If you arrived here from the decision guide, you are in the right place. Live fleet video vs recorded worked out which fleet use cases actually need a live feed. This is its companion: what that decision costs to run, line by line.
One live camera per vehicle, streamed ten hours a day, 26 days a month, at the ~600 kbps one current fleet dashcam documents for its live view. That is 70 GB of mobile data per vehicle per month. Across 200 vehicles, it is 14 TB — and 3.12 million streaming minutes.
The same fleet running live video only when something happens — say ten events per vehicle a month, ten minutes each — uses under half a gigabyte per vehicle. Roughly 150 times less.
That gap is the whole of fleet video monitoring cost. The decision guide concluded that most fleets need live on trigger, not live always. This post prices both, so the business case rests on numbers rather than on how impressive a wall of live feeds looks.
The three lines of a live monitoring budget
Every live fleet video programme pays for the same three things, whoever supplies it.
- Data — the mobile data each vehicle uses to send live video.
- People — the operators who watch and respond. A live feed nobody is watching is a recording with a higher bill.
- Platform — the service that carries the live video to viewers and records it, usually billed by the minute, the viewer or the gigabyte.
The mix changes completely between the two designs, and the most expensive line is often not the one buyers look at first.
Line 1: mobile data
Per camera, the arithmetic is fixed by the stream. At 600 kbps, one live feed is about 0.27 GB per hour; the reasoning behind that number, and why the recording stream never goes live, is in how many camera feeds fit down a moving 4G link.
Design (per vehicle, per month) | Live hours | Mobile data |
|---|---|---|
Always-on, 1 camera, 10 h × 26 days | 260 | 70.2 GB |
Always-on, 4 cameras | 1,040 camera-hours | 280.8 GB |
Live on trigger, 10 events × 10 min | 1.7 | 0.45 GB |
We deliberately leave a price per GB out of this table. Retail mobile data prices differ by an order of magnitude between countries, and fleet M2M plans are negotiated contracts that are not published. Multiply the data column by your own contracted rate — the rest of the argument does not depend on it.
The finding is not the price. It is the ratio: always-on moves roughly 150 times the data of live on trigger, the platform line scales with the same ratio, and staffing follows it.
Line 2: the people who watch
This is the line that decides the business case, and it is the easiest one to leave out of it.
Always-on live video only has value if someone is watching it. Covering one seat around the clock takes about 4.2 full-time staff (168 hours a week divided by a 40-hour week), before leave, training or supervision. And one person cannot meaningfully watch many feeds at once — the operator-to-camera ratio is the constraint that decides whether proactive monitoring pays in fixed-site surveillance, and it transfers directly to a fleet. If an operator can genuinely watch N feeds, a 200-vehicle always-on fleet needs 200 ÷ N seats per shift, each costing 4.2 people.
Live on trigger inverts the model. Operators are sized to events, not feeds. Two hundred vehicles with ten events each a month is 2,000 events; at five minutes of handling each, that is about 167 operator-hours a month — roughly one full-time person's working month (about 173 hours). In practice you still staff for peaks and round-the-clock cover, but the scaling variable is how often something happens, not how many cameras exist.
The trigger has to come from somewhere, and that is where edge AI in the vehicle and the cloud layer earn their place: deciding which moments deserve a human is what keeps the people line small.
Line 3: the platform
Whatever carries live video from vehicle to operator bills on some combination of minutes, viewers and gigabytes. The volume is what drives it.
200-vehicle fleet, per month | Always-on, 1 camera | Live on trigger |
|---|---|---|
Publisher minutes | 3,120,000 | 20,000 |
Live data from vehicles | 14 TB | 90 GB |
Three things turn that volume into a bill. Per-minute providers price exactly the first row, and what enterprise teams actually pay per minute at scale covers how quickly those minutes compound. If you run the media servers yourself, cost follows the viewers rather than the vehicles. And every live minute is usually recorded too, so the retention bill grows with it — a question for the retention policy rather than the live budget.
We deliberately quote no provider rates here. The point is the multiplier: a 150-fold difference in volume, 3.12 million minutes against 20,000, dwarfs any realistic difference between rate cards.
What the budget buys, honestly
A business case needs the other side of the ledger, and here the honest answer is thinner than pitches tend to suggest.
The use cases where live video changes outcomes — panic alarms, assaults, hijacks, thefts in progress — involve real money. Cargo crime in the US and Canada averaged $273,990 per incident in 2025. But we could not find any published, non-vendor measurement of how often live monitoring actually changes the outcome of those incidents. Anyone quoting a prevented-loss percentage should be asked where it comes from.
What is documented is that the safety gains from fleet cameras come from recorded footage and coaching, which neither design has to pay live rates for. So the defensible business case is narrow: live on trigger buys the ability to respond to rare, expensive, time-critical events, at a data and staffing cost that scales with those events. Always-on buys the same ability at roughly 150 times the volume, plus a staffing model that grows with every camera fitted.
One more cost belongs in the model even though it has no line in the table: live only works where there is signal, and the feed dies in dead zones. Budget for the coverage your routes actually have, not the coverage the contract implies.
The Bottom Line
Fleet video monitoring cost is driven by one design decision. Always-on live video uses about 150 times the data and minutes of live on trigger, and it forces staffing to scale with cameras rather than with events.
Price live on trigger for the use cases that need it, keep everything else recorded, and treat any claimed return on always-on live video as unproven until someone shows the data.
What's Next
This is the cost half of a pair; its companion, the decision guide linked at the top, sets the scope priced here. For the other half of the fleet video picture — what happens to footage when the live link is gone — store-and-forward for fleet video picks up where the live feed stops.
Frequently Asked Questions
How much does live fleet video monitoring cost?
It depends mostly on design. One always-on live camera per vehicle uses about 70 GB of mobile data a month, while live on trigger uses under half a gigabyte for ten ten-minute events. Operator staffing and platform minutes scale with the same ratio, so the design choice outweighs any individual price.
How much mobile data does a live fleet camera use?
At a typical live-view bitrate of around 600 kbps, about 0.27 GB per hour per camera. Streaming one camera ten hours a day for 26 days is roughly 70 GB per vehicle per month; four cameras is about 280 GB.
How many operators does live fleet monitoring need?
For always-on video, one round-the-clock seat needs about 4.2 full-time staff, and seats scale with the number of feeds each operator can genuinely watch. For live on trigger, staffing scales with events: 2,000 events a month at five minutes each is about 167 operator-hours, roughly one person's working month, before peak cover.
Is always-on live fleet video worth the cost?
For most fleets, no. The documented safety gains come from recorded footage and coaching, and no published non-vendor study measures how often live monitoring changes incident outcomes. Live on trigger captures the time-critical use cases at a fraction of the data, minutes and staffing.
What does a fleet video platform charge for?
Usually some mix of streaming minutes, viewers and data, plus recording storage. Samvyo, based on SFU architecture, is offered with flat licensing rather than per-minute billing, which changes how the always-on versus on-trigger volume difference shows up in the bill — worth modelling against your own event rates.
How can a fleet reduce live video monitoring costs?
Stream live only on triggers such as panic buttons, collision sensors or AI alerts; show operators events rather than walls of feeds; and keep coaching, claims and audits on recorded footage. Keeping the media path, relays and recording on infrastructure you control, as Samvyo allows, also puts the retention side of the bill under your own policy.
