Buy or lease IPv4 in 2026: a decision framework for hosting companies
Whether to buy or lease IPv4 is one of the bigger infrastructure decisions a hosting company makes, and no single answer fits every business. This article sets out a framework built on four questions – cash flow, time horizon, growth and risk – so you can reach a decision you can explain to your finance team. We deliberately leave market prices out: they move, and the logic of the decision matters more than this month’s quote.
Start with cash flow, not the price tag
Buying IPv4 means paying for the addresses up front and holding them as an asset. Leasing turns the same need into a recurring monthly or annual cost. The real question is less “which is cheaper?” and more “which shape of spending suits the business right now?”
Questions for your finance team:
- Can we commit a significant one-off payment without starving other investments?
- Would we rather keep capital free and accept a permanent line in operating costs?
- How would an IP purchase be treated on our balance sheet? Accounting treatment varies, so check with your accountant rather than assuming.
A useful exercise is to compare the purchase cost with cumulative lease payments over different periods. The point at which total lease payments overtake the cost of buying – your break-even – is the most useful single number in this decision. Include the costs around each option, not just the headline figure: registry membership and administration if you own, setup and change fees if you lease.
How long will you need the addresses?
Time horizon often settles the matter. If you need addresses for a specific contract, a trial market or a project with an end date, leasing is the natural fit: you pay for the period you use and hand the block back.
If the addresses will sit underneath your core platform for many years, buying deserves serious consideration. Beyond break-even, ownership becomes steadily cheaper relative to leasing, although you still carry registry fees and admin. An owned block also remains an asset: if your needs shrink, you can sell it or lease it out.
Be honest about uncertainty. A business plan that says “we will need these for ten years” is a forecast, not a fact. If you can’t say with reasonable confidence that the need will outlast break-even, the flexibility of a lease is worth paying for.
Growth: steady, fast or spiky?
The pattern of growth matters as much as total demand.
- Steady, predictable growth suits buying a block sized for the next few years, with planned additions.
- Fast or uncertain growth suits leasing, because you can add capacity in smaller steps without committing capital to a guess.
- Seasonal or customer-driven spikes suit a hybrid: own the base load, lease the peaks.
The hybrid model is underrated. Owning a core block gives you stable addresses for infrastructure, mail and long-lived customer services, while leased blocks absorb new contracts until demand settles.
Weigh the risks on each side
Neither option is risk-free; the risks are simply different.
| Factor | Buying | Leasing |
|---|---|---|
| Upfront capital | High | Low |
| Ongoing cost | Registry fees and admin | Lease fees; renewal terms can change |
| Control | Full, as registered holder | Defined by the contract |
| Renumbering risk | Low | Present if the lease ends or isn’t renewed |
| Admin burden | Membership, transfer, registry objects | Mostly handled by the lessor |
| Exit options | Sell or lease out | Return at end of term |
When buying, the main risks sit in the transaction: confirming the seller’s right to transfer, checking the block’s routing and reputation history, and completing the registry transfer correctly. In the RIPE region, holding allocated IPv4 normally means being an LIR (a RIPE NCC member), with the fees and administration that brings. Check the current RIPE NCC transfer policy and charging scheme before you commit.
When leasing, the main risks sit in the contract: what happens at renewal, how much notice you get, and whether the block arrives clean. A short lease with weak notice terms can force an unplanned renumbering, which is usually far more disruptive than any difference in price.
Factor in where the addresses will be used
If you run workloads in public clouds, your own address space can change the calculation. AWS has charged $0.005 per hour for every public IPv4 address since February 2024, and Google Cloud and Azure apply the same hourly rate to external IPs on standard VMs and Standard static public IPs respectively. None of the three charges that fee for addresses you bring yourself. At $0.005 per hour, one address costs about $43.80 a year, so a large fleet of instances carries a noticeable IP bill.
Both owned and leased blocks can be brought to these clouds, provided the registered holder creates the required ROA and validation records. If BYOIP is part of your plan, confirm that a lessor will support it before you sign.
A short checklist to finish:
- Calculate break-even between buying and cumulative lease cost.
- Judge how confidently your need extends beyond that point.
- Map your growth pattern: steady, fast or spiky.
- Decide which risk you tolerate better – capital tied up, or possible renumbering.
- Consider whether a hybrid of owned and leased space fits best.
How TACS can help
We work on both sides of this decision, so we can model the options against your own numbers rather than push one answer. If ownership makes sense, see how we handle buying IPv4 and registry membership through our LIR services; if flexibility matters more, start with leasing IPv4.