The platforms' own help pages give different numbers, so we put together a quick rundown, and we'd like to hear what you actually see at home.

What the platforms say (download):

- Netflix: 3 Mb/s for HD 720p, 5 Mb/s for Full HD, 15 Mb/s for 4K.

- YouTube: 1.1 Mb/s for 480p, 2.5 Mb/s for 720p, 5 Mb/s for 1080p, 20 Mb/s for 4K.

- We puts 4K without buffering at 20 to 25 Mb/s, which is a bit more conservative.

Put together, that's roughly 15 to 25 Mb/s for 4K depending on the platform, and a lot less for HD or Full HD. A single Full HD stream needs about 5 Mb/s on both Netflix and YouTube.

One caveat: Netflix says 4K can use up to 7 GB per hour. That's a maximum, which works out to about 15 Mb/s sustained. It isn't what every 4K stream pulls all the time.

YouTube also suggests using a wired connection rather than wireless and closing other apps while you watch.

Your turn:

- What download speed do you have, and on which platform?

- Does 15 to 25 Mb/s for 4K match what you see at home, or do you need more or less?

A new corrective update is available for the nPerf Web App Beta, focusing on what happens when your browser tab moves to the background during a test.

Fixes

• Switching tabs during a test: coming back to the nPerf tab no longer restarts the gauge loading animation. Your current measurement stays visible instead of temporarily disappearing while the gauge resets.
• Result details graph: the curve is now better framed, with more space above peaks and an axis that adapts to the measured data.
• Tooltips: they are no longer cut off when displayed near the edges of the graph.

You can test the latest version directly on the nPerf Web App Beta.

As always, feedback is welcome! 👀

Italy is considering a pretty interesting approach to its next spectrum renewals.

Instead of making operators pay again for frequencies expiring in 2029, regulator AGCOM is proposing to extend the licences until 2037 for free.

But there’s a catch: operators would have to commit to real network investments.

TIM, Fastweb-Vodafone and WindTre would face targets around 5G Standalone deployment, coverage, latency and network capacity, with progress checked over time.

There’s also a competition angle: the proposal would require the largest operators to provide wholesale access to Iliad and MVNOs, potentially including things like network slicing and spectrum sharing.

It’s quite a different philosophy from simply auctioning spectrum to whoever pays the most.

Less money spent on licences → more money invested directly into networks.

At least, that’s the idea.

What do you think about this approach?

Should access to spectrum be increasingly tied to technology investment and network performance, rather than how much an operator is willing to pay for it?

A new Web App Beta update is now available!

This release focuses on UI smoothness, gauge animations, and overall performance during tests.

🛠️ Fixes

• The speed gauge now matches the intended design, with a continuous arc, smoother scale transitions and dynamic zooming of the graduations for a more consistent experience across platforms.
• The test progress bar now stays perfectly synchronized with the graph, following the measurement engine's progression throughout the test.
• Fixed a 3D globe memory leak. Opening the server selection window repeatedly no longer leaves unused graphics contexts behind, preventing browser memory saturation. This continues the iOS/Safari optimization work introduced in version 3.0.5.

✨ Improvements

• Performance optimization: the interface is no longer fully rebuilt after every engine measurement (around 300 updates per test), resulting in a noticeably smoother experience while a test is running.

As always, thank you for testing the Web App Beta and sharing your feedback. Every report helps us improve the experience before the public release.

https://preview.redd.it/0xdwrhiuk4kh1.png?width=1600&format=png&auto=webp&s=16a7e5230a4c9365c4bdb223b4198fb6c8946a83

According to recent data, Bangladesh leads the region with an average fixed broadband download speed of 98.1 Mbps. That’s significantly faster than India (67.8 Mbps), Kazakhstan (66.1 Mbps), and Uzbekistan (62.2 Mbps), with Pakistan trailing at just 21.9 Mbps.

In practical terms, this means a 2 GB movie downloads in under 3 minutes in Bangladesh, compared to nearly 6 minutes in Pakistan. The gap is even more striking for smaller files, 1 GB takes just 1 minute and 20 seconds in Bangladesh, versus almost 6 minutes in Pakistan.

What’s driving this performance gap? Is it infrastructure, policy, or competition among ISPs? And for those of you in the region, does this match your real-world experience?

A new update is rolling out for the nPerf Web App Beta! 🚀

This release focuses on stability improvements, result reliability, and a few quality-of-life enhancements.

🛠️ Fixes

• Improved memory management on iOS / Safari for better stability. Session video recording has been removed, and the 3D globe is no longer loaded on mobile devices (instead of being loaded and simply hidden). More optimizations are planned in future updates.
• Fixed an issue where some test results could be flagged as bot traffic. The Cloudflare anti-bot token is now preserved correctly, ensuring valid results are submitted as expected.

✨ Improvements

• The test server sponsor logo is now clickable and opens the sponsor's website in a new tab.
• Connection type (ARCEP): the displayed technology now reflects the actual access technology (FTTH, xDSL, etc.) instead of the local connection between your device and the router. This continues the removal of the "Wi-Fi" / "Ethernet" labels.

As always, thank you to everyone testing the Web App Beta. Your feedback helps us improve every release!

First release note for the beta 👀

This is mostly a corrective update. If you had issues with geolocation or tests not starting, this should fix most of them. If you’re still running into problems, feel free to contact us.

Main changes:

• The test now always starts → before, ignoring the location prompt could leave the button doing nothing
• The “geolocation not enabled” message is now accurate → no more false alerts caused by timeouts or technical issues
• Location shown = location actually used → fixes wrong city display when using VPNs / proxies
• Translations fixed → no more weird “Start\nTest” formatting issues
• Speed gauge improved → better scaling at the start + smoother animation at the end

If you had issues before, curious to know if it’s working better now 👍

For those lucky enough to access the Web App beta, what are your first impressions?

https://preview.redd.it/gbfdmt095kdh1.png?width=1600&format=png&auto=webp&s=1bb16ba23e2c3bd9b484d16cee17025a7fc7186a

Algunos datos interesantes sobre redes fijas a principios de 2026.

Chile está claramente por delante en velocidad de subida, con un promedio de ~276 Mbps, por encima de Perú (~222 Mbps) y Colombia (~163 Mbps).

La diferencia se nota bastante en el uso real.

Subir un archivo de unos 2 GB en Chile puede tardar menos de un minuto, mientras que en otros países de la región puede acercarse a los 2 minutos… o incluso más.

Después, la caída es bastante clara:

• Argentina → ~119 Mbps
• Paraguay → ~55 Mbps
• Bolivia → ~48 Mbps

Dependiendo del país, los tiempos de subida pueden ser fácilmente 2 a 4 veces más largos.

Lo interesante es que la velocidad de subida suele pasar desapercibida frente a la de descarga.

Pero es clave para cosas como el teletrabajo, las copias en la nube, el envío de archivos grandes o las videollamadas.

Chile no está solo un poco por delante, está consistentemente arriba, incluso mientras otros países mejoran su infraestructura.

Tengo curiosidad:
¿realmente se nota esta diferencia en el día a día o la velocidad de subida sigue siendo algo que la mayoría no percibe?

Saw some data estimating download times for a game like GTA VI (~150 GiB).

Top countries based on fixed download speeds:

• France → ~462 Mbps → ~47 min
• Switzerland → ~455 Mbps → ~47 min
• Singapore → ~434 Mbps → ~50 min
• Slovenia → ~394 Mbps → ~55 min
• Canada → ~354 Mbps → ~1h01

So in the fastest countries, you’re looking at under 1 hour for a huge game.

What’s interesting is how fast the gap grows.

Same 150 GiB download:

• US → ~1h25
• UK → ~2h11

That’s more than double the time compared to the top countries.

And it gets even more noticeable with bigger files (200 GiB+), which is probably where things are going.

It’s a good reminder that for gaming, speed isn’t just about ping.

It directly affects:
preloads, updates, reinstalling games… basically how long you wait before actually playing.

Curious where people here are at.

How long does a ~100–150 GB download take for you right now?

https://preview.redd.it/l3z8pxo6y19h1.png?width=1600&format=png&auto=webp&s=38cce6fefeb7e048645d592aa7886c2744d79f7e

Slovakia comes out on top with ~81.7%, just ahead of Italy and France, both around 79%. The top of the ranking is actually quite tight, but things spread out quickly once you move down the list.

Some countries still sit in the high 70s, while others drop to the mid-60s, with Mexico at ~62.5%. That’s roughly a 20-point gap between the best and the lowest performers.

In practice, that difference is noticeable. Around 80%, streaming is smooth, even in HD or 4K. Closer to 65%, it still works, but buffering and loading times become more frequent, especially on higher quality.

What makes this interesting is that all these countries have similar GDP per capita. The difference isn’t really economic, it’s more about network infrastructure and how well it’s deployed.

Curious whether people actually feel that gap when streaming on mobile, or if it mostly shows up in measurements.