Mbps vs MBps: Why Your Internet Speed Numbers Don't Match

You pay for "100 Mbps" internet. You start downloading a large file and your browser shows 12.5 MB/s. You feel cheated. You are not being cheated. Both numbers are correct, and the difference comes down to a single uppercase letter.

That lowercase "b" in Mbps means bits. The uppercase "B" in MBps means bytes. A byte is 8 bits. Divide 100 by 8 and you get 12.5. Your connection is doing exactly what your ISP promised. The confusion is baked into the notation, and once you see the pattern, the math never surprises you again.

The 8x Factor

The relationship is simple and absolute. One byte equals 8 bits. This is not approximate, not a rounding convention, not something that varies by manufacturer. IEC 80000-13:2008 defines the byte as exactly 8 bits.

The notation follows a straightforward rule:

• Mbps (lowercase b) = megabits per second
• MBps or MB/s (uppercase B) = megabytes per second

To go from Mbps to MBps, divide by 8. To go from MBps to Mbps, multiply by 8. That single operation explains the gap between what your ISP advertises and what your download manager reports.

Here is what common ISP plan speeds translate to in actual megabytes per second:

If you have a gigabit connection, your theoretical ceiling is 125 MBps. Not 1,000 MBps. The math is the same at every speed tier.

Why ISPs Advertise in Bits

ISPs did not pick bits to confuse you, though the marketing advantage is real. The convention comes from how network hardware actually works.

Network equipment transmits data serially, one bit at a time, over a wire, fiber strand, or radio channel. The IEEE 802.3 Ethernet standards, which define wired networking from 10 Mbps all the way up to 400 Gbps, specify all data rates in bits per second. When AT&T, Comcast, or Deutsche Telekom advertise a speed, they are using the same units the underlying hardware uses. The alternative would require them to translate the spec sheets, which would only add another source of potential confusion.

That said, larger numbers do look better in advertising. "100 Mbps" sounds eight times faster than "12.5 MBps," even though they describe the same speed. ISPs are not going to volunteer a smaller number.

Why Your Actual Speed Is Even Lower

Even after dividing by 8, your real download speed will come in below the theoretical maximum. Every piece of data sent over the internet carries overhead.

TCP/IP protocol headers wrap your actual file data in addressing and error-correction information. Ethernet framing adds more. Combined, these protocol layers consume roughly 5 to 10 percent of your raw bandwidth. On a 100 Mbps connection, the usable throughput for file data is typically 90 to 95 Mbps, which works out to about 11.2 to 11.9 MBps rather than the theoretical 12.5.

Wi-Fi makes it worse. Wireless protocols (governed by IEEE 802.11 standards) add collision avoidance mechanisms, encryption overhead, and retransmission when signals degrade. Walls, distance from the router, interference from neighboring networks, and the number of connected devices all eat into throughput. A 100 Mbps Wi-Fi connection delivering 8 to 10 MBps of real file throughput is common and not a sign that anything is broken.

FCC Broadband Benchmarks

In March 2024, the FCC updated its definition of broadband to a minimum of 100 Mbps download and 20 Mbps upload. The previous threshold, 25 Mbps down and 3 Mbps up, had been in place since 2015 and was widely considered outdated. Under the new standard, a connection needs to deliver at least 12.5 MBps download speed to qualify as broadband.

Broadband speeds have climbed steadily as fiber and DOCSIS 3.1 cable deployments expand, but tens of millions of Americans, particularly in rural areas, still fall below the 100 Mbps threshold.

The Storage Side of the Confusion

Bits versus bytes is not the only place where similar-looking prefixes mean different things. Storage has its own version of this problem.

The IEC published standard 80000-13, which introduced binary prefixes to separate powers of 1,024 from powers of 1,000. Under this system, 1 kibibyte (KiB) equals 1,024 bytes, while 1 kilobyte (kB) equals exactly 1,000 bytes. The distinction scales up: 1 gibibyte (GiB) is 1,073,741,824 bytes, while 1 gigabyte (GB) is 1,000,000,000 bytes.

In practice, adoption is inconsistent. Windows displays storage using binary values but labels them "GB," so a 500 GB hard drive shows up as roughly 465 "GB" in File Explorer. macOS switched to decimal reporting in 2009, so the same drive shows the full 500 GB. Storage manufacturers use the decimal definition (1 GB = 1,000,000,000 bytes) because that matches SI conventions and, not coincidentally, produces a larger number on the label.

NIST has published the binary prefix standards and recommends their use, but the computing industry has not uniformly adopted them. The result: "GB" on a hard drive box and "GB" on a Windows screen refer to slightly different quantities, and most people never realize it.

The Short Version

When your ISP says Mbps, they mean megabits. When your browser says MB/s, it means megabytes. Divide by 8 to translate. Subtract another 5 to 10 percent for protocol overhead, and more if you are on Wi-Fi. That is the entire explanation. Nobody is lying to you. The units are just different.

Sources: IEEE 802.3 Ethernet Standards, IEC 80000-13:2008, FCC Broadband Benchmark Report (2024), NIST Prefixes for Binary Multiples, RFC 793 (TCP)