Cost

Buy Blob storage reserved capacity

One capability: commit to the steady, predictable base of Blob and Data Lake Gen2 storage you keep year after year, and pay the reserved rate for it instead of the pay-as-you-go rate, without changing a single byte of how the data is stored.

13 min·10 sections·AZURE

Last reviewed 18 June 2026

Blob storage reserved capacity: the basics

What does Azure Advisor actually mean by 'consider Blob storage reserved instance'?

If you keep a large, steady amount of data in standard Blob storage or Azure Data Lake Storage Gen2, you are almost certainly paying the on-demand rate for capacity you never delete. Azure Storage reserved capacity is a billing discount: you commit to a fixed amount of capacity, measured in units of 100 TiB or 1 PiB per month, for a one-year or three-year term, and Azure charges the matching usage at the reserved rate instead of the pay-as-you-go rate. Nothing about the data moves or changes. The accounts, containers, tiers and redundancy all stay exactly as they are. Only the price of the capacity you were already going to store changes.

Azure Advisor surfaces this as a Cost recommendation, 'Consider Blob storage reserved instance to save on Blob v2 and Datalake storage Gen2 costs' (recommendation ID 0169a2e1-c7bf-4c37-90b8-0714811c82d3). Advisor looks at your recent usage of standard general-purpose v2 and Blob storage accounts, works out the stable base you have stored every hour over the lookback window, and recommends a reservation sized to that base. The recommendation is purely a savings signal: there is no security or reliability change behind it, just an unused commitment discount you are leaving on the table.

The reservation is scoped to a specific region, access tier and redundancy option, and it is a use-it-or-lose-it commitment for each hour. That is the whole shape of the decision: reserve the floor of storage you are confident you will always be holding in a given region, tier and redundancy, and keep paying pay-as-you-go for the variable layer on top. Get the floor right and the discount is risk-free; reserve more than you actually hold and the unused hours are simply wasted.

In this lesson you will learn exactly what Azure Storage reserved capacity discounts and what it does not, how Advisor decides a reservation is worth recommending, how to size the commitment to a safe floor so the discount is risk-free, and how to purchase and govern reservations across the estate. The focus throughout is the saving and the payback, not security: this is a pricing decision about capacity you already hold.

Fun fact

The discount that hides in plain sight

Reserved capacity is unusual among cost levers because it changes nothing about the workload. There is no migration, no tier change, no code path to test, and no risk to availability or durability. The data sits exactly where it was, served exactly as before, and the invoice line for capacity simply drops to the reserved rate. The most common way teams lose this saving is not technical failure but silence: a reservation reaches the end of its term, nobody renews it, and the capacity quietly rolls back onto the full pay-as-you-go rate. Azure sends a notice 30 days before expiry and again on the day, precisely because the lapse is so easy to miss.

Reading the reserved-capacity recommendation across a subscription

Priya runs the platform team at a media company that stores years of rendered assets and a growing Data Lake Gen2 in standard storage. Azure Advisor has raised the Blob reserved-capacity recommendation with a projected annual saving, and the question on the table is whether the committed base is solid enough to act on.

Rather than accept the headline number blind, Priya starts by pulling the Advisor cost recommendation and confirming what stable capacity the estate has actually been holding, so the reservation is sized to a floor the team is confident about rather than to a recent peak.

List the Advisor Cost recommendations and surface the Blob reserved-capacity one, with the saving Advisor has projected against your own usage.

$ az advisor recommendation list --category Cost \ --query "[?contains(shortDescription.solution, 'Blob storage reserved')].{problem:shortDescription.problem, saving:extendedProperties.annualSavingsAmount, region:extendedProperties.region}" -o table

Problem Saving Region

-------------------------------------------- -------- --------

You could save by buying a reserved instance 12480 westeurope

# Advisor sized this against the stable base you already store. Confirm the floor before committing.

Advisor projects the annual saving against your real recent usage in a specific region. Treat it as a sized starting point, then sanity-check the floor you are confident you will hold for the whole term.

How the reserved-capacity discount actually appliesdeep dive

A reservation is a billing construct, not a storage construct. When you buy one, you choose four things: the region, the access tier (hot, cool or archive), the redundancy option (LRS, ZRS, GRS, GZRS, RA-GRS or RA-GZRS), and the size (a number of 100 TiB or 1 PiB units) for a one-year or three-year term. For every hour, Azure matches your actual standard Blob and Data Lake Gen2 capacity usage against reservations with the same attributes and bills the matched portion at the reserved rate. It applies automatically to existing and future data that fits the terms; you do not point it at a specific account or container.

The discount is use-it-or-lose-it per hour. If, in a given hour, you store less than the reserved amount that matches a reservation's region, tier and redundancy, the unused portion of that hour's commitment is forfeited; it does not roll forward. This is why sizing to the floor matters: a reservation set to capacity you reliably exceed is always fully used, while one set above your real usage wastes the gap every hour. Because the attributes must match, a reservation bought for the hot tier in one region does not discount cool-tier data or data in another region; you would buy separate reservations for those.

Several things are explicitly outside the reservation and stay at pay-as-you-go: early deletion charges, per-operation transaction costs, bandwidth and data transfer. The reservation discounts capacity, the gigabytes stored, and nothing else. It also does not cover premium accounts, general-purpose v1 accounts, page blobs, Queue storage or Table storage; Azure Files reserved capacity is a separate, distinct reservation. Knowing the boundary keeps the saving estimate honest: it is a discount on the capacity line, not on the whole storage bill.

What is the impact of leaving the recommendation unactioned?

The direct impact is simply overpayment. Every month the steady base of standard Blob and Data Lake Gen2 capacity is billed at the pay-as-you-go rate when a matching reservation would discount it, the difference is money spent for nothing in return. Because storage bases tend to be large and to grow, the gap compounds: the longer a stable floor sits unreserved, the more pay-as-you-go premium has been paid that a reservation would have avoided. Advisor quantifies exactly this as the projected annual saving on the recommendation.

The second-order impact is forecast noise. Pay-as-you-go capacity is a variable line that drifts with usage, whereas a reservation converts the predictable floor into a fixed, known monthly cost. Reserving the base makes the storage bill easier to forecast and reserves the variability for the genuinely variable layer on top. Leaving everything on demand keeps the whole line fuzzy even though most of it is, in practice, perfectly predictable.

The risk of acting is bounded and well defined: it is the commitment itself. A reservation is paid for whether or not the capacity is held that hour, so reserving more than you reliably store strands part of the commitment. That risk is fully controllable by sizing to the conservative floor, and Azure provides an exchange and refund route, with documented limits including a maximum refund per year, as a safety valve if the estate changes materially. The asymmetry favours acting on a well-sized floor: the saving is certain and recurring, while the only downside is a forecasting error you control.

How do you capture the saving safely?

Work it as a sizing-and-commit loop, not a one-click accept of the Advisor figure. The order matters: confirm the durable floor before committing, so every reserved unit is covered every hour and the discount stays risk-free.

1. Confirm the durable floor, not the recent peak

Plot stored standard Blob and Data Lake Gen2 capacity over the lookback window for each region, access tier and redundancy combination. Identify the floor: the capacity present every month that the business is confident it will keep for the whole term. Advisor's projection is a sized starting point; this step converts it into a commitment you can defend.

2. Match the reservation attributes exactly

A reservation only discounts usage with the same region, access tier and redundancy, so buy one reservation per combination that has a meaningful floor. Hot-tier data in West Europe and cool-tier data elsewhere need separate reservations. Reserve in units of 100 TiB or 1 PiB, sizing each to its own floor.

3. Choose the term and billing plan deliberately

A three-year term gives a deeper discount than one year but commits for longer; a one-year term is the safer choice where the floor is less certain. Billing can be upfront or monthly at no extra cost, so monthly billing captures the same discount without the cash outlay. Leave the variable layer above the floor on pay-as-you-go.

4. Own the renewal and keep the safety valve in reach

Record the reservation, its scope and its expiry date so the saving is renewed deliberately rather than lapsing back to the full rate. If the estate shifts materially, use the exchange route to re-size into a new Azure Storage reservation, or refund within the documented limits, including the maximum refund per year.

# 1. Pull the Advisor reserved-capacity recommendation and its projected saving.
az advisor recommendation list --category Cost \
  --query "[?contains(shortDescription.solution, 'Blob storage reserved')]" -o jsonc

# 2. Price the reservation BEFORE buying: confirm the discount on the floor you sized.
#    BlockBlob is the reserved-resource-type for Blob + Data Lake Gen2 capacity.
#    --sku, --location, --quantity must match the region/tier/redundancy floor you confirmed.
az reservations reservation-order calculate \
  --reserved-resource-type BlockBlob \
  --sku <blob-reservation-sku> \
  --location westeurope \
  --quantity 1 \
  --term P3Y \
  --billing-plan Monthly \
  --applied-scope-type Shared \
  --display-name "blob-reserved-floor-westeurope"

# 3. Purchase once the calculated price and discount are confirmed.
#    --reservation-order-id is the GUID returned by the calculate step (or generate one).
#    Monthly billing captures the same discount with no upfront cash outlay.
az reservations reservation-order purchase \
  --reservation-order-id <reservation-order-guid> \
  --reserved-resource-type BlockBlob \
  --sku <blob-reservation-sku> \
  --location westeurope \
  --quantity 1 \
  --term P3Y \
  --billing-plan Monthly \
  --applied-scope-type Shared \
  --display-name "blob-reserved-floor-westeurope"

Quick quiz

Question 1 of 5

Azure Advisor recommends 'Consider Blob storage reserved instance to save on Blob v2 and Datalake storage Gen2 costs'. What does acting on it actually change?

Keep learning

Go deeper on how Azure Storage reserved capacity is priced, scoped and applied, so you can size and govern reservations with confidence.

You can now treat the Advisor reserved-capacity recommendation as one capability rather than a notification to ignore: confirm the durable floor of standard Blob and Data Lake Gen2 capacity you reliably hold, match a reservation to each region, tier and redundancy combination, choose the term and billing plan deliberately, and own the renewal so the saving never lapses. Nothing about the data changes; only the unit price on the steady base you already store.

Back to the library

Blob storage reserved capacity: the savings view

Turning a steady, predictable storage base into a committed discount with a clear payback

This is one of the cleanest savings levers in Azure. The data you are storing does not change, the way it is served does not change, and the only thing that moves is the unit price on the predictable base of capacity you hold every month. Advisor estimates the saving for you against your own recent usage, so you are not guessing: the recommendation comes with a projected annual saving sized to the floor you have actually been storing.

The trade-off is a genuine commitment. A one-year or three-year reservation is a sunk amount that you pay whether or not you use it that hour, so the risk is not a price risk but a utilisation risk: reserving capacity you later delete. The finance job is to size the reservation to the conservative floor, the amount you are confident you will hold for the whole term, so the committed units are always covered and the discount is effectively risk-free. The variable layer above the floor stays on pay-as-you-go.

The payback is straightforward to reason about because the reserved rate is a fixed percentage off pay-as-you-go for the same capacity, tier and redundancy. The break-even is simply whether you hold the reserved amount for enough of the term to recover the commitment, and for a floor you size conservatively, that is true from day one. Exchanges and refunds exist as a safety valve if the estate shifts, with documented limits.

This lesson is for the finance partner who sees the Advisor reserved-capacity recommendation with its projected saving and wants to know how solid the number is, what the commitment risk really is, and how to size the reservation so the discount is captured without stranding committed capacity. It covers the utilisation risk, the break-even logic, and the exchange and refund safety valve.

Fun fact

The discount that hides in plain sight

How a finance partner pressure-tests the projected saving

Priya is the finance partner for the platform team. The Advisor recommendation lands with a projected annual saving in one region, and the instinct is not to wave it through on the headline figure but to ask one question: how much of that committed base is the business genuinely certain to keep for the whole term? Advisor sized the recommendation against recent usage, which means it can be anchored to a recent peak rather than the durable floor. The finance move is to separate the two.

Priya asks the team to plot stored capacity over the lookback window and identify the conservative floor: the amount of standard Blob and Data Lake Gen2 data that has been present every single month and is not going anywhere, in that specific region, tier and redundancy. The reservation is sized to that floor, not the peak. The pack reads: 'The reserved rate is a fixed discount off pay-as-you-go for capacity we are already storing. Sized to the floor, every committed unit is covered every hour, so the discount is risk-free and recurs for the whole term. The variable layer above the floor stays on pay-as-you-go, and we keep the exchange and refund route open if the estate shifts.' The recommendation goes from a headline number to a sized, defensible commitment with a clear payback.

Why the reserved-capacity saving is a recurring line, not a one-off

The value here compounds in a way one-off savings do not. A reservation discounts the matching capacity every hour for the full term, so a saving captured once recurs for one or three years automatically. Conversely, every month the recommendation sits unactioned is a month of pay-as-you-go premium paid on capacity that was never going anywhere. The Advisor projection is the annualised version of that recurring gap.

The cost model is asymmetric but, unlike a security control, the downside is a number you set yourself. The saving is a fixed, documented discount off pay-as-you-go for matched capacity; the only risk is utilisation, reserving units you then stop storing. Size to the floor and that risk goes to near zero, because the committed amount is covered every hour. The finance role is to anchor the reservation to the durable floor rather than the recent peak, lay out the break-even against the term, keep the variable layer on demand, and hold the exchange and refund route in reserve for genuine estate changes.

What finance can do to capture the saving cleanly

Finance cannot size storage, but it owns the framing that turns an Advisor figure into a sized, defensible commitment with a clear payback and a controlled downside. Three levers.

1. Anchor the commitment to the floor, not the peak

Make explicit that the reservation must be sized to the durable floor, the capacity the business is confident it will hold for the whole term, not to the recent usage peak Advisor may have anchored to. Every reserved unit that sits under real, persistent usage is discounted every hour; anything above the floor is utilisation risk. Sizing to the floor is what makes the discount effectively risk-free, and it is a finance judgement as much as an engineering one.

2. State the break-even and the term trade explicitly

The reserved rate is a fixed discount off pay-as-you-go for matched capacity, so the break-even is simply holding the reserved amount for enough of the term to recover the commitment, which a conservatively sized floor clears from day one. Lay out the one-year versus three-year choice as a depth-of-discount against length-of-commitment trade, and prefer the longer term only where the floor is genuinely certain.

3. Diarise the renewal and keep the exchange route documented

The most common way this saving is lost is a reservation lapsing unrenewed back to the full rate, so the expiry date belongs on the finance calendar with a named owner. Document the exchange and refund route, including the maximum refund per year, as the safety valve if the estate changes, so a commitment never feels like a trap and the team is willing to size it properly in the first place.

Quick quiz

Question 1 of 5

The Advisor recommendation shows a projected annual saving. As the finance partner, how should you treat that number before committing?

Keep learning

Go deeper on how Azure Storage reserved capacity is priced, scoped and applied, so you can size and govern reservations with confidence.

You have finished the finance view of Blob reserved capacity. You know the saving is a fixed, recurring discount off pay-as-you-go for capacity the business already holds, that the only risk is utilisation and it is controlled by sizing to the conservative floor, and that the exchange and refund route is the safety valve for genuine estate changes. Next time the Advisor recommendation lands, you will anchor the commitment to the durable floor, state the break-even and term trade, and diarise the renewal rather than treating the projected number as either gospel or noise.

Back to the library

Blob storage reserved capacity: the headline

A risk-free discount on storage you already hold, gated only by how confidently you can commit

There is a standing discount available on the storage the business already keeps, and the only question is how much of the steady base you are willing to commit to for one or three years. Azure Advisor has already done the sizing against real usage and put a projected saving against it. This is found money on the existing bill, not a new project.

The decision is not technical and carries no operational risk: the data, the accounts and the durability all stay identical. The only exposure is committing to capacity the business later stops needing, which is a forecasting question, not an engineering one. Size the commitment to the floor you are confident about and the discount is essentially free; the variable layer on top stays flexible.

The leadership move is to make sure reservations are reviewed and renewed deliberately rather than left to expire silently back onto the full rate. The saving is real and recurring, and the main way to lose it is inattention.

A short read for the leader who wants to know why a discount on storage the business already holds is sitting unclaimed, what committing to it does and does not risk, and what governance keeps the saving from quietly lapsing at renewal.

Fun fact

The discount that hides in plain sight

What it looks like when a standing discount is actually captured

At a cost review the CFO noticed that storage was one of the larger recurring lines and asked the obvious question: is any of this on a committed discount, or are we paying full rate for data we have held for years? The honest answer at the time was that the estate was entirely pay-as-you-go, even though Advisor had been recommending a Blob reservation with a projected saving for months.

The team's response was to treat the recommendation as a deliberate decision rather than a notification to ignore. They confirmed the durable floor of storage the business holds in its main region, purchased a reservation sized to that floor for a three-year term to capture the deeper discount, and put the renewal date on the finance calendar so it could never lapse silently. Nothing about the service changed: same data, same durability, same performance, lower unit price. The next time the CFO asked, the answer had changed from 'no, it is all on demand' to 'the steady base is on a reserved rate that recurs automatically, the variable layer stays flexible, and the renewal is owned and diarised.' That shift, from leaving a sized discount unclaimed to capturing it as a governed, recurring saving, is the whole value of this capability.

Why a sized, unclaimed discount is a leadership issue

Leaving a sized, low-risk discount unclaimed on a recurring bill is a governance gap, not a technical one. Advisor has already done the work of estimating the saving against real usage; the only reason it goes uncaptured is that committing to it requires a deliberate decision and an owner. This is the rare cost lever where the saving is certain, the operational risk is nil, and the only exposure is a forecasting judgement that leadership can frame.

The cost of inaction is a steady premium on the existing bill, and the cost of action is a commitment sized to capacity the business is confident it will keep. Because the saving recurs and the main way to lose it is letting a reservation lapse unrenewed, the leadership requirement is light but real: own the decision, size it conservatively, and make sure renewals are deliberate rather than accidental.

The leadership move on reserved capacity

The executive handle is not to approve each reservation line by line; it is to require that the standing discount on the steady storage base is captured, sized conservatively, and renewed deliberately.

1. Require the sized discount to be claimed, not just noted

Advisor has already projected the saving against real usage. Ask for a decision on it rather than letting the recommendation sit as a perpetual notification. The default of leaving the steady base on pay-as-you-go is a recurring premium with no offsetting benefit; capturing it is found money on the existing bill.

2. Frame the only real risk as forecasting, and bound it

There is no operational risk: the data and durability do not change. The single exposure is committing to capacity the business later stops needing. Require that reservations are sized to the floor the business is confident about, with the variable layer left flexible, so the commitment is bounded and the discount is, in practice, risk-free.

3. Make renewals deliberate and owned

The saving recurs automatically for the term but does not renew on its own, and Azure only warns 30 days out. Put renewal dates on the finance calendar with an owner so the discount is never lost to silence, and treat the exchange and refund route as the documented safety valve if the estate shifts materially.

Quick quiz

Question 1 of 5

A CFO asks whether any of the steady storage base is on a committed discount. What is the defensible end state to aim for?

Keep learning

Go deeper on how Azure Storage reserved capacity is priced, scoped and applied, so you can size and govern reservations with confidence.

Two takeaways. Reserved capacity is the rare cost lever where the saving is certain and recurring, the operational risk is nil, and the only exposure is a forecasting judgement you can bound by sizing to the floor. And the main way to lose the saving is silence: own the decision to claim it, size it conservatively, and make renewals deliberate so a sized, low-risk discount on storage you already hold is never left unclaimed or allowed to lapse.

Back to the library

Part of the learning path Lock in Azure commitments

Buy Blob storage reserved capacity

Blob storage reserved capacity: the basics

The discount that hides in plain sight

Reading the reserved-capacity recommendation across a subscription

How the reserved-capacity discount actually appliesdeep dive

What is the impact of leaving the recommendation unactioned?

How do you capture the saving safely?

1. Confirm the durable floor, not the recent peak

2. Match the reservation attributes exactly

3. Choose the term and billing plan deliberately

4. Own the renewal and keep the safety valve in reach

Quick quiz

Keep learning

Blob storage reserved capacity: the savings view

The discount that hides in plain sight

How a finance partner pressure-tests the projected saving

Why the reserved-capacity saving is a recurring line, not a one-off

What finance can do to capture the saving cleanly

1. Anchor the commitment to the floor, not the peak

2. State the break-even and the term trade explicitly

3. Diarise the renewal and keep the exchange route documented

Quick quiz

Keep learning

Blob storage reserved capacity: the headline

The discount that hides in plain sight

What it looks like when a standing discount is actually captured

Why a sized, unclaimed discount is a leadership issue

The leadership move on reserved capacity

1. Require the sized discount to be claimed, not just noted

2. Frame the only real risk as forecasting, and bound it

3. Make renewals deliberate and owned

Quick quiz

Keep learning

Related cost lessons