Data usage and carriers for IRL streaming
What a month of real streaming actually costs in data, the hotspot trap that kills plans, and how to pick networks when no carrier is best everywhere.
The short version
Roughly 10–15 GB for a five-hour session — but that is the total across every bonded device, not what each line uses. Split three ways it is more like 3–5 GB per line.
There is no best carrier. Each one prioritises coverage in different places, which is the whole reason bonding works.
Watch the hotspot distinction. Depending on how you connect a device, its traffic may come out of a hotspot allowance that is a fraction of the size of its data allowance.
Real numbers from a real month
Most answers to this question are estimates built from bitrate maths. These are not. This is what actually came off three phones over one month of streaming.
- Streaming time
- 109 hours
- Sessions
- 27 (averaging about 4 hours each)
- Devices bonded
- 3 phones
- Total data used
- ~300 GB — about 100 GB on each phone
- Per hour, all links
- ~2.75 GB
- Per hour, per line
- ~0.9 GB
That works out to an average of roughly 6 Mbps of actual throughput across the bonded connection, which is a normal bitrate for a good-looking IRL stream. A five-hour session at that rate lands between 10 and 15 GB.
We considered building one. The problem is that anything precise enough to be worth using would need your bitrate, your codec, how aggressive your adaptive bitrate gets in bad coverage, and how much retransmission your connections need — and that last one is unknowable in advance, because it depends on where you walked.
A calculator would produce a confident number that was wrong. The measured figures above are more honest, and easier to sanity-check against your own first month.
Total versus per line
This is the distinction that causes the most planning mistakes, because you buy data per line but you consume it as a pool.
When someone says a five-hour stream uses 15 GB, that is 15 GB spread across every device carrying the stream. On three phones, each line sees roughly a third of it. So the question "will 50 GB be enough" has no answer until you say how many lines you are splitting it over.
| Stream length | Total across all links | Per line on 2 phones | Per line on 3 phones | Per line on 4 phones |
|---|---|---|---|---|
| 1 hour | ~2.75 GB | ~1.4 GB | ~0.9 GB | ~0.7 GB |
| 3 hours | ~8 GB | ~4 GB | ~2.75 GB | ~2 GB |
| 5 hours | ~14 GB | ~7 GB | ~4.5 GB | ~3.5 GB |
| 20 hours/month | ~55 GB | ~28 GB | ~18 GB | ~14 GB |
| 100 hours/month | ~275 GB | ~138 GB | ~92 GB | ~69 GB |
These are averages from one setup, not predictions. Higher bitrate, poor coverage, or heavy retransmission will push them up.
Device data or hotspot data
This one is worth understanding properly, because it can quietly decide whether a plan works. There are two ways to bring another device into a bonded stream, and they draw from different buckets.
The device runs the app
If the second phone has the IRL Bonded app installed, you can enable it to join as a bonded device. Its traffic then goes out over that phone's own connection and counts against its regular device data allowance — the big number on the plan.
The device does not run the app
If it does not have the app, the streaming phone can instead connect out over a Wi-Fi connection, which in practice usually means a hotspot shared from another phone. That traffic counts against the hotspot allowance of the phone sharing it.
Plans marketed as unlimited routinely include a hotspot cap that is a small fraction of the device data allowance — and once it runs out, hotspot speeds are usually cut hard enough to be useless for streaming, even though the phone itself still has plenty of data left.
If you are relying on hotspot mode, check that number specifically before a long stream. It is a completely different figure from the one on the front of the plan.
Where you have the choice, running the app on each bonded phone is the better option — you get the full device allowance on every line rather than several small hotspot buckets.
What actually drives usage
Three things move the number, and you control two of them.
- Bitrate, which is linear. Dropping from 6 Mbps to 4 Mbps cuts a third of your data. This is the biggest lever you have, and on a phone screen the difference is smaller than you would expect.
- Codec. HEVC gets comparable quality at a lower bitrate than H.264, so the same-looking stream costs less data.
- Retransmission, which you do not control. When packets are lost they get sent again, so actual usage runs somewhat above your encoder bitrate. Rough coverage costs more data than good coverage for identical footage.
There is no best carrier
Carrier advertising is built around convincing you one network is best. It is not true in any way that helps you, because coverage is a local question and the marketing is a national average.
Each network has made different decisions about where to prioritise capacity. One is genuinely better along your usual route. A different one is better two towns over. Neither fact tells you anything useful about the other place, and no national ranking will capture the specific street you walk down every week.
You cannot know in advance which network will be strongest at any given corner. So rather than betting on one, spread the stream across several and let whichever is doing best at that moment carry more of it. The point is not that bonding gives you more bandwidth — it is that it stops any single carrier's weak spot from being your weak spot.
Which also means putting three lines on the same network largely defeats the purpose. Three SIMs on one carrier share one dead zone.
Priority, and why deprioritised is worse than it sounds
Most plans include an amount of premium data, after which you are deprioritised. People read that as being throttled, and it is not quite the same thing — it is worse in a specific way.
Deprioritised usually means you are not slowed down constantly. You are moved to the back of the queue, and you only feel it when the tower is congested. That sounds mild until you think about when towers get congested: crowded places, events, busy times of day. The exact conditions that make somewhere worth streaming from are the conditions where deprioritised data stops working.
So the premium data threshold matters more for streaming than for ordinary phone use. You are not trying to avoid a slowdown on your couch at 3am. You are trying to still have a usable uplink in a crowd.
Upload is what matters, and nobody sells it
Every plan, every coverage map, and every speed-test screenshot is about download. You need upload, which is a fraction of the download figure and is the first thing to degrade when a tower gets busy.
There is a structural reason for this. The overwhelming majority of mobile users consume far more than they send — video, browsing, social feeds are all download-heavy — so networks are built and tuned around that. Upload capacity gets the smaller share of the allocation. As a streamer you are using the network in an unusual direction, and the infrastructure is not optimised for you.
The practical consequence: a coverage map showing strong service tells you very little about whether you can hold 6 Mbps up from that spot. Test your actual routes rather than trusting the map.
Budget carriers
Smaller carriers resell access to the big networks, often at a fraction of the price. They run on the same towers, so coverage maps look identical — but their customers generally sit at a lower priority than the host network's own.
That makes them a genuinely good way to start. If you are streaming a handful of times a month and not pushing high bitrates, a cheap line on a major network is fine, and it is a sensible way to add a third or fourth link without tripling your bill.
They stop being a good option once you are streaming seriously. Low priority means the network deprioritises you before its own customers, so your connection degrades first in exactly the busy conditions where you needed it most. Fine to begin with, not what you want to be relying on long term.
What to actually buy
If you stream a lot
Lines on the major networks — in the US that means Verizon, AT&T, and T-Mobile — on their premium plans, with the highest priority data tier and no slowdown threshold you are likely to hit. Spread across different carriers, not several lines on one. Check the hotspot allowance separately if you plan to use hotspot mode.
If you stream occasionally
Roughly five sessions a month, moderate bitrate, and you can comfortably use lower-priority or budget plans without exhausting anything. Spend the difference on a third line instead — diversity across networks will do more for your stream than premium priority on two.
Three premium unlimited lines with high-priority data will run you a couple of hundred dollars a month. That is more than most streaming services, more than most software, and quite possibly the largest recurring cost in your setup.
It is worth knowing that before you build around it rather than after. If that number is not workable, fewer lines on better plans beats more lines on plans that get deprioritised the moment you walk into a crowd.
The carrier names above are US-specific, but the reasoning is not. Everywhere, the useful approach is the same: spread your links across genuinely different networks rather than one, buy the highest priority tier you can justify, check the upload rather than the download, and test your actual routes instead of trusting a coverage map.
Common questions
How much data does IRL streaming use?
Across 109 hours of bonded streaming in one month on three phones, total usage came to roughly 300 GB — about 2.75 GB per hour across all three lines, or around 0.9 GB per hour per line.
A five-hour session typically lands between 10 and 15 GB in total across every bonded device, so roughly 3–5 GB on each line. Your numbers will differ with bitrate, codec, and how much retransmission your coverage demands.
Does bonded streaming use hotspot data or regular data?
It depends how the device is connected. A phone running the IRL Bonded app and joining as a bonded device uses its own device data allowance.
If the device does not have the app and the streaming phone connects to it over Wi-Fi instead, that traffic comes out of the hotspot allowance on the phone sharing the connection — usually a much smaller number, and worth checking before a long stream.
Which carrier is best for IRL streaming?
There isn't one, despite the advertising. Each network prioritises coverage in different places, so whichever wins on your route may lose somewhere else entirely.
That is precisely why bonding exists: instead of betting on one network being good everywhere, you spread the stream across several and use whichever is strongest where you are standing.
What is deprioritized data and why does it matter?
Past your plan's premium data threshold you're generally not slowed all the time — you're moved to the back of the queue and slowed only when a tower is congested.
Which means you get slowed specifically in busy places at busy times. That's exactly where and when the streams worth doing tend to happen.
Is it worth putting all my lines on one carrier for a discount?
Usually not. Multi-line discounts are real, but three lines on one network share the same dead zones, the same congested towers, and the same outage.
Diversity across networks is the thing that makes bonding work. A cheaper bill on a setup that drops in the same places is a poor trade.
Carrier plan names, thresholds, and hotspot allowances change frequently. The measured usage figures on this page hold up over time; the specific plan advice is worth re-checking against current offerings before you buy.