Short version: yes, AA wifi has been weirdly unreliable lately, and not just on the routes others mentioned.
A few angles that complement what @sognonotturno and @nachtdromer already covered:
1. It is not always “congestion” or “hardware”
They are right that load and aircraft type matter, but I have had flights where:
- Cabin was half empty
- Portal said “good” connection
- Basic sites like Fastmail and GitHub kept timing out while Speedtest showed 5–8 Mbps
That smells more like aggressive traffic shaping and flaky DNS than pure capacity. In particular, long‑lived connections such as SSH, RDP, certain video platforms, and some gaming protocols get nuked while casual browsing still works. So if your use case leans on persistent sessions, AA wifi will feel reliably bad even when raw throughput is OK.
2. Do not overestimate “monthly pass spreads the risk”
I slightly disagree with the idea that a monthly pass makes the pain go away. It only really helps if:
- You fly AA a lot, and
- Most of those flights are on the same provider mix covered by that pass
If your travel pattern is mixed (AA plus others, or lots of short hops) you can end up with a month of terrible value instead of one annoying but contained flight pass. In that case, sticking to single‑flight purchases and aggressively requesting refunds can actually be more rational.
3. Diagnose at home so you know what is not your fault
You cannot fix AA’s satellite or their backend, but you can remove variables on your side before you ever board:
- Test how your laptop behaves when roaming between access points, how it handles captive portals, and whether it recovers well from brief drops.
- Tools like NetSpot are handy for this at home or the office: they let you visualize signal strength, channel overlap, and roaming behavior.
Pros of NetSpot in this context:
- Clear heatmaps so you can see where your own wifi is actually bad
- Good for uncovering driver quirks and sticky roaming tendencies on your device
- Helps you standardize your wifi settings so you are not debugging your laptop at 35,000 feet
Cons:
- It will not change anything about AA’s satellite link or traffic shaping
- There is a learning curve if you dig into the survey features
- Some advanced options are overkill if you just want a quick yes/no check
The point is not that NetSpot magically fixes inflight wifi. It just means that when AA’s network acts up, you can be far more confident the fault is not your own configuration.
4. Treat “portal works, internet does not” as a distinct failure mode
One thing I see a lot on AA:
- Portal loads fine
- Messaging or cached pages seem OK for a bit
- Anything that needs fresh DNS resolution just dies
When that happens, classic tips like toggling airplane mode or forgetting the network do not always help. A slightly different approach:
- Try a low‑overhead site first, like a simple text‑heavy news page
- If the portal still responds but outside sites stall for more than 10–15 minutes, mentally write off that flight’s connectivity instead of spending half the leg rebooting devices
At that point you are fighting their infrastructure, not your laptop.
5. Plan connectivity as “tiers” of reliability
Instead of thinking “wifi works” or “wifi does not,” I plan AA flights in tiers:
- Tier 1: Email, chat, text‑only web. Usually OK.
- Tier 2: File sync, large attachments, code pulls. Works maybe half the time.
- Tier 3: Real time stuff like Zoom, RDP, cloud IDEs, trading dashboards. Assume it will fail.
If your critical task lives in tier 3, do what @nachtdromer hinted at and build a backup that does not rely on AA at all, such as phone dial in or a plan to do async updates instead of live calls.
6. When you complain, separate device issues from network issues
To add to their refund tips, I have had more success when I clearly differentiate:
- “Signal / portal kept dropping” versus
- “Signal and portal fine, but internet unusable for X minutes / Y portion of flight”
The latter points directly at AA’s network stack, which is harder for them to pin on passenger devices. Attach one or two concise screenshots if you have them, but do not flood them with logs.
So yes, what you are seeing is common, and it is not just you. Between what @sognonotturno and @nachtdromer shared and some basic baseline testing with something like NetSpot on your own gear, you can at least reclaim some sanity and know when it is time to stop troubleshooting and just switch to offline work.