While the Industry Raised Prices, a Few Hosts Promised Never To. Regulation Just Made That a Strategy.

The spring of 2026 will be remembered in hosting as the season everyone raised prices at once. Hetzner lifted them three times since January. OVHcloud went up by as much as 49% on VPS, Netcup by up to 24%, Scaleway across its catalogue, IONOS by unexplained email, and Namecheap’s renewals by up to 51%. The driver was a memory market that has repriced the industry’s core hardware.

Against that backdrop, a small group of providers is running the opposite play, and it deserves more strategic attention than its members’ size would suggest:

• InterServer still locks the rate quoted at signup for the life of the service, today $7 per month for standard shared hosting, disclosed upfront beside a $2.50 first month, measured on its live pricing page on June 11. It has maintained the guarantee for years and, notably, has not withdrawn it through the worst component inflation in a decade.
• HostMetro renews its plans at exactly the signup price.
• Poland’s webh.pl backs a formal five-year price guarantee on its cloud and KVM VPS lines.
• Contabo answered its competitors’ increases by launching better-specified plans at lower prices.

For years, the price-lock was a budget-segment marketing gimmick. Two things changed in the past twelve months: the cost environment made it genuinely expensive to offer, and a wave of auto-renewal regulation quietly made it the cheapest compliance strategy in the industry.

Key facts

• The wave: Hetzner +30–37%, OVHcloud +43–49%, Netcup +24%, Namecheap renewals +51%, part of the broader 2026 repricing wave
• The counter-position: InterServer discloses its renewal rate at signup and locks it for the life of the service ($7/mo standard shared after a $2.50 first month; $6 VPS); HostMetro renews at the identical signup price; Poland’s webh.pl runs a formal 5-year price guarantee; Contabo launched better specs at lower prices instead of raising
• The cost backdrop: DRAM contract prices +58–63% QoQ, RAM projected +250–300% by end-2026 vs September 2025, a price-lock is now a sold inflation put
• The regulatory dividend: New York’s §527-a consent-or-refund rule and the EU DFA’s proposed opt-in renewals only bite when the renewal price changes, a locked price needs no consent flow
• The channel dividend: AI assistants synthesize hosting recommendations from community sentiment, where “no renewal jump” is the most-rewarded attribute
• The risk: a locked back book in an inflationary market is a balance-sheet liability, the Contabo/InterServer position is a bet that efficiency gains outrun component costs

Who Is Actually Doing This

InterServer is the category’s reference case because its guarantee is the oldest, the broadest, and the most explicit: the price quoted at signup is the price at every renewal, as measured on its live pricing page this week, $7 per month for standard shared hosting after a disclosed $2.50 first month, and $6 per VPS slice, with no term-length conditions. Note the structure honestly: even the price-lock reference case now leads with an intro number; what it refuses to do is hide, or later move, the rate that follows. The structurally important detail is what the lock does and does not cover: it protects the existing customer’s rate for the life of that service, while leaving InterServer free to reprice new orders. That makes it a grandfathering machine rather than a price freeze, economically similar to what OVHcloud did defensively this spring when it capped legacy-infrastructure increases at 2–6% while repricing new deployments by up to 49%, and to the two-tier structure Hetzner formalized in its June 15 standardization. The difference is positioning: the incumbents grandfather quietly to limit churn; InterServer leads its marketing with the promise.

HostMetro sells the same mechanic (its published comparison shows the Mega Max plan at $59.40 for the first year renewing at an identical $59.40, against competitors renewing at multiples) with an asterisk to its full terms that buyers should read, since the guarantee’s footnoted conditions are where such promises live or die. Contabo has made no formal lifetime guarantee but has done something arguably harder in 2026: it absorbed the component-cost wave, launched a better-specified lineup at prices starting under $5, and announced no increases, an implicit lock sustained by claimed efficiency gains at its Hub Europe facility. The model has European practitioners too, which matters for the regulatory argument below: Poland’s webh.pl has offered a formal five-year price guarantee since 2023 on its cloud hosting and KVM VPS lines, the company reprices its catalogue once a year, on January 1, but the adjustments do not touch existing customers within the guarantee window, and plan specifications continue to improve while the price stays locked. A time-boxed lock of this kind is the disciplined middle ground between InterServer’s lifetime promise and the industry’s silent jump: long enough to be a genuine commitment, bounded enough to be insurable against a 250% RAM curve. And at the domain layer, Cloudflare Registrar has run the limit case for years: domains sold at wholesale registry cost with no markup at all, a structural rather than promotional promise that renewal price equals cost.

What a Price-Lock Actually Is, Financially

Strip the marketing and a renewal price-lock is a financial instrument: the provider sells the customer a put option on cost inflation. The customer’s exposure to hardware, energy, and licensing cost increases transfers to the provider for the life of the service. For most of the 2010s that option was nearly free to write, hardware got cheaper every year, so the lock cost the provider nothing and signaled confidence. The 2026 environment repriced the option violently: DRAM contract prices up 58–63% quarter-over-quarter, NAND up 70–75%, providers committing to component orders twelve months ahead, and OVHcloud citing supplier projections of 250–300% RAM price increases by end-2026 against September 2025. Every provider holding a locked book through that curve is paying out on the option it sold.

Which is precisely why the surviving locks are informative. A lock is only rational under one of three conditions:

• Efficiency offsets. The provider’s cost per delivered unit falls fast enough, through utilization, automation, owned facilities, and hardware sweating, to absorb component inflation. This is Contabo’s stated position and the entire bet behind its counter-cyclical launch.
• Margin structure. The locked price was set high enough, or the cost base is small enough relative to support and operations, that hardware is not the dominant cost line. That is plausible for shared hosting, where a server’s cost divides across hundreds of accounts, and visibly implausible for GPU or dedicated products, which is why nobody locks those.
• Lifetime-value arithmetic. Churn at the renewal jump is so expensive (the industry’s own data puts price as the top churn driver, cited by 56% of providers) that forgoing the jump and keeping the customer beats taking it and re-acquiring a replacement at 2026’s acquisition costs, especially as the affiliate machinery for acquiring replacements is itself breaking down.

The Regulatory Dividend Nobody Priced In

Here is what changed in the past year that the price-lock providers themselves have barely begun to market. The entire regulatory assault on subscription economics (New York’s GBL §527-a, California’s ARL class-action wave, the FTC’s restarted rulemaking, the EU Digital Fairness Act) shares a single trigger condition: the renewal that differs from what the customer agreed to. New York’s consent-or-refund mechanic activates only “in the event a business wishes to increase the price.” The DFA’s most dangerous proposal for hosting (auto-renewal off by default, with mandatory pre-renewal reminders) is dangerous specifically because it converts a silent repricing into a conscious repurchase decision at three or four times the remembered price.

A price-locked provider walks through all of it untouched. No price increase means no §527-a consent flow, no 14-day refund window, no repricing disclosure to engineer. A mandatory renewal reminder that says “your plan renews at the same price you have always paid” is not a churn event; it is a retention advertisement that a regulator forced the provider to send.

The asymmetry is worth stating in C-level terms: for the renewal-jump majority, the emerging rulebook is a compliance cost; for the price-locked minority, it is a free marketing channel. Every statute that forces renewal pricing into the open widens the visible gap between the two models, and it does so exactly as AI-mediated buying makes that gap unprecedentedly visible: assistants synthesizing recommendations from Reddit threads and community sentiment reward “the price never changed” more reliably than any affiliate placement ever rewarded a commission. InterServer’s two decades of accumulated “still paying what I signed up for” forum posts have become an acquisition asset that cannot be bought retroactively.

The Honest Trade-Offs

The model has real costs, and the balance matters for anyone considering adopting it. A locked book in an inflationary market is a liability that compounds: every year of component inflation widens the gap between the locked rate and replacement cost, and unlike the incumbents’ quiet grandfathering, a public guarantee cannot be walked back without destroying the brand asset it created. The price-lock providers also tend to charge more honestly upfront (InterServer’s locked $7 is the real sticker, not the loss-leader $0.99 of an intro-pricing competitor) which means losing the comparison-table war that, until recently, determined affiliate-driven acquisition. It is no accident that the model’s practitioners are privately held, modestly sized, and absent from the top of sponsored rankings: the renewal jump existed to fund the bounties that bought those rankings. A price-locked provider structurally cannot pay a $65–$200 bounty against a $30 first-year revenue line; the math this publication’s affiliate analysis laid out runs exactly backwards for them. The lock and the affiliate machine were always mutually exclusive strategies. What has changed is that the affiliate machine is dying and the lock’s distribution channel, community trust now amplified through AI citation, is ascendant.

There is also a fine-print caveat buyers and copywriters should both respect: locks cover the service as sold. Plan specifications can still be cut, the Hostinger spec-reduction pattern shows how value erodes without a headline price change, and locks rarely survive product migrations, platform changes, or acquisition by a consolidator, which, in a market with double-digit hosting acquisitions in this half-year alone, is not a hypothetical end state. A price-lock from a provider that gets acquired is worth exactly as much as the acquirer’s brand calculus says it is.

What the Majority Should Take From the Minority

The partial lock is available to everyone. The InterServer mechanic (lock existing customers, reprice new orders) is precisely what OVHcloud and Hetzner already did this spring under churn pressure, without claiming credit for it. Formalizing the grandfather clause as a public guarantee converts a defensive necessity into a brand asset and a compliance simplification at near-zero incremental cost, because the back book was not going to be repriced aggressively anyway once §527-a-style rules attach consent flows to every increase.

Rate insurance is a sellable product. If a renewal lock is a put option on inflation, it can be priced as one: a premium tier or paid add-on that fixes the renewal rate for three or five years. Enterprise buyers already demand multi-year price protection in contracts; selling it as a product for SMB hosting converts pricing risk into a revenue line and self-selects exactly the long-retention customers the LTV math wants. This is not hypothetical, webh.pl’s five-year guarantee is precisely this product, already shipping, and its survival through the 2026 cost wave is the proof-of-concept the rest of the market can underwrite against.

The reminder is coming; choose what it says. Between New York, California, and the DFA trajectory, the pre-renewal notification will be mandatory across most of the addressable market within the planning horizon. Every provider should decide now whether that legally compelled message will read as a price increase requiring consent, with the churn that entails, or as proof of a promise kept. The handful of hosts that spent the worst cost inflation in a decade refusing to raise prices have already written their version. The rest of the industry just acquired a regulatory deadline for writing theirs.