This is not a hands-on benchmark and does not declare a permanent speed winner. VPN speed changes with the device, protocol, server distance, server load, ISP route, time of day, and the activity being measured. A result from one lab or one home connection is evidence about that setup—not a promise for every subscriber.
The protocol difference
NordVPN's NordLynx is built around WireGuard and adds NordVPN's own privacy design. ExpressVPN's Lightway is its own protocol, available across the provider's major apps and described by ExpressVPN as designed for speed, reliability, and easier auditing. ExpressVPN has also published a Praetorian Lightway security assessment, which is evidence about that assessment's scope and date—not proof that every future client or route is risk-free. These are not independent proof that one is faster for every route.
Both providers also expose fallback protocols or settings for networks where the default does not connect. Use the app's recommended modern protocol first, then test a fallback only when compatibility requires it.
What changes the result more than the brand name
- Endpoint distance: A nearby server usually has less latency than a distant one.
- Route and congestion: ISP peering, Wi-Fi, time of day, and VPN-server load can dominate a protocol difference.
- Device and app: CPU, operating system, browser, background traffic, and client implementation matter.
- Use case: Video calls and games need stable latency and low jitter; downloads and streaming need sustained throughput; browsing needs responsive DNS and page loading.
- Policy and location: A server that is fast may not be the right legal, privacy, or content-access choice for the task.
A repeatable buyer-side test
- Record a baseline on the same device and connection with the VPN off.
- Test both services in the same nearby region, then in the region you actually need.
- Use each provider's recommended protocol and repeat each measurement at least three times.
- Record latency, jitter, download/upload throughput, DNS behavior, video-call stability, and the time to connect.
- Repeat at another time of day and on the network where you will actually use the VPN.
- Test sleep/wake, Wi-Fi changes, reconnects, and the apps that matter—not only a speed-test page.
Do not average unlike tests or present a single number as a provider-wide result. If the money-back window permits, run this test with the exact Mac, Windows, phone, router, and destinations you care about.
How to report a result honestly
If we publish a hands-on result later, the test record should include the date, city or region, ISP, connection type, device, operating-system/app versions, protocol, server location, baseline speed, VPN speed, latency, jitter, and number of repetitions. Report median and range rather than the single best run. A result should be labeled as “on this setup” and separated from provider-documented features. Do not turn a short test into claims about privacy, streaming access, gaming quality, or universal speed.
Choosing between them
Choose based on the route and controls you need, not a permanent “fastest” label. Compare current device limits, protocol choices, server locations, kill-switch behavior, privacy documentation, support, renewal price, and refund terms. A provider can be fast on one ISP and disappointing on another; a nearby endpoint and stable route may matter more than a small protocol benchmark difference.
Bottom line
NordLynx and Lightway are both current, purpose-built protocol options, but this article does not claim to have measured a universal winner. Test both on your own connection, record the conditions, and choose the service that delivers the stability, location, privacy evidence, controls, and total cost your use case requires.