For three consecutive years, residential proxies only became cheaper — the market got used to the idea that each subsequent rate was more advantageous than the previous one. In 2026, this trend reversed: according to a market study by Proxyway, based on a survey of 13 providers conducted in March-April 2026, the rates for residential proxies increased by approximately 25% and returned to 125-133% of the baseline levels of 2023. Major players — IPRoyal, Decodo, Oxylabs — quietly removed coupons offering discounts of 40-50% that had attracted new customers in recent seasons. Let's explore what exactly happened to prices, why the supply of residential IPs shrank, and how to budget for traffic now.
What the Numbers Showed: A 75% Drop Followed by a Reversal
Context is important. From 2023 to 2025, rates for residential proxies shrank by 75% — this was a price war where providers bought market share at the expense of margins. It was during this time that the industry developed the habit of thinking: "wait a month — it will get cheaper." In 2026, this logic ceased to function. Proxyway notes stabilization and a reversal: prices began to rise, and promotional mechanisms that effectively made the price list decorative were rolled back.
The average prices per gigabyte in the market for 2026 are as follows (median from surveyed providers):
| Package Size | Average Price per GB |
|---|---|
| 5 GB | $4.16 |
| 50 GB | $3.00 |
| 500 GB | $2.28 |
| 1,000 GB | $1.94 |
The gap between "trying" and "working seriously" is more than twofold. Providers with aggressive pricing stand out: DataImpulse operates in the range of $0.49–1.00 per GB, while Proxyrack charges $0.80–1.20. This does not mean that a cheaper gigabyte is automatically worse, but it does mean that one should compare not the price, but the cost of a successful request — more on that below.
At the same time, the sizes of the pools did not grow proportionally to the price. Oxylabs holds the maximum in residential networks — around 175 million addresses, while the market median is 54 million. In the mobile segment, the leader is SOAX with ~33 million, and the median is 16 million. Thus, the market did not suddenly become richer in IPs; it became more expensive with approximately the same capacity.
Why the Supply of Residential IPs Shrunk
Prices rise where supply contracts — and in 2026, several blows hit the channels for supplying residential addresses.
Two Major Takedowns in Six Months
In January 2026, Google, together with the FBI, took down the IPIDEA network — one of the largest residential proxy networks. In July 2026, NetNut was dismantled: according to published data, the infrastructure relied on more than 2 million compromised devices — smart TVs, streaming boxes, and Android gadgets. Each such episode removes millions of "cheap" exit nodes from the market while simultaneously raising the compliance bar for those who remain.
Device Manufacturers Close the Loophole in Firmware
On July 2, 2026, researchers from Spur published an analysis of proxy SDKs in television applications: more than 42% of applications in the LG webOS store contained code that turned the TV into an exit node, while in Samsung Tizen, the figure was over 25%. On July 21, LG Electronics USA announced that it would remove applications that did not eliminate such SDKs. LG's Senior Vice President John Taylor stated the position clearly: a residential proxy network is not what LG TVs are intended for. Bright Data, whose SDKs dominated this sample, responded that users consent on a separate screen and receive value in return, citing an independent audit by PwC.
The debate over "real consent" is not academic here: Spur researcher Trevor Sutter pointed out that a checkbox within a TV application does not equate to meaningful transparency, especially when a child in the household checks it. For the market, the practical conclusion is that the cheapest way to acquire residential IPs is becoming legally and reputationally toxic. We discussed how the cost structure of a pool works and how much the owner of a home channel actually earns in the article where residential proxies come from.
Demand is Growing Faster than Capacity
The second half of the equation is consumption. In the State of Web Scraping 2026 report (Apify in collaboration with The Web Scraping Club, published on January 29, 2026, based on a survey of communities in December 2025), the picture is as follows:
- 65.8% of respondents used more proxies last year than the year before;
- 58.3% noted an increase in spending on proxies;
- over 62% reported an increase in infrastructure costs overall — primarily due to tightening anti-bot protections;
- 43.1% work simultaneously with 2-3 providers, while another 12.1% work with 4-5.
Note the connection: expenses grew during the period when prices were falling. This means that the bill was increasing not due to price, but due to volume and complexity — more retries, more "expensive" targets, more sessions for the same task. Now, this is compounded by the price reversal.
Provider revenues confirm that demand has not disappeared: Bright Data showed about 50% annual growth with an ARR of around $300 million, NetNut — 28%, Webshare — plus 30%, Rayobyte — about 25% (the best result since its founding). A market growing at such rates has no motivation to continue a price war.
The Mobile Segment Went in the Opposite Direction
The most counterintuitive aspect of the report is mobile proxies. While residential proxies became more expensive, mobile proxies plummeted: Proxyway reports a decrease of up to 98% for some providers (in the case of Rayobyte), while for Proxidize and Proxyrack, mobile traffic fell below $2 per GB. For the first time in a long time, a mobile gigabyte may be cheaper than a residential one — despite the fact that mobile IPs are traditionally considered more "trusted" by platforms due to CGNAT and the common pool of operators.
Against this backdrop, Bright Data's maneuver is telling: since April 2026, the company has stopped selling mobile proxies to new clients — the product page redirects to the general section of proxy types, and the pricing page returns a 404 error; there was no official announcement, and everything is known from support and observed changes on the website. Existing subscriptions continue to operate. Simultaneously, on April 1, 2026, the company updated its Acceptable Use Policy, removing support for account management scenarios on social media, including TikTok and Instagram — precisely the scenarios for which mobile proxies were most often purchased. New clients are offered residential and ISP addresses.
The practical conclusion is that if your scenario involves working with accounts and mobile applications, there is now a rare window when mobile proxies are priced comparably to residential ones, while some major players in this segment are exiting. For scraping large volumes of web content, the economy still favors residential addresses.
What to Do with Your Traffic Budget Right Now
The price reversal changes not so much the choice of provider as the discipline of calculations. Here are several practical implications.
- Calculate the cost of a successful request, not per gigabyte. A provider at $1.00/GB, where a third of requests result in 403 errors and retries, is more expensive than a provider at $2.50/GB with a clean pool. Take your actual share of successful responses on the target site and divide your spending by that — only this number is comparable between providers.
- Reassess your package step. The difference between 5 GB and 500 GB in the market average is more than double ($4.16 versus $2.28). If you consistently choose a volume, fractional purchases are the most expensive way to operate.
- Do not expect discounts to return. Coupons of 40-50% were a tool of the price war, not a permanent mechanism. Planning your 2026 budget based on promotional prices from 2024 is a sure way to miss by one and a half times.
- Multi-provider usage is now the norm, not paranoia. 55% of respondents use more than one provider. This is insurance not only against bans but also against the possibility that a specific provider will discontinue their product or revise their AUP, as happened with Bright Data's mobile direction.
- Ask about the origin of the addresses. After the takedowns and the SDK issue in TVs, the question "where do you get your IPs and how do you vet nodes" is no longer a formality: too cheap residential traffic in 2026 likely means a pool that could be targeted by law enforcement tomorrow.
- Cut volume before cutting price. The biggest savings are usually not in the rate but in the traffic: blocking image and font loading, caching, avoiding re-crawling unchanged pages. We showed how to calculate such economics step by step in our analysis of the cost of scraping a million pages.
Conclusion
The residential proxy market has gone through a full cycle: the dumping of 2023-2025 has ended, and in 2026, prices returned to levels roughly a quarter higher than three years ago. The reason is not the greed of providers, but the fact that cheap supply is physically shrinking — networks on compromised devices are being dismantled by law enforcement, and device manufacturers are closing the path of SDKs in firmware. At the same time, demand is growing: nearly two-thirds of practitioners are using more proxies than a year ago.
For teams relying on proxy traffic, this means a shift from the logic of "finding the cheapest gigabyte" to the logic of "achieving predictable results for understandable money": counting successful requests, maintaining more than one provider, checking the origin of the pool, and trimming unnecessary traffic on their side. The price war is over — winning will now require engineering.