It's been a while since we had a VPN review on Meta Game Guides, so I'm pretty excited about this Atlas VPN review. To give you an overview of Atlas VPN, the company was founded in 2019 and entered the VPN market as a freemium service. In 2021, Nord Security (NordVPN, currently merged with Surfshark) acquired Atlas VPN, which I would call a brilliant M&A move to enter the medium-tier VPN market.
Nord already owns a significant portion of the market with NordVPN, even more so with the merger with Surfshark in 2021. Acquiring a relatively new VPN service to hold a significant market share of the low-cost VPN market makes perfect sense, especially since Atlas VPN “is trusted by more than 6 million users,” according to their about page.
Atlas VPN Overview
What you should expect from this review and what I expected as I had deep dine into AtlasVPN is a VPN service with all the features you would expect from a standard, modern VPN, but none of the bells and whistles you will find in premium VPN services.
The main advantage Atlas brings to the table is its untarnished reputation, compared to all the services mentioned earlier.
At the same time, Atlas comes with a few unique features (e.g., SafeSwap, the single-server equivalent of Surfshark's new feature, Nexus, and passwordless login).
So, what you should expect from Atlas VPN is decent all-around performance and affordability first and foremost. Let's see who Atlas performed under my tests then.
Atlas VPN Features
The table below showcases some exciting features available with Atlas VPN, but I would like to mention a few more:
- Ad and malware block, which is both useful and uncommon.
- WireGuard® protocol—Till 2021 you could only find WireGuard support only in a handful of medium to premium tier VPNs.
- Multihop+ offering for a combination of optimal latecny and speed along with privacy and security by smartly rotating VPN locations. I expect Multihop+ to only be available in a limited selection of countries that offer multiple locations.
- SafeSwap servers rotating user IPs within the same server as an extra privacy feature.
However, one of the first things that came to my mind was the number of servers in tandem with the number of users served by Atlas VPN. So I went through a round of several inquiries among VPN providers to check the correlation between the number of servers and active userbases:
- NordVPN: 10-14 million users, 5,300+ RAM only servers.
- ExpressVPN: Undisclosed amount of users (I would expect to be close to Nord's numbers), 3,000+ RAM only servers.
- Surfshark: Over 1 million users, 3,200+ RAM only servers.
- Ivacy: Over 2 million users, 3,500+ servers.
- CyberGhost: Over 15 million users, 6,100+ servers.
- PIA: Over 1 million users, 3,000+ servers.
- Atlas VPN: Over 6 million users, 750 servers.
Note: One of my affiliate contacts from Ivacy reported 50,000+ active worldwide users when I asked him, a further indication that the above numbers count the overall accounts number than the active accounts number for each VPN.
Based on these numbers, some of which may be totally made up, either AtlasVPN has access to server technology from the future, or they count inactive subscriptions towards those 6 million users.
Or they used a random seed to come with that number. However, assuming the number is valid, even if it counts inactive subscriptions and considers Atlas VPN's young age, it must contain many active subscriptions.
So before we proceed with the performance section of this review, I guess this is one of the slowest VPNs I have reviewed so far. Let's see!
|Number of countries (includes virtual servers)||33 (39 locations)|
|Number of servers||750|
|Log Policy||No logs|
|Ad and Malware Blocker||Yes|
|External Audits||Yes (VerSprite)|
|Number of devices per license||Unlimited|
|Money-back guarantee||30 days|
|24/7 Support||No & no live chat support|
|Tor/Onion over VPN||Yes (no extra features)|
|Headquarters Location||USA (5 Eyes Jurisdiction)|
|Average Download Speed Dropoff|
|Average Upload Speed Dropoff|
Atlas VPN Performance
As you will see in the video below, my tests got off to a rough start, as the Windows client kept disconnecting no matter what server I chose to connect to. That was an excellent opportunity to have a first chat with Atlas VPN support though.
Below you can see the issue at hand:
This brings us to another problem that needs I must emphasize at this point. Even with a premium subscription, Atlas VPN clients don't have access to live chat support. So after 30 minutes, I received the following response:
Essentially, Atlas VPN supports only two protocols on Windows:
In my case, switching from Auto to IKEv2 fixed the issue, but it left me with a single protocol option on Windows. Additionally, I prefer to review VPN services using next-gen protocols whenever available, so that one was a bummer.
The real bummer, though, was having to wait 30 minutes for a 10s answer, which then led to the follow-up “Why can't I use WireGuard on Windows? Why don't you support OpenVPN?” with the answer coming 5 minutes later.
Choosing to adopt a next-gen protocol without ensuring that you support the most popular and well-documented one does not seem wise to me.
Regardless, with the issue resolved, let's get down to the speed test results with IKEv2 enabled.
Speed Test results
From there on, I went through one of the most frustrating reviewing experiences of my life. And it had nothing to do with performance itself. However, after running an initial test on Romania Bucharest that showcased blazing fast speed and low latency, I made the switch to Madrid – Spain.
This caused Atlas VPN's client to get stuck trying to connect to a VPN server in Spain. As a result, I had to do the following two things: disable the Kill Switch (it was not on by default) to stop losing my connection while Atlas was hanging while trying to connect, close the VPN client and connect to my next desired location.
Now let's take a look at the numbers:
|Location||Ping [ms]||Download Speed [Mbps]||Download Speed Dropoff||Upload Speed [Mbps]||Upload Speed Dropoff|
|Baseline – Athens, Greece, no VPN||7||77.17||N/A||10.24||N/A|
|Greece – Athens||N/A||N/A||N/A||N/A||N/A|
|Romania – Bucharest||24||72.35||6.25%||9.54||6.84%|
|Spain – Madrid||72||71.35||7.54%||9.2||10.16%|
|France – Marseille||54||72.17||6.48%||9.52||7.03%|
|UK – London||Stuck||Stuck||Stuck||Stuck||Stuck|
|Netherlands – Amsterdam||73||72.23||6.40%||8.52||16.80%|
|Germany – Frankfurt||61||72.06||6.62%||9.65||5.76%|
|USA – New York||118||72.2||6.44%||8.98||12.30%|
|USA – Los Angeles||183||70.42||8.75%||8.72||14.84%|
|Canada – Toronto||128||71.63||7.18%||8.64||15.63%|
|Australia – Melbourne||293||67.32||12.76%||9.1||11.13%|
|Average Download Dropoff||9.92%||Average Upload Dropoff||14.07%|
The first important observation here is that the dropoffs I experienced put Atlas VPN in NordVPN, Surfshark, and ExpressVPN. It almost feels like Atlas VPN has leased or borrowed a fraction of NordVPN's infrastructure to achieve impressive performance.
I don't know if the above is true, but if it is done in an isolated environment, I see nothing wrong with Atlas VPN customers experiencing a glimpse of Nord.
But that's about the only positive comment I can make about Atlas VPN so far.
Reviewing AtlasVPN has been like someone asking you to test a software product in alpha, rather than a finished product that's been on the market for 3 years.
Based on what I witnessed in this review, I can say the following about Atlas's Windows client.
- I would not trust it for location switches as I there is something definitely wrong with disconnecting from one server and connecting to another. It raises questions about the digital footprint your activity leads during the switch.
- I find it hard to believe that the Kill Switch operates as it should when I have to close the client to connect to a new location. There is real possibility though that this is not an automatic Kill Switch but a manual way to pull the plug and shut down any ongoing connections.
- Leaving the client open for several hours really slowed down my PC. Memory leaks?
- I could not evaluate the tracker and ad blocker as it required the WireGuard protocol, which did not work on my system.
- WireGuard support seems like a marketing decision that backfired.
Even if my issues are just an outlier, this was a first for me as a reviewer, and even at this low price tag, there are far more robust solutions than AtlasVPN.
I would go as far as saying that maybe Atlas VPN should consider switching to Nord's client, changing the logo, stripping down the bells and whistles not supported by Atlas, and delivering a higher quality experience to the users.
Ease of Use
From a user experience standpoint, Atlas VPN's client is one of the most straightforward you will encounter. I would say that it is a bit too simplistic due to the lack of premium options.
You will find three tabs on the server selection screen: all servers, streaming optimized, and privacy pro. The last one contains the servers using SafeSwap or Multihop+ technology. Considering the limited selection of locations and servers, I would love to see latency and load stats for each server, but sadly there is no such option.
Client App – Servers Tab
Contains all available servers and locations.
Contains streaming optimized servers with SmartDNS.
Contains MultiHop+ and SafeSwap servers.
Switching protocols is also as simple as it can get in the Settings section (Protocols tab):
Note that activating the Track Blocker is not an option while connected under IKEv2. You will have to disconnect and make the switch:
I would rate Atlas's ease of use very high if everything worked as intended, but everything went the other way around in this review. My actual experience did not come close to the product's designated intention. Having to meddle around with the app's settings to get it to work is disappointing as a reviewer.
Atlas VPN Safety, Leaks & Encryption
First, enabling the kill switch would completely shut down any data transfers instead of being triggered only when there is a VPN connection interruption. As a result, I proceeded with the DNS, IPv6, and WebRTC leak evaluation with the Kill Switch disabled.
The results were alarming on multiple tools, but here is a great summary:
Here is a rundown of the test results:
- There are definitely leaks of Greek DNS servers while I'm connected to Romania so there is a DNS in place.
- The IP address 184.108.40.206 is indeed a Romanian one, but it has been associated with many US Atlas VPN users that it is flagged as US-used IP.
- There is also an IPv6 leak in place.
Here is what I found out about the Romanian IP address used:
The IP 220.127.116.11 is tied to Atlas VPN, so it is probably blacklisted on several websites. The IP itself is owned by a data center owned by M247 Ltd in Bucharest, Romania. VPN providers, especially lower-tier ones, are known to rent data centers to run their operations rather than have their own network, and that is the case with Atlas too.
What is alarming here is that the use of this IP has been observed in Noblestown US, a clear sign that the leak has occurred too many times to associate an IP Romania with usage from a US location.
While Atlas VPN follows modern encryption standards, it is pretty clear that its infrastructure and client bugs deem it a VPN service that I would not advocate for if you a looking for a high level of privacy and security.
Atlas VPN Streaming
I don't think stream unblocking is a selling point for VPNs anymore. Nowadays, streaming platforms associate the client's billing information with a specific distribution area. For instance, I cannot access the US Netflix library using a VPN anymore. I wouldn't be able to even if I was in the US physically, as my Netflix account is associated with the Greek library.
The primary use case of VPNs, in this case, is limited to those traveling abroad. If you are in a different country and want to access your original library, you must use a VPN. Otherwise, you will be stuck with the Netflix Originals library.
The obvious concern here is that with so many leaks in place, it will be effortless for the streaming service to identify your physical location and redirect you to a non-violating version of its movie & TV show catalog.
Regardless, I had no issue accessing the following service during this review:
- BBC iPlayer
- Sky UK
- Netflix UK
- Disney+ UK
- HBO Max
- Netflix US
- Disney+ US
However, while I was able to unblock the above services, the DNS leak was ongoing, so I cannot predict the long-term effects.
VPN providers typically develop browser extensions to isolate the VPN's execution and avoid leaks, but Atlas only offers OS-level applications. This is a strategy I would endorse if it were functioning 100% as it should, but in this case, it's far from the truth.
Overall, I would say that Atlas VPN can handle unblocking pretty well; what I'm questioning here is the consistency of its success when there is a leak present.
No logs Policy and Audit
Atlas VPN undoubtedly has an unambiguous no-logs policy and doesn't even store encrypted passwords thanks to its 2FA-driven approach towards the login process. Despite complying with US laws, the stored data is very little, even in the case of a potential disclosure to US authorities.
As far as the independent audit, I believe that I found more issues in a handful of days than VerSprite did during their audit, so I cannot read too much into it right now.
While the performance wasn't nearly as impressive as the Windows app, the overall UX was much better on Android. No unresponsiveness, no lags, no memory issues, and slowdowns.
Here are some performance measurements:
- US (New Jersey): Ping: 122ms, Download Speed: 25.3Mbps,Upload Speed: 9.75Mbps, Jitter: 2ms, Loss: 0.7%.
- UK (London): Ping: 56ms, Donwload Speed: 21.5Mbps, Upload Speed: 9.67Mbps, Jitter 3ms, Loss: 0.0%.
- Australia (Sydeny): Ping: 294ms, Download Speed: 24.7Mbps, Upload Speed: 9.70Mbps, Jitter: 18ms, Loss: 1.7%.
This time, I used WireGuard and the Kill Switch without issues.
Atlas VPN Pricing
Atlas VPN comes with three pricing plans:
- Monthly: $10.99/month
- Yearly: $3.29/month
- 3-year: $1.99/month
There is one key observation to make here—Atlas VPN's pricing is too close to Surfshark's for no reason at all, based on what I experienced.
For comparison, here are Surfshark's pricing plans:
- Monthly: $12.95/month
- Yearly: $3.99/month
- 2-year: $2.49/month
The price tag is so similar on the yearly plan that I would rather spend an extra $0.70*12, or $8.14, for a much smoother and premium experience than the one I go from Atlas VPN.
At the same time, it's worth mentioning Ivacy here, as it's a mid-tier VPN that does most things well without shining in any department, although its performance is pretty solid.
- Monthly: $9.95/month
- Yearly: $3.99/month
- 5-year: $1.19/month
While Ivacy is also a big no from me on its monthly and yearly plan, I would much rather spend $71.64 for five years with the possibility of eventually regretting rather than spend $71.49 for three years with the possibility of eventually regretting it.
Should you buy Atlas VPN
I believe the answer to this question is evident throughout my whole review of Atlas. No. Not by any means in its current shape and form. There is promise in terms of features. The mobile experience was on point, but I experienced on Windows more about tinkering and troubleshooting than reviewing a final product.
I really hope Atlas VPN will pay attention to this review and develop significant improvements to the Windows client. Until then, I cannot advocate for a product that is essentially broken on one of the most popular operating systems. I liked the performance, I think SafeSwap and Multihop+ are interesting, but there is something inherently wrong with the native Windows client.
At the same time, I hope that Atlas VPN will contact me to re-evaluate their product after they are confident they have fixed its shortcomings, as this was a painful reviewing process.
While I called the acquisition of Atlas VPN a brilliant move at the start of this review, Atlas feels more like a stain to Nord's reputation in my eyes after this review. I guess even the best can get it wrong.