Ethernet should be faster than Wi-Fi. When it isn’t, something specific is wrong. This article is about the common situations where a wired Ethernet connection tests noticeably slower than the Wi-Fi connection on the same machine — and how to fix them. If you want the full context, see our Complete Guide to Fixing Windows, Browser, and Software Errors.
First, rule out the obvious: run a speed test while connected via Ethernet, then again on Wi-Fi. Do this a few times each and take averages. If Ethernet is consistently slower, not just one random test: there’s a real issue. If it’s similar or Ethernet is actually faster: this page might not be what you need.
The most common cause — responsible for a majority of cases — is duplex mismatch. The Ethernet adapter and the router or switch can’t agree on speed and duplex settings, so they fall back to half-duplex at a lower speed. Traffic in both directions has to take turns rather than running simultaneously, cutting effective throughput by 50% or more. Test this first before anything else.
Fix duplex mismatch immediately
Device Manager → Network adapters → your Ethernet adapter → right-click → Properties → Advanced tab → find “Speed & Duplex” in the list → change from “Auto Negotiation” to an explicit value: 1.0 Gbps Full Duplex (or 100 Mbps Full Duplex for older hardware) → OK → reconnect and retest.
If speed improves: auto-negotiation was failing. The explicit setting forces the correct mode. You can try switching back to Auto Negotiation after updating your Ethernet adapter driver — sometimes the negotiation issue resolves with a newer driver, letting you use auto again.
Cable and port quality
A bad Ethernet cable is the second-most-common cause. Cat5e cables handle 1 Gbps at up to 100 meters. Cat5 (without the ‘e’) only reliably handles 100 Mbps. Cat6 and Cat6a handle 10 Gbps. Damaged cables, cables with bent pins, or cables with poor termination fall back to lower speeds.
Test with a known-good cable. If speed jumps: the old cable is damaged or poor quality. The symptom is usually consistent — it doesn’t vary much between tests if the cable is the issue.
Port quality also matters. If you’re going through a switch or hub: swap to a different port. Old unmanaged switches occasionally have one or two ports with degraded performance. If the Ethernet port is on a docking station or USB-to-Ethernet adapter: test directly to the router — docking station Ethernet is often USB 2.0-backed, which caps around 300-400 Mbps regardless of your actual line speed.
Network adapter driver
Windows 11 installs generic Ethernet drivers that work but don’t always enable full performance for specific hardware. Intel, Realtek, and Aquantia (AQTION) all release specific drivers with performance optimisations.
Identify your adapter: Device Manager → Network adapters → note the name of the Ethernet adapter (e.g., “Intel I225-V,” “Realtek PCIe GbE Family Controller”). Then download the latest driver directly:
- Intel: intel.com → download centre → Ethernet drivers → match to your model
- Realtek: your motherboard manufacturer’s support page is better than Realtek’s own site
- Aquantia/Marvell: marvell.com → support → search for your adapter model
After installing the manufacturer driver: run the duplex fix again (set to explicit full duplex) and retest. The combination of correct driver and explicit duplex resolves most adapter-side performance problems.
Power saving on the Ethernet adapter
Windows enables power saving by default on network adapters, which throttles throughput when the adapter thinks it’s underutilised. Test with power saving disabled:
Device Manager → Network adapters → your adapter → Properties → Power Management tab → uncheck “Allow the computer to turn off this device to save power.” Also check the Advanced tab → look for “Energy Efficient Ethernet” or “Power Saving Mode” → disable these.
Energy Efficient Ethernet (EEE / 802.3az) specifically can cause performance inconsistency — the adapter enters low-power state during quiet periods and takes a moment to return to full speed. Some adapters and switches don’t implement EEE handshaking cleanly, causing stuttering throughput. Disabling EEE on the adapter side often improves measured speed noticeably.
Our guide on Wi-Fi troubleshooting covers the wireless side if you’re also experiencing Wi-Fi problems, and our Ethernet connectivity troubleshooting covers cases where Ethernet isn’t connecting at all rather than just being slow. For Ethernet performance monitoring and bottleneck identification, nPerf’s network testing tools provide more detailed throughput analysis than basic speed tests.
Router Ethernet port speed
The port on the router matters too. Many ISP-provided routers have 100 Mbps LAN ports, not Gigabit. A port labelled “LAN” isn’t necessarily Gigabit — 100 Mbps is still standard on budget routers in 2026. If your internet plan is 500 Mbps but Ethernet gives 95 Mbps: the router port is almost certainly 100 Mbps.
Check: look at the router’s spec sheet (search the model number + “specifications”). If it says “Fast Ethernet” or “10/100” for LAN ports: you’re capped at ~95 Mbps. “Gigabit” or “10/100/1000” ports can deliver 1 Gbps.
This also explains why Wi-Fi might test faster — your laptop’s Wi-Fi 6 running on a clear 5 GHz channel achieves 400+ Mbps, while Ethernet through the 100 Mbps port tops out at 95 Mbps. Ethernet would actually be faster if you replaced the router with one that has Gigabit LAN ports.
USB-to-Ethernet adapter bottlenecks
If you’re using a USB-to-Ethernet adapter (common on thin laptops without a built-in Ethernet port), the USB generation matters enormously:
| USB generation | Theoretical max | Realistic Ethernet max |
| USB 2.0 | 480 Mbps | ~250-300 Mbps (shared bus) |
| USB 3.0 / 3.1 Gen 1 | 5 Gbps | ~900-950 Mbps |
| USB 3.1 Gen 2 / USB-C | 10 Gbps | Near-Gigabit, minimal overhead |
A USB 2.0 adapter can’t deliver Gigabit Ethernet. Period. If you have a MacBook-style thin laptop with only USB-C ports and are using a dock or dongle with Ethernet: check whether the dock uses USB 2.0 or 3.0 internally. Budget docks often have USB 2.0 Ethernet chips regardless of the USB-C connection. The specification page for the dock will say “100 Mbps Ethernet” for USB 2.0 implementations.
Network stack optimisation
Windows 11 has several TCP/IP settings that affect throughput. Default settings are reasonable but can be improved for high-speed connections:
Open PowerShell as Administrator and run:
netsh int tcp set global autotuninglevel=normal
netsh int tcp set global chimney=disabled
netsh int tcp set heuristics disabled
netsh int tcp set global rss=enabled
netsh int tcp set global fastopen=enabled
RSS (Receive Side Scaling) distributes incoming network packets across multiple CPU cores, improving throughput. FastOpen reduces connection latency. These settings are safe to apply and can be reverted with netsh int tcp set global autotuninglevel=default if needed.
Also useful: netsh int ip set global taskoffload=enabled → enables TCP offload to the network adapter (hardware-level packet processing rather than CPU). On modern adapters this provides 5-15% throughput improvement under load.
MTU settings
The MTU (Maximum Transmission Unit) affects how large each network packet is. Default is 1500 bytes for Ethernet. Mismatched MTU between the adapter and the network causes excessive fragmentation, which tanks throughput.
Check current MTU: netsh interface ipv4 show subinterfaces → look at the MTU value for your Ethernet adapter. If it’s not 1500: netsh interface ipv4 set subinterface "Ethernet" mtu=1500 store=persistent (replace “Ethernet” with your adapter’s exact name from the output).
Some ISPs use PPPoE which reduces effective MTU to 1492. VPN connections further reduce it. If your ISP uses PPPoE and standard MTU is 1500, large packets get fragmented. Changing to 1492 or 1480 for PPPoE connections can improve throughput by eliminating fragmentation overhead.
Background activity consuming bandwidth
If speed tests show reasonable Ethernet speed but real-world browsing or downloads feel slow: something else on the machine or network is consuming bandwidth in the background.
Check: Task Manager → Performance → Ethernet → look at current send/receive rates while you think nothing is running. If the bars show significant activity: find what’s using the bandwidth via Processes → sort by Network column. Windows Update, OneDrive sync, Steam background downloads, and Windows Delivery Optimization (sharing updates with other PCs) are common heavy consumers.
For Delivery Optimization specifically: Settings → Windows Update → Advanced options → Delivery Optimization → “Allow downloads from other PCs” → set to “Devices on my local network only” or disable entirely. This prevents your machine from uploading Windows updates to internet peers, which can consume significant upload bandwidth and affect overall connection performance.
| Ethernet speed issue | Likely cause | Targeted fix |
| ~50% of expected speed, inconsistent | Duplex mismatch | Force 1 Gbps Full Duplex in adapter advanced settings |
| Caps at ~95 Mbps | Router or switch has 100 Mbps LAN ports | Verify router specs; upgrade if needed |
| Caps at ~250-300 Mbps | USB 2.0 adapter or dock | Upgrade to USB 3.0 dock or built-in Ethernet |
| Inconsistent speed, stutters | Energy Efficient Ethernet or power saving | Disable EEE and power saving in adapter advanced settings |
| Suddenly slower than before | Driver update or Windows update | Reinstall manufacturer Ethernet driver |
| Fine speeds but slow overall | Background bandwidth consumption | Task Manager → Processes → find culprit; disable Delivery Optimization |
Ethernet being slower than Wi-Fi usually points to a hardware configuration issue rather than a complex software problem. Duplex mismatch plus power-saving settings together resolve the majority of cases — and both are fixed in the same Advanced tab of the network adapter properties, which makes them easy to address in one go.
Using iperf3 for accurate Ethernet testing
Speed tests measure your internet connection, not your Ethernet connection specifically. A bottleneck at the ISP or speed test server can mask good local Ethernet performance. For accurate local network measurement: use iperf3.
Install iperf3 on two Windows machines (or one Windows machine and a router with iperf3 support): iperf.fr → download for Windows. On one machine, run iperf3 -s (server mode). On the other, run iperf3 -c [server-IP]. This measures raw Ethernet throughput between two local devices, completely bypassing the internet connection and its limitations.
Expected result on a working Gigabit Ethernet connection: 900-950 Mbps. If you see 400-500 Mbps: power saving or Energy Efficient Ethernet. If you see ~100 Mbps: 100 Mbps port or duplex mismatch. If you see ~250 Mbps: USB 2.0 adapter. These ranges map directly to the causes and fixes above.
When Wi-Fi is actually better — and that’s okay
Sometimes Wi-Fi is genuinely faster than Ethernet for specific scenarios. Wi-Fi 6 can sustain 600+ Mbps in ideal conditions. If your internet plan is 500 Mbps, a modern Wi-Fi 6 connection can max it out — and for that use case, Ethernet’s additional latency stability and full-duplex advantages matter less.
Ethernet’s real advantages over Wi-Fi are:
- Lower and more consistent latency (important for gaming, VoIP, video calls)
- No wireless interference or signal degradation
- Full-duplex communication (simultaneous upload and download)
- More reliable connection — no dropped frames from interference
- Security — no wireless eavesdropping
If a speed test shows Wi-Fi at 450 Mbps and Ethernet at 420 Mbps on a 500 Mbps internet plan: both are essentially maxing out the ISP connection. For practical purposes, Ethernet isn’t “slower” — it’s within measurement margin of an ISP-capped connection. The meaningful Ethernet advantage in that scenario is in latency and consistency, not raw speed.
Gigabit Ethernet with 100 Mbps internet
Many users don’t realise that a Gigabit Ethernet port has no value above your ISP’s plan speed for downloads from the internet. A 1 Gbps port with a 100 Mbps DSL plan will never go above 100 Mbps on internet speed tests — the ISP is the bottleneck, not the Ethernet connection.
Where Gigabit Ethernet matters even on a slow internet plan: local network transfers (copying files between two PCs, accessing a NAS, streaming from a local media server). These transfers are purely local and benefit from Gigabit speeds regardless of internet plan. A 10 GB video file copies in ~90 seconds over Gigabit versus 13+ minutes over 100 Mbps — a meaningful difference for anyone doing regular local file work.
Understanding this distinction prevents hours of troubleshooting “why is my Gigabit Ethernet giving me only 80 Mbps” when the answer is simply that the ISP plan only offers 80 Mbps, and the Ethernet connection is performing perfectly.
If after applying all the fixes above — duplex forced to Gigabit Full Duplex, EEE disabled, power saving off, latest manufacturer driver installed, tested with a known-good Cat5e or Cat6 cable — Ethernet still tests 50% below Wi-Fi: the network adapter itself may be defective, the Ethernet port on the router may be damaged, or the cable run has a fault somewhere. A USB 3.0 Gigabit Ethernet adapter (Plugable, UGREEN, StarTech — around £20-£30) used as a test device quickly isolates whether the problem is the built-in adapter or the network path. If the USB adapter gives correct speed, the built-in adapter is the issue. You might also run into Ethernet Slower Than WiFi in Windows 11.
