VPN Speed Test Guide 2026: How to Test & Optimize Your Connection

“My VPN is slow.” It is the most common VPN complaint — and in 90% of cases, the problem is not the VPN itself. It is how you are testing, which server you are connected to, or which protocol you are using. A poorly configured fast VPN will always lose to a well-configured mediocre one.

This guide teaches you how to accurately test VPN speed, understand what the numbers actually mean, identify what is causing slow performance, and optimize your connection for maximum speed. Whether you are troubleshooting a sluggish connection or evaluating which VPN to buy, these are the same testing methods we use for our VPN speed test results — the methodology behind 500+ speed tests across 10 server locations.

The VPNs We Recommend (Based on Speed Testing)

Before diving into methodology, here are the three VPNs that consistently deliver the best speed performance in our testing:

---

How VPN Speed Works: The 60-Second Explanation

When you connect to a VPN, three things happen that affect your speed:

1. Encryption overhead. Your data gets encrypted before leaving your device and decrypted at the VPN server. Modern protocols (WireGuard) do this extremely efficiently — adding only 3-8% overhead. Older protocols (OpenVPN) add 15-25%.

2. Server routing. Instead of traveling directly from your device to the website, your data routes through the VPN server first. If that server is nearby (same country), the detour adds 5-20ms of latency. If it is across an ocean, the detour adds 80-200ms.

3. Server load. VPN servers are shared resources. If 10,000 users are connected to the same server, each gets less bandwidth. Premium VPNs mitigate this with large server fleets and dynamic load balancing. Cheap or free VPNs pack users onto fewer servers.

---

Step-by-Step: How to Accurately Test Your VPN Speed

Most people test VPN speed wrong. They run a single speed test, see a number, and draw conclusions. That is like checking the weather once and declaring it always rains. Here is the methodology we use for consistent, reliable results.

Step 1: Establish Your Baseline (Without VPN)

Before testing your VPN, you need to know your actual internet speed without it.

1. Disconnect from your VPN completely. Not just disconnected from a server — fully quit the VPN application. Some VPNs run background processes that can affect speeds even when “disconnected.”
2. Close bandwidth-consuming applications. Quit streaming, cloud syncing, large downloads, and other devices on your network doing the same.
3. Run 3 speed tests on Speedtest.net and Fast.com. Record download, upload, and latency (ping) for each test.
4. Average your results. Your baseline is the average of these 6 tests (3 from each platform).

Why use two testing platforms? Speedtest.net and Fast.com use different server networks and testing methodologies. Averaging across both gives you a more accurate baseline and helps identify if one platform’s test servers are temporarily congested.

Example baseline:

Test	Download	Upload	Latency
Speedtest #1	247 Mbps	42 Mbps	8 ms
Speedtest #2	253 Mbps	45 Mbps	7 ms
Speedtest #3	251 Mbps	43 Mbps	8 ms
Fast.com #1	240 Mbps	40 Mbps	9 ms
Fast.com #2	245 Mbps	41 Mbps	8 ms
Fast.com #3	248 Mbps	42 Mbps	9 ms
Average	247 Mbps	42 Mbps	8 ms

Step 2: Connect to Your VPN (Nearest Server, Best Protocol)

1. Open your VPN app and connect to the nearest server — ideally in your own country or a neighboring country. This minimizes the routing distance penalty.
2. Select the fastest protocol. Use WireGuard (Surfshark, Proton VPN) or NordLynx (NordVPN). If your VPN does not offer WireGuard, use IKEv2. Avoid OpenVPN for speed testing — it is significantly slower.
3. Wait 10-15 seconds after connecting before testing. The connection needs a moment to stabilize.

Step 3: Run VPN Speed Tests

Run the same 6-test battery you ran for your baseline:

• 3 tests on Speedtest.net
• 3 tests on Fast.com

Important: On Speedtest.net, manually select a test server near your physical location, not near the VPN server. By default, Speedtest auto-selects the closest server to your apparent IP (the VPN server’s location), which can give misleading results.

Step 4: Calculate Your Speed Retention

Speed retention percentage = (VPN speed / Baseline speed) x 100

Example with NordVPN (NordLynx, nearest server):

Test	Download	Upload	Latency
VPN Average	228 Mbps	38 Mbps	14 ms
Baseline	247 Mbps	42 Mbps	8 ms
Retention	92.3%	90.5%	+6 ms

Interpreting the results:

• 90-95% retention: Excellent. This VPN is barely noticeable. Top-tier performance.
• 80-90% retention: Good. Suitable for everything including 4K streaming and video calls.
• 70-80% retention: Acceptable. Fine for browsing and standard streaming, may struggle with latency-sensitive tasks.
• Below 70% retention: Poor. Either the server is too distant, the protocol is outdated, or the VPN provider has congestion issues.

Step 5: Test Multiple Server Locations

Repeat Steps 2-4 for at least 3 additional server locations:

• Same country, different server (tests server load variation)
• Neighboring country (tests short-distance routing)
• Different continent (tests long-distance performance)

This gives you a complete performance profile rather than a single data point.

---

What Affects VPN Speed: The 7 Factors

Understanding what controls VPN speed helps you diagnose problems and optimize performance. Here are the seven factors, ranked by impact.

1. VPN Protocol (Highest Impact)

The protocol is the single biggest controllable factor in VPN speed. Here is what our testing shows:

The takeaway: WireGuard-based protocols are 25-40% faster than OpenVPN across all VPNs tested. If your VPN is slow, the first thing to check is your protocol setting.

How to change protocols:

• NordVPN: Settings > VPN Protocol > NordLynx
• Surfshark: Settings > VPN Protocol > WireGuard
• Proton VPN: Settings > Protocol > WireGuard

2. Server Distance (High Impact)

Physics is undefeated. Data traveling across the Atlantic takes longer than data traveling across town.

Server Distance	Typical Retention	Latency Added	Example
Same city / nearby	88-95%	+3-10 ms	New York → New Jersey
Same country	85-93%	+5-20 ms	New York → Los Angeles
Neighboring country	80-90%	+10-30 ms	US → Canada
Same continent (far)	70-85%	+20-60 ms	US → UK
Different continent	50-70%	+80-200 ms	US → Japan

The takeaway: Always connect to the nearest server unless you specifically need an IP address in another country (for streaming geo-restricted content, for example).

3. Base Connection Speed (Medium Impact)

Your VPN cannot make your internet faster than it already is. But interestingly, faster base connections tend to show higher percentage retention because the VPN’s fixed overhead (encryption processing time) becomes a smaller fraction of total throughput.

Base Speed	Typical VPN Speed (WireGuard, nearby)	Retention
25 Mbps	20-23 Mbps	80-92%
100 Mbps	85-93 Mbps	85-93%
250 Mbps	218-238 Mbps	87-95%
500 Mbps	435-475 Mbps	87-95%
1 Gbps	800-920 Mbps	80-92%

The takeaway: If your base connection is under 50 Mbps, you will notice VPN overhead more. If your base is 100+ Mbps, modern VPNs add nearly imperceptible drag.

4. Server Load (Medium Impact)

VPN servers have capacity limits. When too many users connect to the same server, speeds drop for everyone. Premium VPNs mitigate this with massive server fleets:

• NordVPN : 6,400+ servers in 111 countries
• Surfshark : 3,200+ servers in 100 countries
• Proton VPN : 4,600+ servers in 110+ countries

How to check server load: NordVPN and Surfshark both display server load percentages in their apps. Choose servers with load under 50% when possible. Proton VPN does not show load explicitly but auto-selects the fastest available server.

Peak hours matter. Server load increases dramatically during evening hours (7-11 PM) in any timezone. If you are testing VPN speed, test during off-peak hours for best results, or test during peak hours if you want to know worst-case performance.

5. Your Device (Low-Medium Impact)

VPN encryption requires CPU processing. On modern laptops and flagship phones, this is negligible. On older devices, budget phones, or IoT devices, encryption processing can bottleneck your speed.

Devices that handle VPN encryption well: Any laptop from the last 5 years, flagship phones (iPhone 12+, Samsung S21+, Pixel 6+), gaming consoles (via router).

Devices that may bottleneck: Budget phones (under $200), older tablets, smart TVs (use VPN on your router instead), Raspberry Pi or similar.

6. ISP Throttling (Variable Impact)

Some ISPs actively throttle VPN traffic by detecting VPN protocol signatures and reducing bandwidth. This is more common in countries with internet censorship (China, Russia, Iran) but also occurs with some ISPs in unrestricted countries.

Signs of ISP throttling:

• VPN speeds are significantly slower than expected based on your base speed
• Speeds are worse during certain times of day
• Different VPN protocols perform very differently (obfuscated protocols work better)

How to test for throttling: Run speed tests on multiple protocols. If WireGuard performs significantly worse than OpenVPN (which is unusual — WireGuard should be faster), your ISP may be identifying and throttling WireGuard traffic. Try NordVPN’s obfuscated servers or Surfshark’s NoBorders mode.

7. Network Conditions (Variable Impact)

WiFi interference, network congestion at your ISP, international backbone congestion, and even weather can affect speeds. These factors exist with or without a VPN and are outside your control.

Mitigation: Test on ethernet when possible (eliminates WiFi variables). Test at multiple times of day. If speeds are slow, retest in an hour before blaming your VPN.

---

How to Optimize Your VPN for Maximum Speed

Based on our testing, here are the optimization steps in order of impact. Each step alone can significantly improve your VPN speed, but the combination delivers the best results.

Optimization 1: Switch to WireGuard / NordLynx Protocol

Expected improvement: 20-40% faster than OpenVPN

This is the single most impactful change most people can make. If your VPN is set to “Automatic” protocol selection, it may be choosing OpenVPN for compatibility reasons rather than WireGuard for speed.

NordVPN users: Switch to NordLynx (NordVPN’s customized WireGuard implementation). Go to Settings > VPN Protocol > NordLynx. NordLynx adds a double NAT system for privacy while maintaining WireGuard’s speed advantage.

Surfshark users: Switch to WireGuard. Settings > VPN Protocol > WireGuard.

Proton VPN users: Switch to WireGuard. Settings > Protocol > WireGuard.

Optimization 2: Connect to the Nearest Server

Expected improvement: 10-40% faster than connecting to a distant server

Most VPN apps have a “Quick Connect” or “Fastest Server” button that auto-selects the nearest, least-loaded server. Use it. Manually selecting a server in another country because you think it might be “faster” almost never works — proximity is king.

Exception: If you need to appear to be in a specific country (for streaming or accessing geo-restricted content), connect to the nearest server in that country, not just the nearest server globally.

Optimization 3: Use Ethernet Instead of WiFi

Expected improvement: 5-20% more consistent (reduces variability)

WiFi adds its own layer of latency and packet loss. Running your VPN over an ethernet connection eliminates WiFi as a variable and gives you the clearest picture of actual VPN performance. If you are troubleshooting slow VPN speeds, plug in via ethernet first to rule out WiFi issues.

Optimization 4: Enable Split Tunneling

Expected improvement: Faster non-VPN traffic, reduced VPN load

Split tunneling routes only specific apps or websites through the VPN, letting everything else bypass it. This reduces the load on the VPN tunnel and eliminates speed overhead for traffic that does not need encryption.

Use cases for remote workers:

• Route work apps (Slack, email, company portals) through VPN for security
• Let streaming, gaming, and personal browsing bypass the VPN for full speed
• Route video calls outside VPN if latency is causing issues (only on trusted networks)

All three providers support split tunneling:

• NordVPN: Settings > Split Tunneling (app-based)
• Surfshark: Settings > Bypasser (app-based and website-based)
• Proton VPN: Settings > Split Tunneling (app-based)

Optimization 5: Try a Different Server in the Same Region

Expected improvement: Variable (0-30% if server was overloaded)

If your nearest server is slow, try the second or third nearest. Server load varies by time of day and the server you happened to connect to may be at capacity. Switching to a different server in the same city or country often resolves speed issues immediately.

Optimization 6: Restart Your VPN Connection

Expected improvement: Variable (resolves stale connection issues)

VPN connections can degrade over long sessions. If you have been connected for hours and speeds have dropped, disconnect and reconnect. This forces a fresh handshake, potentially connecting you to a less-loaded server and re-establishing optimal routing.

Optimization 7: Update Your VPN App

Expected improvement: Variable (newer versions often have performance improvements)

VPN providers regularly release updates that improve connection speed, fix bugs, and optimize protocol implementations. Running an outdated app may mean missing performance improvements that have already been released.

---

Common VPN Speed Myths (Debunked)

Myth: “Free VPNs are just as fast as paid ones”

Reality: Free VPNs almost universally throttle speed to push users toward paid plans. They also run fewer servers, meaning higher congestion. Proton VPN ’s free tier is the only reputable exception — and even it limits you to servers in 3 countries with speed caps.

Myth: “More servers = faster VPN”

Reality: Server count matters less than server quality and distribution. A VPN with 1,000 high-capacity 10Gbps servers will outperform one with 10,000 budget 1Gbps servers. That said, having servers in more locations does mean you are more likely to find one near you.

Myth: “A VPN can speed up my internet”

Reality: A VPN can only restore speeds that your ISP is artificially throttling. If your ISP throttles Netflix traffic from 100 Mbps to 10 Mbps, a VPN hides that traffic and restores the full 100 Mbps. But if your ISP honestly delivers 100 Mbps, a VPN will deliver slightly less (85-95 Mbps with WireGuard). A VPN cannot create bandwidth that does not exist.

Myth: “The VPN server nearest to the website is fastest”

Reality: The fastest path is: Your device → nearest VPN server → website. The proximity that matters most is between you and the VPN server, not between the VPN server and the website. Websites use CDNs that serve content from edge locations globally, so the website’s distance from your VPN server rarely matters.

Myth: “Double VPN / multi-hop is twice as slow”

Reality: Multi-hop routes your traffic through two VPN servers instead of one, but the speed penalty is typically only 10-20% additional — not 50%. Both NordVPN and Proton VPN offer multi-hop features. In our testing, NordVPN’s Double VPN retained about 75-80% of base speed versus 88-95% for single-server connections.

---

Speed Requirements by Activity

Not all internet activities need the same speed. Use this table to determine whether your VPN-connected speed is adequate for your needs:

Activity	Minimum Speed	Comfortable Speed	VPN Impact
Web browsing	5 Mbps	15+ Mbps	Imperceptible
Email and messaging	1 Mbps	5+ Mbps	Imperceptible
SD video streaming	3 Mbps	8+ Mbps	Imperceptible
HD video streaming	5 Mbps	15+ Mbps	Minimal
4K video streaming	25 Mbps	40+ Mbps	Minimal
Zoom / video calls (HD)	3.8 Mbps	10+ Mbps	Minimal
Zoom / video calls (group)	8 Mbps	20+ Mbps	Low
File uploads (50-200MB)	10 Mbps	25+ Mbps	Low
Online gaming	5 Mbps (low latency critical)	15+ Mbps	Moderate (latency)
Large downloads (1GB+)	25 Mbps	100+ Mbps	Low
Torrenting	10 Mbps	50+ Mbps	Low-Moderate

The practical takeaway: If your base connection is 50+ Mbps and you use a modern VPN with WireGuard protocol on a nearby server, your VPN-connected speed will exceed the “comfortable” threshold for every activity in this table. VPN speed is a concern primarily for users with slower base connections (under 25 Mbps) or those connecting to very distant servers.

---

VPN Speed Test Checklist

Use this checklist every time you test VPN speed for consistent, comparable results:

Pre-test preparation:

• Close all bandwidth-consuming applications
• Disconnect other devices from the network (if possible)
• Use ethernet instead of WiFi (if available)
• Fully quit any VPN application before baseline tests

Baseline testing (VPN off):

• Run 3 tests on Speedtest.net (record download, upload, latency)
• Run 3 tests on Fast.com (record download, upload, latency)
• Average all 6 results for your baseline numbers

VPN testing:

• Connect to nearest server using WireGuard / NordLynx
• Wait 10-15 seconds for connection to stabilize
• Run 3 tests on Speedtest.net (manually select server near your location)
• Run 3 tests on Fast.com
• Average all 6 VPN results

Analysis:

• Calculate retention: (VPN average / Baseline average) x 100
• 90%+ retention = Excellent
• 80-90% retention = Good
• 70-80% retention = Acceptable
• Below 70% = Investigate (wrong protocol? distant server? ISP throttling?)

Optional additional tests:

• Repeat with a server in a different country
• Repeat with OpenVPN to compare protocol performance
• Test at different times of day (morning, afternoon, evening)

---

Our Speed Testing Results (Summary)

For our complete testing data with 500+ speed tests across 10 locations, see our detailed VPN speed test results 2026 article. Here is the high-level summary:

Speed rankings (nearby server, WireGuard):

1. NordVPN — 91% average retention. NordLynx protocol gives it a measurable edge.
2. Surfshark — 86% average retention. Best speed-per-dollar at $2.19/month.
3. Proton VPN — 78% average retention. Trades some speed for verifiable Swiss privacy and open-source transparency.

---

Troubleshooting Slow VPN Speeds

If your VPN speed is below expectations after following the optimization steps above, work through this diagnostic checklist:

Is your base connection slow?

Test without VPN first. If your base speed is under 25 Mbps, VPN overhead will be more noticeable. Consider upgrading your internet plan or finding a better WiFi connection before blaming the VPN.

Are you using OpenVPN?

Check your protocol setting. OpenVPN is 25-40% slower than WireGuard in our testing. Switch to WireGuard/NordLynx immediately.

Are you connected to a distant server?

Check which server you are connected to. If you are in the US connected to a server in Japan, speeds will be 30-50% lower than connecting to a US server. Switch to the nearest server.

Is the server overloaded?

Try a different server in the same region. If speeds improve significantly, the original server was congested. NordVPN shows server load percentages — aim for servers under 50% load.

Is your ISP throttling VPN traffic?

Try these steps in order:

1. Switch VPN protocols (if WireGuard is throttled, try IKEv2 or OpenVPN)
2. Try NordVPN’s obfuscated servers (designed to bypass VPN detection)
3. Try Surfshark’s NoBorders mode
4. If nothing works, contact your ISP — they may have VPN-specific throttling policies

Is your device the bottleneck?

On older or budget devices, VPN encryption can max out CPU capacity. Check CPU usage while connected to VPN. If it is at 90%+, your device is the bottleneck. Consider running the VPN on your router instead, which offloads encryption from your device.

Is it a DNS issue?

Slow DNS resolution can make everything feel sluggish even if raw throughput is fine. Use your VPN’s built-in DNS (all three providers offer this) or configure a fast public DNS (Cloudflare 1.1.1.1 or Google 8.8.8.8).

---

When VPN Speed Does Not Matter

It is worth noting that for many common activities, the speed difference between VPN-on and VPN-off is completely imperceptible:

• Web browsing: A page that loads in 0.5 seconds without VPN loads in 0.55 seconds with VPN. You will not notice.
• Email: Sending and receiving email is unaffected by a 5-15% speed reduction.
• Messaging (Slack, WhatsApp, Telegram): Text messages are kilobytes. VPN overhead is measured in milliseconds.
• Standard video calls: Zoom HD requires 3.8 Mbps. Even a 50 Mbps connection with 85% VPN retention delivers 42.5 Mbps — 11x what Zoom needs.
• Music streaming: Spotify uses 0.32 Mbps for high quality. Any VPN handles this effortlessly.

VPN speed becomes a meaningful consideration only for: 4K streaming on slower connections (under 40 Mbps base), large file transfers (where time savings are measurable), online gaming (where latency matters more than throughput), and torrenting (where sustained throughput affects download times).

---

Pros and Cons of VPN Speed Impact

---

Final Recommendations

If speed is your top priority: NordVPN with NordLynx protocol delivers the fastest VPN speeds we have tested — 91% average retention across all locations. At $3.39/month on the 2-year plan, it is a premium worth paying for consistently fast performance. Read our full NordVPN review.

If you want the best value: Surfshark at $2.19/month delivers 86% speed retention with unlimited device connections. The speed-per-dollar ratio is the best in the industry. Read our full Surfshark review.

If privacy matters more than raw speed: Proton VPN trades some speed (78% retention) for Swiss jurisdiction, open-source apps, and Secure Core multi-hop routing. If verifiable privacy is non-negotiable, the speed tradeoff is modest. Read our full Proton VPN review.

For all three: Use WireGuard/NordLynx protocol, connect to the nearest server, and you will experience speeds that are fast enough for any task short of competitive gaming at the highest level.

For detailed speed data across 10 server locations, see our VPN speed test results 2026. For help choosing between these providers head-to-head, see our NordVPN vs Surfshark comparison. And for the best VPN setup for travel, explore our best VPN for travel guide.