AB-006802.11ac Basics
802.11ac Maximum Speed: Real-World Data Rates
How the 802.11ac maximum speed is calculated from channel width, spatial streams, and modulation, with the real Mbps figures behind 6.93 Gbps.

What is the maximum speed of 802.11ac?
The maximum theoretical speed of 802.11ac is 6.93 Gbps (6933 Mbps), reached only with the largest configuration the standard defines: 8 spatial streams, a 160 MHz channel, and 256-QAM modulation at the highest defined coding rate. The IEEE 802.11ac-2013 amendment was written to deliver significantly higher throughput than its predecessors for operation below 6 GHz (https://standards.ieee.org/ieee/802.11ac/4473/), and this figure is the ceiling that follows from its channel widths, spatial stream limit, and modulation scheme. No deployed consumer or enterprise hardware reaches this number in practice: real products stop well short of 8 streams, and real throughput is always lower than any PHY rate. The sections below show where the number comes from and what narrower, more common configurations actually deliver.
How is that maximum data rate calculated?
The data rate for one spatial stream is the number of usable data subcarriers in the channel, multiplied by the bits per symbol the modulation carries, multiplied by the forward error correction code rate, divided by the OFDM symbol duration. Widening the channel adds subcarriers, a higher-order modulation like 256-QAM adds bits per symbol compared with 64-QAM, and a higher code rate keeps more of those bits as usable data rather than error correction overhead. Adding spatial streams then multiplies the single-stream figure directly, since each stream carries an independent copy of this calculation over its own antenna path. The final 6.93 Gbps figure is really three multipliers stacked together: channel width, modulation and coding, and spatial stream count.
What data rate does a single spatial stream deliver at each channel width?
Using the highest modulation and coding scheme available at each width (256-QAM at the top code rate, with the short guard interval vendors commonly use for headline figures), one spatial stream delivers:
| Channel width | Highest coding scheme | Single-stream rate |
|---|---|---|
| 20 MHz | MCS8 (256-QAM) | 86.7 Mbps |
| 40 MHz | MCS9 (256-QAM) | 200.0 Mbps |
| 80 MHz | MCS9 (256-QAM) | 433.3 Mbps |
| 160 MHz | MCS9 (256-QAM) | 866.7 Mbps |
Notice that 20 MHz tops out one modulation and coding step below the others. For one, two, four or eight streams, the 802.11ac amendment does not define MCS9 at 20 MHz, because the number of data bits per symbol would not come out as a whole number, so those 20 MHz links are capped at MCS8. Three- and six-stream links are the exception: at 20 MHz they can use MCS9. This is also why 20 MHz delivers less than exactly a quarter of the 80 MHz figure, even though a quarter is a reasonable first guess. For more on how width is chosen in practice, see channel width in 802.11ac.
How many spatial streams does real hardware actually support?
The 802.11ac specification allows up to 8 spatial streams, but almost no shipped hardware uses that many. Most consumer routers and access points implement 3 or 4 streams on their 802.11ac radio; many client devices, including most phones and laptops, implement only 1 or 2. An 8-stream AP would also need 8 antennas dedicated to 802.11ac alone, which is uncommon outside specialized enterprise gear. In practice, the real ceiling on any given link is set by whichever side, AP or client, supports fewer streams: a 4-stream AP talking to a 2-stream laptop is limited to 2-stream performance on that connection.
What speed can you expect from typical 3-stream or 4-stream hardware?
A common 80 MHz, 3-stream configuration, the basis for many mid-range 802.11ac routers, reaches 1300.0 Mbps (3 x 433.3 Mbps). A 4-stream radio at the same 80 MHz width reaches 1733.3 Mbps. These are the PHY rates that most buyers actually encounter: well below the 8-stream, 160 MHz ceiling, and still well above anything a single client will see as real throughput once overhead and distance are accounted for.
Full table: maximum theoretical speed by channel width and spatial stream count
| Streams | 20 MHz | 40 MHz | 80 MHz | 160 MHz |
|---|---|---|---|---|
| 1 | 86.7 Mbps | 200.0 Mbps | 433.3 Mbps | 866.7 Mbps |
| 2 | 173.3 Mbps | 400.0 Mbps | 866.7 Mbps | 1733.3 Mbps |
| 3 | 288.9 Mbps | 600.0 Mbps | 1300.0 Mbps | 2340.0 Mbps |
| 4 | 346.7 Mbps | 800.0 Mbps | 1733.3 Mbps | 3466.7 Mbps |
| 8 | 693.3 Mbps | 1600.0 Mbps | 3466.7 Mbps | 6933.3 Mbps |
The bottom-right cell, 8 streams at 160 MHz, is the 6.93 Gbps headline figure. Almost every other cell is the same per-stream rate from the earlier table, multiplied by stream count. The two exceptions are both on the 3-stream row: at 20 MHz, three streams can use MCS9 (288.9 Mbps), while at 160 MHz the standard does not define MCS9 for three streams, so that link tops out at MCS8 (2340.0 Mbps).
Why is real-world throughput always lower than these theoretical numbers?
These figures are PHY, or physical layer, rates: what the radio can put on the air during a single transmission, not what an application measures. Real throughput loses ground to inter-frame spacing, acknowledgement frames, retransmissions, and contention with other devices on the same channel, all before distance enters the picture. Signal quality falls with distance and obstacles, which forces the link down to a lower coding scheme long before it drops to zero, so a client far from the AP may never reach the top modulation at all. Mixed environments with legacy 802.11n or 802.11a clients add further overhead, since the AP has to accommodate slower negotiation and protection mechanisms for those devices. As a rule of thumb used across the industry, expect real single-client throughput in the range of 50 to 70 percent of the PHY rate under good conditions, and less in a crowded or obstructed environment.
How does modulation (QAM) affect the maximum achievable speed?
Modulation sets how many bits each OFDM symbol can carry. 802.11n tops out at 64-QAM, which carries 6 bits per symbol; 802.11ac raises that ceiling to 256-QAM, carrying 8 bits per symbol, a one-third increase in raw bit density before any other factor is considered. The trade-off is reliability: 256-QAM packs the signal's amplitude and phase into finer distinctions, so it needs a cleaner, higher signal-to-noise ratio to decode correctly. That is why the top coding scheme is only usable close to the AP or in a quiet RF environment, and why the headline 6.93 Gbps figure describes a best-case link, not a typical one. For a wider comparison of what changed between generations, see 802.11ac vs 802.11n.
How does 802.11ac's maximum speed compare with Wi-Fi 4 and Wi-Fi 6?
802.11n (Wi-Fi 4) caps out at 600 Mbps with 4 streams and 64-QAM, well under even the single-stream 160 MHz rate of 802.11ac. Wi-Fi 6 (802.11ax) keeps 802.11ac's 256-QAM as a baseline and adds an optional 1024-QAM mode, which raises the per-stream ceiling further, alongside OFDMA scheduling that changes how that capacity gets shared among multiple devices rather than how fast any single transmission runs. For the generation-by-generation picture, see Wi-Fi generation labels explained and 802.11ac vs Wi-Fi 6: what changed.
Conclusion
The maximum theoretical speed of 802.11ac, 6.93 Gbps, comes from stacking its three main levers: up to 160 MHz of channel width, up to 8 spatial streams, and 256-QAM modulation at the highest defined coding rate. Real hardware uses a fraction of that ceiling, typically 3 or 4 streams at 80 MHz, and real throughput sits well under even that PHY rate once overhead, distance, and contention are accounted for. Knowing the per-stream figures at each channel width makes it possible to read a product's advertised speed and understand what configuration actually produced it, rather than treating any single number as a promise of real-world performance.

