Education · Guide · 10 min read
Are free VPNs safe in 2026? What the data shows
A research-based look at the risks, the exceptions, and what to check before installing one. Built on peer-reviewed studies and regulatory records, not marketing.

Most free VPNs are not safe. Independent research shows the bulk of free apps carry trackers, resell data, or ship malware. A handful of credible exceptions exist, and this guide separates the two.
The short answer
Most free VPNs are not safe. Independent research consistently shows that free VPNs collect more data than paid ones, embed tracking libraries, sell user traffic to data brokers, or in the worst cases, ship outright malware. The 2017 CSIRO Australia study tested 283 free VPN Android apps and found that 38% contained malware or malvertising and 75% embedded third-party tracking libraries. That risk profile has not improved in 2026.
There are exceptions. A handful of credible providers (Proton VPN, Windscribe, hide.me, TunnelBear) offer free tiers as marketing for their paid plans. These free tiers are not the same product as the bulk of free VPN apps on the Play Store or App Store.
The short version: avoid free VPN apps from unknown developers. If you need a free tier, use one of the four credible providers above and understand its limits.
Our research methodology
This guide synthesizes peer-reviewed academic research, audit reports, regulatory complaints, app store policy changes, and aggregated user feedback. We do not run our own VPN tests. Read our research methodology for the criteria and weighting we apply. Specific claims about apps, providers, and data practices are sourced inline to peer-reviewed papers, court filings, or vendor documentation.
What "free" actually means in the VPN business
A VPN runs on a network of servers in dozens of countries. Each server costs money: hardware, bandwidth, maintenance, security audits, customer support, legal compliance. A paid VPN charges users to fund that infrastructure.
A free VPN does not. It still pays the same costs. The question is: where does the money come from?
The answer falls into four categories, all documented in independent research and regulatory action over the past decade. The short summary: someone is paying for your free VPN. If it is not you, it is most likely your data, your bandwidth, or your attention.
This is not a generalization. It is the actual business model of free VPNs as documented by researchers at ICSI Berkeley, CSIRO Australia, the Federal Trade Commission in the United States, and consumer protection agencies in the European Union. We cite the specific studies and cases in the sections below.
This is also why a free VPN from an established paid provider behaves differently. They subsidize the free tier from paid revenue. The free user is not the product.
The four ways free VPNs make money
There are four documented monetization models for free VPNs. A given free app may use one or several at once.

1. Data harvesting and resale. Free apps log user activity (sites visited, apps used, location, device IDs) and sell this aggregated data to advertisers, data brokers, or market research firms. The 2017 CSIRO study found that 75% of the 283 free VPN Android apps tested contained third-party tracking libraries, undermining the privacy claim that motivated the user to install a VPN in the first place. The 2019 Top10VPN free VPN risk index found 72% of free apps contained third-party trackers, often the same Facebook and Google analytics libraries the user was trying to escape.
2. Bandwidth resale (residential proxy networks). Some free VPNs reroute paying customers' traffic through free users' devices, turning the free user into an exit node for paid clients. Hola VPN's 2015 case is the most cited example: free user devices were sold as proxies, including for botnet attacks. The free user typically does not know this is happening and cannot detect it from the app interface alone.
3. Aggressive advertising and malvertising. Many free apps display intrusive ads, sometimes outside the VPN app itself, sometimes embedded in the network stream. The CSIRO 2017 study found 38% of free Android VPN apps contained malware or malvertising. Some apps inject ads into the user's normal web browsing through the VPN tunnel, replacing the publisher's ad inventory with their own.
4. Premium upsell as the honest model. This is the only model that does not turn the user into the product. Proton VPN, Windscribe, hide.me, and TunnelBear offer a limited free tier (data caps, fewer servers, slower speeds) as marketing for their paid plans. The free tier is real but constrained. They do not log, sell data, or ship trackers. This model is funded entirely by paid subscribers.
The first three models are documented in dozens of free apps on the Google Play Store and Apple App Store in 2026. The fourth is the exception.
Documented risks from independent research
Independent research has documented specific risks repeatedly over the past decade. The findings are remarkably consistent across studies and across years.
CSIRO Australia 2017. The most cited free VPN study, tested 283 free Android VPN apps. Findings: 38% contained malware or malvertising, 75% embedded third-party tracking, 18% did not encrypt traffic at all, 38% leaked the user's real IP address through DNS or WebRTC. Eighteen percent leaked because they functioned as proxies, not actual VPNs.
ICSI Berkeley research. Studies of VPN privacy claims have repeatedly found discrepancies between what apps promise and what they do. Apps claim to encrypt all traffic while logging metadata, claim no logs while having logging code in their binaries, claim anonymous while linking accounts to email and payment data.
Top10VPN free VPN risk index (updates 2023 to 2025). A rolling index of free VPN apps tested against privacy and security criteria. Recurring findings: most free apps fail at least one major test (DNS leak, IP leak, missing kill switch, tracker presence, jurisdiction transparency). Some popular free apps are owned by companies based in jurisdictions with mandatory data retention or limited consumer privacy protection.
Regulatory action. The Federal Trade Commission has settled multiple cases with VPN providers for misrepresenting privacy and logging practices. App stores have removed dozens of free VPN apps over the past three years for violating user data policies.
A critical pattern: the risks are not theoretical. They are documented, reproducible, and have led to legal settlements, app store removals, and academic publications. A free VPN downloaded from an unknown developer in 2026 carries a risk profile substantially worse than the risk it claims to mitigate.
What free VPNs are missing
Beyond the documented risks, free VPNs are usually missing features that matter for actual privacy and usability.
Independent audits. A no-logs claim is meaningless without a third-party audit. None of the bulk free apps have published audits. The exceptions (Proton VPN, Windscribe) have, but only for their paid services in some cases.
Kill switch. A kill switch cuts internet access if the VPN drops, preventing accidental exposure. Most free apps either omit this feature or implement it unreliably.
DNS leak protection. A VPN that leaks DNS requests to your ISP defeats the privacy purpose. Free apps frequently leak. Read our VPN speed troubleshooting guide for context on how protocols handle DNS.
Streaming and torrenting support. Free apps generally cannot unblock Netflix US, Disney+, BBC iPlayer, or other major streaming services because their IP ranges are detected and blocked. Free apps also generally disable peer-to-peer traffic.
Speed. Free apps share a small number of servers among many users, causing congestion. The result: slow speeds, frequent disconnects, and a poor browsing experience.
These missing features are not flaws. They are deliberate design choices in the free monetization model.
The exceptions. Free tiers from credible providers
Four free VPN tiers in 2026 are operated by companies that primarily make money from paid subscriptions, audit their no-logs policy, publish transparency reports, and treat the free tier as marketing rather than as the product.

| Provider | Data cap | Servers | Jurisdiction |
|---|---|---|---|
| Proton VPN | No cap | 3 countries | Switzerland |
| Windscribe | 10 GB/mo | 14 countries | Canada |
| hide.me | 10 GB/mo | 8 countries | Malaysia |
| TunnelBear | 2 GB/mo | 47 countries | Canada |
Proton VPN free. No data cap, no time limit. Servers in three countries (United States, Netherlands, Japan). Slower speeds than paid plans. Owned by Proton (Switzerland), behind Proton Mail. Independent audits and open-source apps. The strongest free VPN in 2026 if your need is privacy on basic web browsing.
Windscribe free. 10 GB per month data cap. Servers in 14 countries. Standard speeds. Canadian company, has published transparency reports and resisted legal pressure to log on multiple occasions.
hide.me free. 10 GB per month data cap. Servers in 8 countries. Slower than paid. Malaysian jurisdiction. Established paid service with audits.
TunnelBear free. 2 GB per month data cap. Servers in 47 countries (paid tier servers are accessible on free, just rate-limited). Canadian company, owned by McAfee since 2018. Has published independent audits.
What these four have in common: real paid subscribers, real revenue, real audits, real transparency. The free tier is constrained on purpose to convert paying users. What they cannot do: unblock streaming reliably, handle torrents, or replace a paid plan for sustained use. They are an honest free option, not a free version of a paid VPN.
What to check before using a free VPN
If you must use a free VPN, the following checks separate the honest from the dangerous. Each check takes a minute or two.
The app store listing should disclose the publisher. Search the company name. If you cannot find a real company, a real address, and a real privacy policy on their website, do not install.
Look for language about data sharing with third parties, about anonymous usage statistics, about retention periods. Vague policies that allow data sale to unspecified partners are red flags.
A no-logs claim without an audit is just marketing. Bulk free apps cannot show one. The four credible free tiers in the previous section can.
Jurisdiction tells you which legal system applies. A free VPN based in a country with mandatory data retention is, by law, logging.
Search the app name plus "controversy," "data breach," "FTC," "lawsuit." If others have reported problems, you will find a trail.
Once installed, run a DNS leak test and an IP leak test (free tools like dnsleaktest.com). If the VPN leaks, it is not protecting you regardless of the marketing.
If any of these checks fail, uninstall.
Better alternatives that cost less than you think
The paid VPN market in 2026 has compressed prices significantly. The trade-off between free and paid is no longer the same as it was five years ago.
Two to seven dollars per month on a 2-year plan. Major paid VPNs offer 2-year plans at roughly $2 to $7 per month. That is less than a streaming subscription. For that price you get independent audits, kill switch, streaming unblocking, full server fleets, no logs, and customer support. See our NordVPN review and our ExpressVPN review for full breakdowns of the two most-audited paid options.
Free credible tiers. Proton VPN free (no data cap), Windscribe free (10 GB), hide.me free (10 GB), TunnelBear free (2 GB). For light occasional use, these are honest options.
Browser-built VPNs. Some browsers (Opera, Brave Premium, Firefox) ship VPN features at low or no cost. These are not full-featured but cover basic browsing privacy.
Self-hosted. A self-hosted WireGuard or OpenVPN server on a $5 per month VPS gives you a private VPN with zero third-party trust. Technical setup required, but achievable for an evening of work.
A paid VPN is no longer a luxury. A bulk free VPN is no longer a money saver. It is a privacy trade.
FAQ
Final word
A free VPN is not free. The question is what the user pays in. With most free apps in 2026, the user pays with their data, their bandwidth, or their attention. With four credible exceptions (Proton VPN, Windscribe, hide.me, TunnelBear), the free tier is honest marketing for a paid plan.
If your need is occasional, light, privacy-oriented browsing, one of the four credible free tiers is fine. If your need is streaming, sustained use, torrenting, or genuine privacy on a daily basis, a paid VPN at $2 to $7 per month is a substantially better deal than the risk profile of a bulk free app.
The internet has moved past the era when free meant good enough. Independent research, regulatory action, and audit transparency now divide the VPN market into two tiers. Pick from the smaller tier.
For a deeper look at how a VPN actually works, see our explainer what is a VPN.
Related reading
Sources
This guide synthesizes the following public sources, consulted in May 2026:
- Ikram, Vallina-Rodriguez, Seneviratne, Kaafar, Paxson. "An Analysis of the Privacy and Security Risks of Android VPN Permission-enabled Apps." CSIRO and University of New South Wales, 2017.
- ICSI Berkeley research on VPN privacy claims, multiple publications 2018 to 2023.
- Top10VPN Free VPN Risk Index, 2023 to 2025 updates.
- Federal Trade Commission VPN enforcement actions, 2019 to 2024.
- Vendor transparency reports: Proton VPN, Windscribe, hide.me, TunnelBear (verified May 2026).
- Hola VPN residential proxy controversy, 2015, documented in academic and tech press coverage.
Study findings, app counts, and provider details reflect published figures at the time of writing and may change. Verify current provider terms before installing any VPN.
Published: May 29, 2026 · Author: Simon Phillips

