Unlimited Bandwidth ISP Proxies: What "Unlimited" Really Means (With Real Data)

Updated 30 October 2026: added the bulk price ladder for unlimited bandwidth proxies, the effective cost per GB for our example workloads, an unmetered vs per-GB table and six new FAQs.

Unlimited bandwidth ISP proxies are static IPs you pay for per IP per month, with no gigabyte counter: you can push as much traffic through each IP as your job needs and the bill stays the same. That is the opposite of rotating residential and mobile proxies, which are billed per GB. The catch is that many providers write "unlimited" on the pricing page and then add a fair-use threshold in the help center that slows you down or starts charging after a set number of GB per IP. CleanProxies ISP proxies and ISP subnets come with unlimited bandwidth and unlimited threads, with no throttling and no overage fees, from $0.75/IP.

This guide explains what "unlimited" really means across the market, shows how much data real workloads use (we measured it on 29 October 2026), and gives you a short Python script that tells you when per-IP unlimited beats per-GB billing for your own job.

TL;DR

• Unlimited bandwidth = you pay per IP, not per GB. Heavy jobs cost the same as light ones. On CleanProxies ISP proxies there is no GB counter at all: unlimited bandwidth, unlimited threads, no throttling, no overage fees.

• "Unlimited" on other sites often has a fair-use threshold. On their own help pages, several large providers cut sessions, slow speed or start charging after 50–100 GB per IP per month (sources in the table below).

• How you fetch decides how much you use. In our test, a full browser load used about 11× more data than a compressed HTML fetch of the same pages (311 KB vs 29 KB on average).

• Break-even is low. 100 CleanProxies ISP IPs cost $149/month. The same money buys about 150 GB of per-GB traffic, so once your job moves more than about 1.5 GB per IP per month, unlimited is cheaper. A steady scraper on 100 IPs in our model moved 749 GB to 8 TB a month.

• Test it yourself with the $4.99 trial (5 ISP IPs for 1 hour), then size your plan with the proxy cost calculator.

"Unlimited" at a glance

What each provider's own pages say about bandwidth on static ISP / static residential proxies (checked 29 October 2026):

None of these limits is hidden in a bad-faith way; they are on the providers' own help pages. But they change the math. A plan that is "unlimited" until 50 GB per IP is, for a heavy workload, a metered plan with extra steps.

Metered, unmetered and "fair use": the three models

Metered (per GB). You buy data. Every request, retry, image and script counts against your balance. This is how almost all rotating residential and mobile proxies are sold, including ours: real home and mobile IPs cost the network real money per GB, so per-GB pricing is fair there. CleanProxies rotating residential is as low as $0.63/GB and mobile as low as $0.79/GB, and the data never expires.

Unmetered (per IP). You rent a fixed set of IPs for a month. Traffic is not counted. This is how static ISP proxies work when they run on serious infrastructure: the IP is announced from a carrier ASN but served from data-center-grade hardware, so traffic is cheap to carry. CleanProxies ISP proxies run on 100Gbps network ports with <20ms latency and 99.9% uptime, which is why we can leave the meter off.

Fair use (per IP with a threshold). You rent IPs, but after a set number of GB per IP (or per plan), the provider cuts concurrent sessions, throttles speed or bills the extra. The threshold is usually in a help article, not on the pricing card.

If you are new to static IPs, start with what a static residential proxy is. The short version: an ISP proxy is a dedicated IP registered to a real internet provider, which stays yours all month.

How much data do real workloads use? Our measurements

"Unlimited" only matters if you would otherwise hit a limit, so we measured. On 29 October 2026 we fetched six public pages two ways from our server in the US:

1. Plain HTTP fetch with curl, once asking for gzip compression and once without, recording the bytes received (body + headers).

2. Full browser load in headless Chromium (Playwright), a fresh browser context (empty cache) per page, counting the encoded bytes of every request the page made (HTML, CSS, JS, images, fonts) until the network went idle, plus 1.5 seconds.

All six pages returned HTTP 200. We fetched them directly from the server, without a proxy, so the numbers show the page weight itself; a proxy adds only a little connection overhead on top. Each page was measured once.

What the numbers mean:

• Compression is free bandwidth. On the four servers that support it, asking for gzip cut the transfer 4.2× to 7.6×. Always send Accept-Encoding: gzip (curl: --compressed; Python requests and httpx do it by default).

• Browsers are the big consumer. A full page load averaged 311 KB, about 11× the compressed HTML, because it downloads scripts, styles and images too. Your own target pages may be much heavier: product pages with large image galleries can pass several MB.

• Your numbers will differ. These are six public pages, measured once. Measure your real targets with the same method (code below) before you choose a plan.

For a full sizing method (how many IPs, request rates, GB budgets), see how many proxies do I need.

Measure your own pages

# bytes received for one compressed HTML fetch (body + headers)
curl -s -o /dev/null --compressed -w "%{http_code} %{size_header} %{size_download}\n" \
  -x http://USER:PASS@HOST:PORT https://example.com/

Replace USER:PASS@HOST:PORT with a line from your proxy list. Run it on 20–50 real pages and average the result. To measure full browser loads, this is the Playwright script we used (Node.js, npm i playwright):

// page-weight.mjs: total encoded bytes per full page load
import { chromium } from "playwright";
const browser = await chromium.launch(); // add { proxy: { server, username, password } } to measure through a proxy
for (const url of process.argv.slice(2)) {
  const ctx = await browser.newContext(); // fresh context = empty cache
  const page = await ctx.newPage();
  const cdp = await ctx.newCDPSession(page);
  await cdp.send("Network.enable");
  let bytes = 0, requests = 0;
  cdp.on("Network.loadingFinished", (e) => { bytes += e.encodedDataLength; requests++; });
  const res = await page.goto(url, { waitUntil: "networkidle", timeout: 45000 });
  await page.waitForTimeout(1500);
  console.log(url, res?.status(), requests, "requests", (bytes / 1000).toFixed(0), "KB");
  await ctx.close();
}
await browser.close();

Run it with node page-weight.mjs https://example.com/ https://example.org/.

Worked example: when does unlimited pay off?

Take a price-monitoring job on 100 static IPs, each sending a polite 6 requests per minute, around the clock, for 30 days:

100 IPs × 6 requests/min × 60 × 24 × 30 = 25.9 million requests a month

Now the bill:

*Traffic multiplied by CleanProxies' own rotating residential unit rates at the 500 GB ($0.89/GB) and 5,000 GB ($0.74/GB) tiers, used here only to show how per-GB billing scales. Per-GB data is bought up front in tiers (it never expires on CleanProxies), so a real purchase rounds up to a tier. ISP prices are the standard ladder; AT&T has its own ladder from $2.00/IP. Rotating residential is still the right tool when you need many locations or a fresh IP per request.

And against the fair-use thresholds in the table above: the browser scraper's 81 GB per IP is past Oxylabs' 50 GB point, where sessions drop to 10 per IP. Whether that slows the job depends on how many parallel connections you open per IP; browser jobs, which open many connections per page, are the ones that feel it. Double the request rate and the job passes the 100 GB per IP marks at Bright Data, Decodo and IPRoyal too, where the next step is reduced speed, fewer threads or an extra bill. On CleanProxies the same job keeps the same price and the same thread allowance.

The break-even script

This prints the monthly traffic per IP above which a given per-IP price is cheaper than a given per-GB rate, plus the estimated cost of your own workload both ways (traffic × unit rate, before rounding up to a GB tier).

# unlimited_vs_per_gb.py: when do unlimited ISP proxies beat per-GB billing?
def monthly_gb(ips, req_per_min_per_ip, kb_per_request, hours_per_day=24, days=30):
    requests = ips * req_per_min_per_ip * 60 * hours_per_day * days
    return requests * kb_per_request / 1_000_000  # 1 GB = 1,000,000 KB

def compare(ips, price_per_ip, price_per_gb, req_per_min_per_ip, kb_per_request):
    flat = ips * price_per_ip                     # unlimited: same bill at any traffic
    gb = monthly_gb(ips, req_per_min_per_ip, kb_per_request)
    metered = gb * price_per_gb
    breakeven_gb_per_ip = price_per_ip / price_per_gb
    print(f"traffic: {gb:,.0f} GB/month ({gb / ips:,.1f} GB per IP)")
    print(f"unlimited per-IP plan: ${flat:,.2f}")
    print(f"per-GB plan:           ${metered:,.2f}")
    print(f"break-even: {breakeven_gb_per_ip:.2f} GB per IP per month")

# 100 CleanProxies ISP IPs at $1.49/IP vs per-GB at $0.99/GB (100 GB tier)
compare(ips=100, price_per_ip=1.49, price_per_gb=0.99,
        req_per_min_per_ip=6, kb_per_request=28.9)

Output with our measured HTML average:

traffic: 749 GB/month (7.5 GB per IP)
unlimited per-IP plan: $149.00
per-GB plan:           $741.60
break-even: 1.51 GB per IP per month

(The per-GB figure is higher here than in the table above because the script uses one rate, $0.99/GB, for all traffic; in practice you would buy a bigger, cheaper GB tier.) The 1.51 GB figure is specific to $1.49/IP against $0.99/GB; change the inputs to your own prices, request rate and page weight. With CleanProxies' 1,000-IP price ($0.99/IP) against the 1,000 GB rate ($0.84/GB), or the 10,000-IP price ($0.75/IP) against the 10,000 GB rate ($0.63/GB), it comes out at about 1.2 GB per IP per month. Our modelled HTML scraper (7.5 GB per IP per month) passes that point in about 6 days.

When unlimited bandwidth matters most

One thing unlimited bandwidth does not change: target sites still limit how fast each IP may send requests. Unlimited means no data meter and no thread cap on the proxy side; your success rate still depends on pacing requests per IP, handling 429s and keeping sessions consistent (see how to fix proxy errors).

• Browser automation. Playwright, Puppeteer and Selenium load every asset, so data per page is 5–15× higher than plain HTTP. Our Python rotation guide and the Playwright setup work with any CleanProxies ISP line.

• Steady, round-the-clock scraping on a known set of sites. Price monitoring, stock checks and catalog syncs from US retailers. See the web scraping use case.

• SEO rank tracking from the US with a fixed set of IPs. See SEO proxies.

• Account and storefront management where each account keeps the same IP all month, and the session loads full pages, images and dashboards all day.

• Large files and media: downloading reports, product images or your own video assets for QA, where one job can move gigabytes.

• Large fleets. Teams with hundreds or thousands of IPs, where a per-IP threshold multiplies into terabytes. ISP subnets (whole /24 blocks of 255, 510 or 1,020 IPs, from $1.19/IP) are unlimited too.

When is per-GB the better choice? When you need many locations or a new IP on every request: rotating residential covers 190+ countries with city and state targeting and a 50M+ IP pool; mobile covers 150+ countries. Light, occasional jobs that move a few GB a month also fit per-GB well. Many teams run both: CleanProxies ISP for the heavy, steady US work and rotating residential for broad geo coverage.

Why CleanProxies ISP proxies

• Truly unlimited: unlimited bandwidth and unlimited threads on every plan, from 5 IPs to 10,000. No fair-use threshold, no throttling, no overage fees.

• Real US carrier IPs: static IPs on US carrier ranges, one fixed IP each all month (no state or city selection).

• The cheapest pricing in the market: as low as $0.75/IP at 10,000 IPs, $0.99/IP at 1,000 and $1.49/IP at 100. ISP subnets from $1.19/IP. (AT&T has its own price ladder.)

• Cleanest IPs: 0 fraud score, super clean ranges on Tier-1 US carrier ASNs (Frontier, AT&T, Comcast, Cox, RCN, T-Mobile).

• 100% dedicated: your IPs are never shared with another customer, so nobody else's activity touches their reputation.

• Fast: 100Gbps network ports, <20ms latency, 99.9% uptime.

• Instant delivery: in our live orders, ISP proxies were active 6–19 seconds after payment.

• Flexible auth: user:pass, IP whitelist, or SOCKS5 user:pass from the dashboard.

• Great 24/7 support.

Compare us side by side on the compare page, or read the bulk ISP proxy price comparison.

How to check any provider's "unlimited" claim

1. Search the help center, not the pricing page, for "fair use", "fair usage", "threshold" and "throttle".

2. Find the number: GB per IP, GB per plan, or concurrent sessions per IP.

3. Find the consequence: fewer threads, lower speed, proxies stop, or an overage bill.

4. Check the minimum order for unlimited: some providers only unlock it above a set number of proxies.

5. Multiply by your workload: run the script above with your own request rate and page weight.

6. Test the real thing. Start the CleanProxies $4.99 trial: 5 ISP IPs for 1 hour (or 100 MB of rotating residential). Run your heaviest job through it for the hour, check the IPs with our proxy checker, and measure the speed and success rate on your own targets. An hour won't show a full month of traffic, but it shows real throughput and page weights, which is all the script above needs. Start the $4.99 trial.

Unlimited bandwidth proxy prices at volume

Our unlimited bandwidth proxies are billed per IP, so the cost per GB falls the more traffic you push. Here is the bulk ladder for CleanProxies ISP proxies, with the effective cost per GB for the two example workloads in the worked example above, which use our measured page weights: a compressed HTML scraper (7.5 GB per IP per month) and a full-browser scraper (81 GB per IP per month).

Price per IP divided by GB per IP, rounded. Prices from the pricing page on 30 October 2026 (standard carrier ladder; AT&T has its own). The heavier the job, the cheaper each GB gets, because the bill never changes.

Unmetered vs per-GB proxies at a glance

Pricing

See every tier on the pricing page.

Ready to stop counting gigabytes?

• Buy unlimited bandwidth ISP proxies

• See all pricing

• Check your current proxies for free

Written and tested by the CleanProxies team. Measurements taken 29 October 2026 from a US server; competitor terms read from each provider's own pages the same day.

Frequently Asked Questions

Team CleanProxies

Proxy Infrastructure Team

Share