“How long will this take?” is the internet’s oldest question, and its answer is one division — once the units stop lying to you. This Upload Time Calculator divides your file size by your connection speed and returns the duration as hours, minutes and seconds, with the real-world overhead figure beside it so the progress bar’s behaviour stops being a betrayal.
Worked Example: 10 GB at 100 Mbps
The file carries 10 × 1,000 = 10,000 MB, which in the bits that network speeds are sold in is 80,000 megabits. At 100 megabits per second the transfer takes 80,000 / 100 = 800 seconds — 13 minutes 20 seconds at the theoretical line rate. The calculator’s note immediately applies the real-world correction: protocol overhead, congestion and the far server’s own limits typically deliver about 90% of the nominal rate on a healthy connection, stretching the honest expectation to about 14 minutes 49 seconds. Two traps hide in this short arithmetic, and both are famous. The first is the bits-versus-bytes factor of eight: file sizes are measured in bytes, speeds in bits, and the user who divides 10 GB by “100” expecting 0.1 of something is off by that factor of eight before anything else goes wrong — the calculator exists largely to make this conversion invisible. The second is asymmetry: most home connections advertise the download speed in letters a foot high and the upload speed in the footnotes — a “500 Mbps” cable plan may upload at 20 Mbps, turning the example’s quarter-hour into five and a half hours. Speed-test your connection in the direction you actually intend to move the file, and enter that number: the calculator is exact, and deserves an exact input.
The Units, Stated Once and Precisely
File sizes and speeds carry a second, quieter ambiguity: the prefixes themselves. This calculator uses the decimal convention of the storage and networking industries — 1 GB = 1,000 MB = 1,000,000,000 bytes, and 1 Mbps = 1,000,000 bits per second — which matches how drives are marketed, how operating systems increasingly report, and how ISPs sell. The binary convention (1 GiB = 1,024 MiB) lives on in memory sizing and some operating-system displays, and it differs by about 7% at the gigabyte scale and 10% at the terabyte — small next to the factor-of-eight trap, but large enough to explain a progress bar that disagrees with a hand calculation by a few minutes per hour. The duration formula itself is unit-clean: time = file size in bits ÷ speed in bits per second, and everything else is prefix bookkeeping. A related figure for planning: the same division run backwards gives the size a connection moves in a fixed window — one hour at a sustained 100 Mbps is 45 GB — which is the budgeting form of the question for backup windows, livestream durations (a 25 Mbps stream writes 11.25 GB an hour) and satellite-link scheduling. However the question is dressed, it is this one division, and the only ways to get it wrong are the factor of eight, the direction of the connection, and a prefix read carelessly.
Why Real Transfers Miss the Line Rate
The 90% planning factor folds in several honest taxes. Protocol overhead: TCP/IP framing, TLS encryption and the transfer protocol’s own chatter consume a few percent of raw capacity before your file’s first byte moves. Latency-bound protocols: a single TCP stream across a long, fat pipe can be limited by the bandwidth-delay product rather than the link itself — which is why parallel-stream transfer tools exist and why a transoceanic upload underperforms a local one at identical nominal speeds. Congestion and shared media: cable segments, Wi-Fi airtime and the destination server’s other customers all take their turns; Wi-Fi deserves its special mention, because a weak-signal wireless hop between you and the router is routinely the slowest link in the whole path, and no ISP upgrade repairs it. Small-file penalties: a million tiny files transfer far slower than one archive of the same total size — per-file handshakes dominate — so the calculator’s figure assumes a single stream of bulk data; for many-small-files jobs, zip first and compute second. And throttling: some services cap per-connection or per-account rates below the line’s capacity. The planning habit that survives all five: compute the line-rate figure here, budget 90% of the speed for bulk transfers on healthy wired connections, 70–80% when Wi-Fi or small files are involved, and treat any transfer that beats the line-rate number as a sign the file was smaller than advertised.
The Reverse Budget: What Fits in the Window?
The same division, aimed the other way, answers the planner’s form of the question: how much data fits a fixed window at a fixed speed? An overnight backup window of 8 hours on a 25 Mbps upload moves 25 × 28,800 / 8 = 90,000 MB = 90 GB at line rate — about 81 GB at the 90% planning factor — which is the number that decides whether a backup strategy is nightly-complete or incrementally honest. Scale the example to the classic home-backup problem: a 1 TB photo library’s first full upload on a 20 Mbps uplink takes 1,000,000 × 8 / 20 = 400,000 seconds — 111 hours, about 4.6 days of continuous, perfectly behaved transfer. That arithmetic, run before the subscription is bought, is why seed drives, local-first strategies and delta-sync tools exist; run after, it is a complaint. The livestream version: a 3-hour event streamed at 12 Mbps writes 12 × 10,800 / 8 = 16.2 GB of recording — the file the upload calculator then prices for the trip to the archive. Windows, events and libraries are the same formula wearing schedules: size in bits divided by rate, with the efficiency factor as the voice of experience.
Choosing What to Measure: Speed Tests Honestly
The calculator is only as truthful as its speed input, and speed figures come in three grades of honesty. The plan’s headline is a marketing ceiling for the download direction under ideal conditions — the least useful number for planning. A speed test’s result is a real measurement to a nearby server over a few seconds, in whichever direction you selected — much better, and the right input when taken wired, at the hour you will actually transfer, in the right direction. A sustained-transfer measurement — the throughput a real long upload settles into against the real destination — is the gold figure, and it is routinely 10–30% below the speed test’s sprint, because distant servers, congestion and per-connection caps only reveal themselves over minutes. Measure once against your actual archive, backup service or client portal, note the sustained figure, and reuse it as the input here; the calculator’s 90% default then becomes a cross-check rather than a guess. For shared connections — a household’s uplink shared with video calls and game downloads — the honest input is the share you can claim, and a transfer scheduled into the house’s quiet hours is worth more than a faster plan fought over at dinner time.
Common Mistakes
Forgetting the factor of eight between bytes and bits (the eight-fold error that built this page). Entering the plan’s download speed for an upload. Mixing decimal and binary gigabytes in the same estimate. Trusting a speed test taken over Wi-Fi at the far end of the house as the wired truth. Estimating a folder of ten thousand small files as if it were one file of the same size. And believing the progress bar’s first-minute extrapolation — most bars average over a window that includes the handshake, which is why their early estimates swing wildly before settling near this page’s figure.
Related Tools
Continue with the RAID Calculator, the CIDR Calculator, the Depth of Field Calculator, or browse the full Other Calculators hub.
Frequently Asked Questions
How long does it take to upload 10 GB at 100 Mbps?
800 seconds — 13 minutes 20 seconds at the full line rate, and about 14 minutes 49 seconds at a realistic 90% efficiency. Remember the factor of eight: 10 GB is 80,000 megabits, not 10,000.
Why is my upload so much slower than my download?
Most consumer connections are asymmetric by design — the headline number is the download rate, while upload may be a tenth of it. Always enter the speed for the direction you are transferring; for uploads that is the small number in the plan’s fine print.
Should I use GB = 1,000 or 1,024 MB?
This calculator uses the decimal 1,000 convention of the storage and networking industries. The binary (GiB) convention differs by about 7% at this scale — enough to explain small disagreements, never an eight-fold one.
How much can I transfer in one hour at 50 Mbps?
50 × 3,600 / 8 = 22,500 MB — 22.5 GB at line rate, or roughly 20 GB at the 90% planning factor. The same formula sizes backup windows and stream archives.
Why do many small files take so much longer?
Each file costs handshakes, metadata and often a fresh connection — overhead that dwarfs the data for tiny files. Archive them into one file first; the calculator’s single-stream assumption then applies.
Does Wi-Fi change the calculation?
The formula is the same; the honest speed input is not. A weak Wi-Fi hop is often the true bottleneck — budget 70–80% of the nominal rate, or measure the speed end to end and enter the measured figure.