Unlimited Bandwidth Proxies: What the Word Leaves Capped
Contents

In short: On a proxy plan, unlimited bandwidth means traffic volume is not billed. Concurrent sessions, speed and a fair-use threshold are three more limits the word does not cover. Per-gigabyte billing costs less for light or occasional work, per-IP billing once one address moves tens of gigabytes a month. On ProxyJam traffic is not metered; each proxy holds up to 100 concurrent sessions, 10 after 50 GB in a calendar month (Terms of Service, section 2.4).
On a proxy plan, unlimited bandwidth usually means one thing: nobody counts the gigabytes you move. It rarely means that nothing else is limited. If you are comparing plans for a static address, our ISP proxies page shows what we sell, and what follows explains which limits sit behind the word and how to check them before you pay.
What does unlimited bandwidth mean on a proxy plan?
It means traffic volume is not billed. Four separate limits can apply to a proxy, and a plan can be generous on one and tight on another:
- Traffic volume: how many gigabytes pass through the proxy in a period. This is the limit the word usually refers to.
- Concurrent sessions: how many connections the proxy holds open at the same time. Sellers call them threads, sessions or connections.
- Speed: how fast data moves, set by the port, by the upstream network, or by deliberate throttling.
- Fair use: a threshold past which one of the other three changes. Speed is reduced, connections are cut back, or a charge begins.
The word unlimited answers for the first of these only. The other three decide whether a heavy job finishes.
Mobile data plans went through the same argument years ago. In 2019 a US mobile carrier agreed to pay $60 million to settle charges from the Federal Trade Commission that it sold unlimited data plans and then throttled them after as little as 2 GB in a billing period. The settlement requires it to disclose any material restriction on speed or amount prominently, outside the fine print. That order covers one carrier, and its question is the right one to put to a proxy seller: what changes, and when?
Per GB or per IP: which costs less?
The two models charge for different things. Per-gigabyte pricing, the usual model for rotating residential pools, charges for the traffic you move. Per-IP pricing, the usual model for static ISP addresses, charges rent on an address for a period and does not count the traffic through it. ISP vs residential proxies explains why each product ends up with its own model.
Which one costs less depends on how many gigabytes the job moves through each address in a month. Our per-IP prices start at $4.30 per IP per month. Take that as the fixed side and compare:
- at 1 GB a month, a pool costs less whenever its rate is under $4.30 per gigabyte;
- at 10 GB a month, the pool has to charge under a tenth of $4.30 per gigabyte to match;
- at 100 GB a month, it has to charge under a hundredth.
The break-even rate falls as traffic grows. Per-gigabyte billing suits small and occasional work, and per-IP billing suits work that is heavy and runs every day. The per-IP bill is also fixed at checkout, while a per-gigabyte bill depends on what the target sends back. Prices for each country are on the pricing page.
How much traffic does a job move?
Estimate it from how the job fetches pages, because the method changes the total more than a hundredfold. The HTTP Archive Web Almanac 2025 measured the median desktop home page at 2,862 KB in July 2025. Only 22 KB of that was the HTML itself. The rest is scripts, images, fonts and video, which a browser fetches and a plain HTTP request does not.
Take a price monitor that checks 500 public product pages four times a day. That is about 60,000 page loads a month:
- fetching only the HTML, at the median 22 KB, it moves about 1.3 GB a month;
- rendering every page in a headless browser, at the median 2.9 MB, it moves about 172 GB.
Product pages are often lighter than home pages (the same report finds less than half the image weight on inner pages), so treat the browser figure as an upper estimate. The first job fits any plan and stays a small line on per-gigabyte billing. The second is where per-IP pricing with unmetered traffic pays off, and where the remaining caps start to matter.
Which caps does an unlimited plan leave out of the headline?
Most plans that advertise unlimited bandwidth state the volume and leave the rest to the terms, so the terms are where to look. Ours are in section 2.4 of the Terms of Service, and they cover the four limits like this:
- volume: traffic is not metered and not billed per gigabyte;
- concurrency: each proxy may hold up to 100 concurrent sessions;
- fair use: once a proxy has moved 50 GB in a calendar month, its ceiling falls to 10 concurrent sessions for the rest of that billing period, and returns to 100 when the next one starts;
- speed: we do not throttle transfer speeds, interrupt transfers in progress, or charge for traffic past any threshold.
A change to these thresholds is published on that page before it takes effect. The same answer, in short, is on our static residential proxies page.
Run the browser-rendered price monitor through one of our proxies and the cap shows up on the ninth day. At about 5.7 GB a day the proxy passes 50 GB then, and from that point it holds 10 connections at once, down from 100. When you plan, count connections: a browser opens about six connections to each domain over HTTP/1.1. A monitor that loads one page at a time fits under 10. One that opens dozens of pages in parallel does not, and needs several addresses or a slower pace.
Which jobs need unlimited traffic?
Jobs that load whole pages or media, every day, from the same place. Three common ones, each a use our Acceptable Use Policy allows when you have the authority to do the work:
- Region testing with media. Checking that video on your own site plays at the right quality, with the right regional catalogue, from inside a market. Apple’s HLS authoring specification lists 6,000 kbit/s for a 1080p H.264 stream, about 2.7 GB per hour of playback. An hour each working day comes to about 54 GB over 20 working days, past the 50 GB line near the end of the month. Our region testing page covers the rest of that job.
- Ad verification with screenshots. Recording what a placement served means loading the whole page and its creatives, video ones included, before the capture. The screenshot is made on your own machine and adds nothing to proxy traffic; the page load is where the gigabytes go. Our ad verification page describes the setup.
- Price monitoring of public pages. The monitor from the worked example, where the site’s terms and robots rules allow it. Section 3 of the policy says public availability alone does not authorise collection.
Other jobs do not need it: a one-off crawl of a few hundred pages, an hourly uptime check of one address, or anything that moves a gigabyte or two a month. Those cost less on per-gigabyte billing. A job that needs a fresh address for every request needs a rotating pool, which we do not sell yet; our rotating residential page explains that model.
What does ProxyJam sell?
ProxyJam sells static ISP addresses in 77 countries, each on a named carrier, priced per IP per period. Traffic is not metered, and the fair-use rule is the concurrency limit in section 2.4 of the Terms of Service, written down where you can read it. There is no free trial. A refund within 24 hours covers up to 10 IPs of a first order, and using the proxies during that window does not void the request (Refund Policy, section 2).
How do you check an unlimited bandwidth offer?
Estimate the gigabytes each address will move in a month, put that next to the per-gigabyte quote you hold, and then ask the questions in the Before You Buy checklist. A plan that answers all of them in its published terms has told you what unlimited means there. If the job is light, a per-gigabyte plan will cost less; if it is heavy and continuous, start with the ISP proxies page. If the term ISP proxy is new to you, the definition of ISP proxies explains what the address is.
