Fibre download speed: why 1 Gbps only gives you 125 MB/s

Your plan advertises 1 Gbps, your download shows 110 MB/s, and you are wondering where the other nine tenths went. Nowhere: a speed advertised in megabits per second becomes a download speed when you divide it by eight, because one byte contains eight bits. A 1 Gbps plan therefore downloads at best 125 MB/s, and around 110 MB/s once network overhead is deducted. Here is the rule, a conversion table built on the speed tiers actually sold in Belgium, and how long a film, a game or an update really takes. Figures recorded in August 2026.

Mbps or MB/s: what is the difference, and why does my download show eight times less?

Two units coexist, and nobody bothers to point it out. Providers, speed tests and adverts count in bits per second: Mbps, Gbps. The software that actually downloads — your browser, Steam, a console, a backup client — counts in bytes per second: MB/s.

One byte equals eight bits. That is the whole difference, and it hangs on a single capital letter: Mb is the megabit, MB the megabyte. The conversion rule is disarmingly simple.

MB/s = Mbps ÷ 8 — and conversely, Mbps = MB/s × 8.

That one division explains almost every "my fibre is slow" thread on provider forums. A 100 Mbps subscriber seeing 12 MB/s in their download bar does not have a problem: they have exactly what they paid for. To place your plan in context beyond the headline downstream figure, use our ranking of the best fibre internet deals in Belgium.

How many MB/s are 100, 300, 500 Mbps and 1 Gbps?

Here is the conversion applied to the speed tiers actually sold in Belgium in August 2026, rather than to generic theoretical values. The "realistic" column takes off 10 to 15%: that is the share consumed by network protocols, encryption and retransmissions, however good your line is.

One thing jumps out of the second column: above one gigabit, the line stops being the bottleneck. A mechanical hard drive writes at around 100 to 150 MB/s, a gigabit Ethernet port caps at 125 MB/s, and the vast majority of download servers will never push 900 MB/s to a single customer. Buying 8.5 Gbps without a multi-gigabit network card and a fast SSD is like fitting a lorry engine to a bicycle.

How long does it take to download a film, a game or an update?

That is the question that actually matters, and it takes one division to answer: time (in seconds) = file size (in MB) ÷ speed (in MB/s). Here is the result for four volumes everybody runs into, at the speeds sold here. These are theoretical durations: add 10 to 15% for real-world conditions.

The table reads both ways. Going from 100 to 500 Mbps divides the wait on a large game by five: two hours forty becomes half an hour, which genuinely changes an evening. Going from 1 to 2 Gbps saves eight minutes on that same game, and twelve seconds on a TV episode — a gain nobody notices day to day.

One field observation, valid everywhere: the source caps you more often than your line does. Console update servers, especially at peak hours, rarely push more than 30 to 60 MB/s to any one subscriber. If you measure 110 MB/s on a speed test and 40 MB/s on a PlayStation download, the anomaly is not at your end.

What counts as a good download speed in 2026?

There is no good number in the abstract, but there are thresholds. Below 25 Mbps (about 3 MB/s), a modern household starts to feel the limit: one 4K stream and one video call at the same time no longer fit comfortably. Between 100 and 300 Mbps, virtually every domestic use is covered, including with several people at once. Above 500 Mbps, you are no longer buying smoothness but raw download speed.

The consumption figures make this plain. Netflix states roughly 7 GB per hour in 4K Ultra HD, 3 GB in HD, 700 MB in SD and 300 MB on its data-saver setting, according to its official help centre. Seven gigabytes an hour is an average throughput of about 16 Mbps. In other words, four screens streaming 4K simultaneously stay under 70 Mbps — a fifth of a 300 Mbps plan.

Is 300, 400 or 500 Mbps enough for your household?

Yes, in the overwhelming majority of cases — and the table below gives an order of magnitude rather than a marketing promise. Add up what you use simultaneously at peak time, not your total usage.

The only serious argument above 500 Mbps is download volume, not screen count. If you install a large game every week, the gigabit earns its keep. Otherwise the money is far better spent on the price once the promotion ends, which weighs vastly more over twelve months than one extra speed tier.

Why is your real download speed lower than your contract?

Four causes stack up, and only one is the provider's doing.

The conversion to bytes already divides the figure by eight — that is not a loss, it is a change of unit. Protocol overhead (TCP/IP headers, acknowledgements, TLS encryption) then takes 10 to 15% of raw bandwidth: unavoidable, and it explains the gap between 125 and 110 MB/s on a gigabit. Wi-Fi commonly costs another 20 to 30% compared with an Ethernet cable, and considerably more through two load-bearing walls. Finally, the source and your own hardware: a throttled server, a slow hard drive, a 100 Mbps network port or an ageing router will cap your download regardless of your subscription.

That leaves the fifth cause, the only one worth complaining about: a line persistently below what was sold. The advertised speed is a theoretical maximum, but a large and sustained shortfall, measured over Ethernet at several times of day, is legitimate grounds to contact technical support. The official comparison tool from BIPT, the Belgian telecoms regulator, lets you check the speeds advertised for the offers available at your address, and the Ombudsman Service for Telecommunications handles disputes that a provider has not resolved.

Why does Wi-Fi distort your download speed measurement?

Because what you are measuring is your local network, not your line. Wi-Fi shares a radio channel between all your devices and your neighbours', loses power through walls, and falls back to a slower band as soon as the signal weakens. A test run from a phone upstairs has no diagnostic value whatsoever.

The method that is worth something takes four steps: connect the computer to the router with an Ethernet cable, close background applications, check that the network port supports gigabit, and repeat the test in the evening around 8 pm. That last point is decisive in Belgium: on coaxial cable, capacity is shared between households on the same neighbourhood segment and speeds drop at peak hours, whereas a dedicated FTTH line should stay stable. The gap between your 3 pm and 8 pm readings is the best free indicator of which technology serves you — the architectural difference is set out in our comparison of FTTH fibre and coaxial cable.

What about upload speed — should you look at that too?

Yes, and often more closely than the downstream figure. Upload speed follows exactly the same conversion rule: 50 Mbps of upload is 6.25 MB/s. But it varies far more between plans: on Belgian coaxial cable it caps between 30 and 50 Mbps whatever the advertised download speed, whereas FTTH reaches 200, 500 Mbps or full symmetry. The plan-by-plan detail is in our article on fibre upload speed in Belgium.

Which speeds are actually available where you live?

The contract figure is worthless if the network does not reach your door, and this is where Belgium is at its most uneven. FTTH deployment reaches, as an order of magnitude in mid-2026, roughly 78% of homes in Brussels, 71% in Flanders and 44% in Wallonia — moving figures, to be checked at the address rather than at the municipality.

In concrete terms: in Brussels and the large Flemish cities, 1 and 2 Gbps FTTH tiers are widely available. In Wallonia, a large share of the territory is still served by the coaxial cable of the former VOO network now run by Orange, where gigabit downstream exists but upload is capped at 50 Mbps. And in the least dense areas of all three regions, copper VDSL still tops out around 50 to 100 Mbps.

There is also a Belgian quirk that few comparison sites explain: which provider you can actually subscribe to depends on the wholesale network deployed at your address — Proximus, Fiberklaar or Wyre. Two neighbours on the same street can face different catalogues and different speed tiers. The only reliable answer is an address-level check: our guide to checking fibre availability at your address covers the tools, including the regulator's atlas, and the fibre page from Proximus offers a quick test on the country's largest network.

In short, fibre download speed is not mysterious: it is the contract figure divided by eight, minus what the protocols take, minus what Wi-Fi wastes, and bounded by the server at the far end. Once that arithmetic is settled, choosing the right plan becomes far easier — because the question stops being about the biggest number and starts being about the one you will actually use.