is removing touchpoints

Jared argues that many drone teams think they're already automated because they fly pre-planned missions. But they still build the flight plans by hand, pull SD cards, and manage the files themselves. All of those steps take time. Hand them to the software and, in his words, "you can 10x your productivity easily."

He sees it in how customers buy. Teams that start with one dock to test it often come back within a month asking for three or five more. That's especially true when they pair docks with beyond visual line of sight (BVLOS) or one-to-many operations, where one operator in an office manages docks spread across a state.

FlightHub isn't only for docks, either. Clients with fleets of M400S carrying lidar, M350S, or Matrice 4 series aircraft can still push missions from FlightHub. The person on site just delivers the hardware, and a remote operations center flies it, watches the capture in real time, and makes decisions as the data comes in.

Jared is also honest about a trade-off. DJI faces restrictions in the US, including limits on new hardware, which makes the Canadian market an interesting place to keep up with new dock generations.

Drones aren't the whole picture

Adam brings examples that go well past inspection. In agriculture, a dock could spot disease hotspots and a DJI spray drone could treat just those areas. A condo corporation in London, Ontario asked him about window-washing drones. And after regulatory changes in Canada, which Adam believes came last November, utilities are discussing drones that fly under the corridor to spray non-compatible vegetation instead of just flying over it to look for fall-in risks.

He also mentions a Pittsburgh company that automates inspection below solar arrays, checking bolts, inverters and electrical parts that an aerial flight can't see. Combine the two, and you get both sides of the asset. Then a site manager only goes out once enough issues have piled up, instead of walking every row of panels on a schedule.

Jared links this to years of work on wind blades. The question is how to get a complete view of an asset by combining tools. As he puts it, "drones aren't the end-all, be-all, but it's a tool to get you the information from a new perspective." The value comes from pairing aerial data with ground robotics and the monitoring systems you already have.

Data work off the flight line too

Not everything in the episode is about drones. Adam talks about an AI workshop where the group compared an ideal client profile with open bids and turned an RFP package into something close to a full proposal in about an hour, work that used to take weeks. He now gets a daily email digest at 7 a.m. of Canadian RFPs that match his keywords. The longer-term goal is to feed that data, along with the dock inspection data, into a Smartsheet workflow.

The hosts close by introducing their next guest, Zachary Roach from Xcel Energy. His team runs docks and remote operations across a utility that covers wind, solar, battery, nuclear, transmission, distribution, and substations in about eight states.

Press play to hear the full conversation, including Jared's walkthrough of the three FlightHub options.

Key takeaways

• DJI FlightHub 2 comes in three versions: DJI-hosted cloud, self-hosted on-prem in your own AWS, and an all-in-one Lenovo server with the software preinstalled.
• The all-in-one works for fleets under about 20 docks; operators with dozens to roughly a hundred docks are better off with full on-prem.
• Most teams start on DJI's cloud to prove the value, then move to a different setup if their IT requirements call for it.
• The biggest productivity gains come from removing manual steps like building flight plans and pulling SD cards, not just from flying pre-planned missions.
• Drones are one data source among many, and pairing aerial data with below-the-asset and ground data gives the most complete picture of an asset.