Mbps vs MB/s: Why Your Internet Is Slower Than the Number You Pay For
Your provider sells the plan in megabits per second, but your download window counts in megabytes per second, and there are 8 bits in every byte. So a 100 Mbps line tops out near 12.5 MB/s before anything else touches it. Then overhead, Wi-Fi, and congestion shave the rest, which is why the real number lands well below the one on your bill.
Why is my download speed so much lower than my plan?
Because the two numbers use different units. Your plan is measured in megabits, and your download manager reports megabytes, and one byte holds 8 bits. That alone makes the download figure roughly one-eighth of the plan figure. A 100 Mbps connection was never going to show 100 anything in your browser.
How do I convert my plan speed into real megabytes per second?
Take the advertised Mbps, divide by 8 for the theoretical byte ceiling, then trim about 10% for real-world friction. That second step matters because overhead and Wi-Fi never let you hit the raw maximum. Here is the full conversion on a common 300 Mbps plan, start to finish.
- Start with the plan number — 300 Mbps, straight off the bill.
- Divide by 8 — 300 ÷ 8 = 37.5 MB/s, the theoretical maximum byte rate.
- Trim for friction — apply roughly 90% for protocol overhead and line loss: about 33.7 MB/s in practice.
- Turn it into a time — a 5 GB (5,000 MB) file at ~33.7 MB/s takes about 148 seconds, roughly two and a half minutes.
That gap between 37.5 and 33.7 MB/s is not your provider cheating you. It is the tax every connection pays to move data reliably, and it shows up on every plan at every speed.
What steals the rest of the speed after the bits-to-bytes math?
Four things, stacked on top of each other. Protocol overhead eats roughly 5–10%. Wi-Fi can cut a connection in half versus wired. Evening congestion slows shared lines. And the server you are pulling from sets its own ceiling. Each one is small alone; together they explain the missing throughput.
Wi-Fi is the biggest and least obvious drain. The rate printed on a router box is the PHY, or raw radio, figure. Actual usable throughput is far lower. A Wi-Fi 5 router advertising 1,200 Mbps often delivers only 500–600 Mbps of real TCP throughput in good conditions, and worse through walls or on old client hardware. A wired connection sidesteps most of that.
How much of my advertised speed should I actually get?
Somewhere between 60% and 90% of the plan figure is normal once you account for everything. The FCC's Measuring Broadband America program found many major providers deliver at or above their advertised download speed at peak, and others land at 86–90%. A gap under 30% is routine; a gap over 50% points to a real fault.
| Plan (Mbps) | Max byte rate (÷8) | 5 GB file at max |
|---|---|---|
| 25 Mbps | 3.1 MB/s | ~27 min |
| 100 Mbps | 12.5 MB/s | ~6.7 min |
| 300 Mbps | 37.5 MB/s | ~2.2 min |
| 500 Mbps | 62.5 MB/s | ~1.3 min |
| 1,000 Mbps | 125 MB/s | ~40 sec |
Those times are the best case at the raw divide-by-8 rate. In practice a real download runs about 10–30% longer once overhead, Wi-Fi, and the source server take their cut, so treat the table as a floor for the clock, not a promise.
Why is my upload so much slower than my download?
On most home connections, it is slower on purpose. Cable, DSL, and fixed-wireless plans are asymmetric by design, giving download 5 to 10 times more capacity than upload. Providers build them that way because typical users pull far more data than they push. Fiber is the exception and usually runs symmetric.
| Connection type | Download vs upload | Typical ratio |
|---|---|---|
| Cable | Strongly asymmetric | 10:1 or higher |
| DSL | Asymmetric | 5:1 to 10:1 |
| Fixed wireless | Asymmetric | 5:1 to 10:1 |
| Fiber | Usually symmetric | ~1:1 |
If you upload video, back up to the cloud, or run video calls, that ratio matters more than the headline download number. A 400/20 cable plan looks fast until you try to send a large file and hit the 20 Mbps ceiling, which is only about 2.5 MB/s before overhead.
How do I read a speed test correctly?
A speed test reports in megabits per second to match how your plan is sold, so divide the result by 8 before comparing it to any MB/s figure your download shows. Ping is separate: it measures responsiveness in milliseconds, not throughput. Run the test wired, and at different times of day, to separate Wi-Fi and congestion from the line itself.
Frequently asked questions
Is 100 Mbps the same as 100 MB/s?
No. 100 Mbps is 100 megabits per second, and 100 MB/s is 100 megabytes per second, which is 8 times more data. A 100 Mbps plan downloads at a theoretical maximum of 12.5 MB/s, and a bit less after overhead.
What is the difference between Mb and MB?
The lowercase "b" is a bit and the uppercase "B" is a byte, and a byte is 8 bits. So MB is 8 times larger than Mb. Internet plans use bits (Mbps); file sizes and download speeds use bytes (MB, MB/s).
Why does my download show a smaller number than my internet plan?
Two reasons. First, your plan is in megabits and the download is in megabytes, so it is already about one-eighth of the plan number. Second, protocol overhead, Wi-Fi, and network congestion trim another 10 to 40% off the top.
How do I convert Mbps to MB/s?
Divide the Mbps figure by 8. A 500 Mbps plan becomes 62.5 MB/s at the theoretical maximum. To go the other way, multiply MB/s by 8: a 30 MB/s download is running at about 240 Mbps.
Is it normal to get less than my advertised speed?
Yes. Getting 60 to 90% of the advertised figure is typical once you factor in overhead, Wi-Fi, and congestion. A shortfall under 30% is routine. If you consistently see less than half your plan on a wired test, contact your provider.
Why is my upload speed so low?
Most cable, DSL, and fixed-wireless plans are asymmetric, giving upload 5 to 10 times less capacity than download because people download far more than they upload. Fiber is usually the exception and offers matching upload and download speeds.
Does Wi-Fi make my internet slower?
Often, yes. The rate on a router box is the raw radio speed, not usable throughput, and real Wi-Fi performance can be half of that or less through walls and interference. A wired Ethernet connection is more consistent and usually faster.
Once you internalize the divide-by-8 rule, the gap between the plan and the download stops feeling like a scam and starts making sense. To skip the mental math on any file, drop your real speed and the file size into the file download duration estimator and it will tell you exactly how long the transfer should take.