Skip to main content

Data Transfer Converter

Type a number in the Enter Speed Value field, choose your unit from the Select Speed Unit dropdown, and the results grid converts to all 34 transfer rate units — bit-based (Mbps, Gbps) and byte-based (MB/s, GB/s). Or pick any unit below for its dedicated conversion page.

Universal Speed Converter

Bits Per Second
bps
Kilobits Per Second
Kbps
Gigabits Per Second
Gbps
Terabits Per Second
Tbps
Petabits Per Second
Pbps
Exabits Per Second
Ebps
Zettabits Per Second
Zbps
Yottabits Per Second
Ybps
Bytes Per Second
B/s
Kilobytes Per Second
KB/s
Megabytes Per Second
MB/s
Gigabytes Per Second
GB/s
Terabytes Per Second
TB/s
Petabytes Per Second
PB/s
Exabytes Per Second
EB/s
Zettabytes Per Second
ZB/s
Yottabytes Per Second
YB/s
Kibibytes Per Second
KiB/s
Mebibytes Per Second
MiB/s
Gibibytes Per Second
GiB/s
Tebibytes Per Second
TiB/s
Pebibytes Per Second
PiB/s
Exbibytes Per Second
EiB/s
Zebibytes Per Second
ZiB/s

Browse by Speed Unit

How Data Transfer Speeds Work

Data transfer speed measures how much data moves from point A to point B in a given amount of time. But there are two distinct concepts that get conflated constantly: bandwidth and throughput.

Bandwidth is the theoretical maximum capacity of a connection — think of it as the width of a pipe. A Gigabit Ethernet port has a bandwidth of 1 Gbps. It physically cannot push bits through the wire any faster than that.

Throughput is what you actually get — the amount of useful data transferred per second after accounting for protocol overhead, congestion, latency, and error correction. On that same Gigabit Ethernet link, real-world throughput typically lands between 900-940 Mbps for TCP traffic.

The other source of confusion is the unit itself. Networks, ISPs, and hardware specs almost universally measure in bits per second (bps, Kbps, Mbps, Gbps). But when you download a file, your browser or file manager shows progress in bytes per second (B/s, KB/s, MB/s). Since 1 byte = 8 bits, you need to divide the bit-rate by 8 to get the byte-rate:

  • 100 Mbps bandwidth = 12.5 MB/s maximum throughput
  • 1 Gbps bandwidth = 125 MB/s maximum throughput
  • 10 Gbps bandwidth = 1.25 GB/s maximum throughput

This 8x factor is the single most common source of confusion in networking. When your ISP advertises "500 Mbps" internet, your peak download speed in a browser is about 62.5 MB/s — and that's before any real-world overhead reduces it further.

Complete Transfer Speed Reference

The tables below cover all standard bandwidth and throughput units, along with a practical reference showing how long each speed takes to transfer 1 GB of data.

Bandwidth Units (Bits per Second)

Unit Symbol Bits per Second Time to Transfer 1 GB
Bit per secondbps1~254 years
Kilobit per secondKbps1,000~92.6 days
Megabit per secondMbps1,000,000~8 seconds
Gigabit per secondGbps1,000,000,000~8 ms
Terabit per secondTbps1,000,000,000,000~0.008 ms
Petabit per secondPbps1015~8 ns

Throughput Units (Bytes per Second)

Unit Symbol Bytes per Second Time to Transfer 1 GB
Byte per secondB/s1~31.7 years
Kilobyte per secondKB/s1,000~11.6 days
Megabyte per secondMB/s1,000,000~16.7 minutes
Gigabyte per secondGB/s1,000,000,000~1 second
Terabyte per secondTB/s1,000,000,000,000~1 ms

Bits per Second vs Bytes per Second — Why the Confusion?

Bits per second (bps, Mbps, Gbps) measure raw data transmission speed. Network equipment and ISPs typically advertise speeds in bits per second because the larger numbers appear more impressive in marketing. Learn more: RAID calculator for storage planning.

Bytes per second (B/s, MB/s, GB/s) represent the actual file transfer rate you experience. Since 1 byte = 8 bits, a 100 Mbps connection delivers approximately 12.5 MB/s of actual throughput. See also: calculate how long data transfer takes.

Understanding this distinction is critical for accurate transfer time calculations. When your ISP promises "1 Gbps" internet, your maximum download speed is roughly 125 MB/s under ideal conditions — and real-world performance is typically 70-80% of theoretical maximum due to protocol overhead. See also: StorageMath calculators.

The bits-vs-bytes confusion has deep historical roots. In the early days of telecommunications, the fundamental unit of data transmission was the baud — one signal change per second. Since early modems encoded one bit per baud, the telecom industry standardised on bits per second as the canonical speed measurement. This convention carried forward into Ethernet standards, Wi-Fi specifications, and ISP marketing.

File systems and operating systems, on the other hand, work with bytes. When you copy a file, the OS tracks progress in bytes because that's how files are addressed on disk. This created two parallel conventions that have never been reconciled:

  • Network layer (everything from the ISP to the NIC): measures in bits per second
  • Application layer (file managers, download tools, backup software): measures in bytes per second

The marketing angle: ISPs have no incentive to switch to bytes per second. "500 Mbps" sounds eight times more impressive than "62.5 MB/s" — even though they describe the same speed. This isn't deceptive (bits per second is the legitimate industry standard), but it does mean consumers need to understand the conversion.

Real-world overhead makes it worse: Even after dividing by 8, you won't see the full byte-rate. TCP/IP headers, Ethernet framing, TLS encryption, and protocol negotiation all consume bandwidth. Typical real-world throughput is 70-85% of the rated speed:

Advertised Speed Theoretical Max (B/s) Typical Real-World Speed Overhead Loss
100 Mbps12.5 MB/s9-11 MB/s~15-25%
500 Mbps62.5 MB/s45-55 MB/s~15-25%
1 Gbps125 MB/s90-110 MB/s~15-25%
2.5 Gbps312.5 MB/s230-280 MB/s~15-25%
10 Gbps1.25 GB/s1.0-1.1 GB/s~10-20%

For local network transfers (NAS backups, file server copies), jumbo frames (9000 MTU) can reduce overhead and push throughput closer to the theoretical maximum. For internet transfers, the overhead is largely unavoidable.

Common Transfer Speed Examples

Your ISP says 200 Mbps — how fast do files actually download?

Start with the advertised speed and divide by 8 to convert bits to bytes:

200 Mbps / 8 = 25 MB/s (theoretical maximum)

Apply real-world overhead (roughly 80% efficiency):

25 MB/s x 0.80 = 20 MB/s (typical download speed)

So a 1 GB file takes about 50 seconds. A 50 GB game takes roughly 42 minutes. These are single-connection estimates — if multiple devices share the connection, each gets a fraction of the total bandwidth.

How long to transfer a 50 GB game at 100 Mbps?

Convert the speed: 100 Mbps = 12.5 MB/s theoretical, ~10 MB/s real-world.

Convert the file size: 50 GB = 50,000 MB.

50,000 MB / 10 MB/s = 5,000 seconds = ~83 minutes

At 1 Gbps (common fiber internet), the same download drops to about 8-9 minutes. At 10 Gbps, it's under a minute.

Wi-Fi 6 vs Ethernet: theoretical vs real speeds

Wi-Fi 6 (802.11ax) advertises speeds up to 9.6 Gbps — but that's the aggregate maximum across all spatial streams and channels. A single client device typically connects at 1.2-2.4 Gbps with a 160 MHz channel. Real-world throughput for a single device usually lands at 500-800 Mbps (62-100 MB/s) under ideal conditions.

Gigabit Ethernet consistently delivers 900-940 Mbps (112-117 MB/s) for TCP traffic with near-zero latency variation. For bandwidth-critical tasks like NAS backups, virtual machine storage, or large file transfers, wired connections are still significantly faster and more reliable than Wi-Fi.

How long to back up 2 TB over USB 3.0 vs Thunderbolt 4?

USB 3.0 (5 Gbps rated, ~400 MB/s real-world): 2,000,000 MB / 400 MB/s = 5,000 seconds = ~83 minutes

Thunderbolt 4 (40 Gbps rated, ~2,800 MB/s real-world with fast SSDs): 2,000,000 MB / 2,800 MB/s = 714 seconds = ~12 minutes

The bottleneck is often the drive speed, not the interface. A SATA SSD maxes out around 550 MB/s regardless of the connection type. An NVMe SSD can saturate USB 3.0 but still has headroom on Thunderbolt 4.

Network Speed Tiers — What You Actually Get

Marketing numbers and real-world performance are two different things. This table shows what you can realistically expect from common connection types, including the time to transfer 1 GB of data:

Connection Type Rated Speed Typical Real-World Speed Time for 1 GB
4G LTE100 Mbps25-50 Mbps (3-6 MB/s)3-6 minutes
5G (sub-6 GHz)1 Gbps100-300 Mbps (12-37 MB/s)27-80 seconds
5G (mmWave)10 Gbps1-4 Gbps (125-500 MB/s)2-8 seconds
Wi-Fi 5 (802.11ac)3.5 Gbps200-400 Mbps (25-50 MB/s)20-40 seconds
Wi-Fi 6 (802.11ax)9.6 Gbps500-800 Mbps (62-100 MB/s)10-16 seconds
100 Mbps Ethernet100 Mbps94 Mbps (11.7 MB/s)~85 seconds
1 Gbps Ethernet1 Gbps940 Mbps (117 MB/s)~8.5 seconds
2.5 Gbps Ethernet2.5 Gbps2.35 Gbps (293 MB/s)~3.4 seconds
10 Gbps Ethernet10 Gbps9.4 Gbps (1.17 GB/s)~0.85 seconds
USB 3.05 Gbps3.2 Gbps (400 MB/s)~2.5 seconds
USB 3.2 Gen 210 Gbps7-8 Gbps (900 MB/s)~1.1 seconds
Thunderbolt 440 Gbps22 Gbps (2,800 MB/s)~0.36 seconds

A few things to note from this table:

  • Wireless speeds vary wildly. The "rated speed" for Wi-Fi and cellular is a best-case lab number. Your actual speed depends on signal strength, interference, distance from the access point, number of connected devices, and channel congestion.
  • Wired connections are predictable. Ethernet and direct-attached interfaces deliver consistent throughput that's close to the rated speed. That's why data centers, NAS setups, and professional studios rely on wired connections.
  • The drive is often the bottleneck. A 10 Gbps Ethernet link can push 1.17 GB/s, but a SATA SSD only reads at ~550 MB/s. You need NVMe storage on both ends to saturate high-speed network links.
  • Latency matters too. For large sequential transfers, throughput is what counts. But for interactive workloads (remote desktops, databases, API calls), latency is often more important than raw bandwidth.