Summary

  • Backblaze estimates that two initial CoreWeave order forms, lasting five and seven years, have an aggregate value of about $335.1 million. The filing warns that actual amounts depend on storage capacity used and other factors.
  • The agreement added roughly $313.4 million to remaining performance obligations after about $21.7 million of customer-warrant value was deducted. That subtraction is an editorial calculation from the disclosed figures, not a separate company forecast.
  • At 30 June, Backblaze reported $396.0 million of total RPO, of which $70.4 million was due within one year and $325.6 million thereafter. It had recognised no revenue from the new arrangements by that date.
  • Delivery creates another ledger: revenue starts only as committed capacity becomes available or managed storage is deployed. Hardware leases, data-centre costs and up to 4,194,876 warrant shares determine what the contract ultimately contributes to cash and ownership.

A backlog appeared before the service

The most striking number in Backblaze’s June-quarter accounts is not revenue. It is $396 million of remaining performance obligations, up sharply after the company signed a storage agreement with CoreWeave on 16 June. The announced total contract value is about $335 million. For a company that recorded $42.7 million of quarterly revenue, the scale is large enough to dominate a first reading.

But it did not dominate the quarter’s sales. Backblaze says revenue recognition for the new committed-capacity and managed-storage arrangements was expected to begin in the third quarter. None was recognised in the three or six months ended 30 June. The contract entered the backlog before it entered reported revenue.

That is not a contradiction, and it is not an accounting footnote to be waved away. It is the commercial mechanism. CoreWeave has ordered two different forms of storage service: capacity on Backblaze’s B2 platform in Backblaze data centres, and a managed storage solution installed in CoreWeave data centres. Backblaze must make the first capacity available and deploy the second service before the corresponding revenue clock starts.

The $335.1 million headline therefore belongs to the first of four ledgers. It is an estimate of value across initial order forms with terms of five and seven years. Backblaze’s 8-K explicitly says the amount actually payable will depend on the storage capacity used and other factors and may differ from the estimate. It is neither cash in the bank nor an invoice for the whole period.

The second ledger is smaller by $21.7 million

The second ledger is RPO. Backblaze defines it to include deferred revenue and future committed revenue under existing customer contracts. That makes it broader than the $35.4 million of deferred revenue on the June balance sheet. Deferred revenue is the invoiced portion for work still outstanding; RPO also extends to future commitments before an invoice is issued.

The CoreWeave agreement does not enter this ledger at the gross headline. The 10-Q says the approximately $335.1 million aggregate contract value is reduced by about $21.7 million attributable to the fair value of warrants issued to the customer. The difference is about $313.4 million. Backblaze’s investor presentation rounds that contribution to roughly $313 million.

This treatment describes customer acquisition consideration, not a financing inflow. Backblaze granted CoreWeave warrants to buy up to 4,194,876 Class A shares at an exercise price of $7.60. Accounting rules treat the grant-date value expected or probable to vest as a reduction of the customer contract’s transaction price. That reduction will run through revenue in the same pattern as the related service revenue.

The result is easy to miss in a headline. A strategic customer can improve demand visibility while receiving a potentially valuable equity-linked benefit. The gross contract estimate measures the service relationship before that consideration. RPO records the contract after the warrant value has been taken into account. Neither figure shows how much capacity is ready today.

Timing reinforces the distinction. Of Backblaze’s $396.0 million total RPO at quarter-end, $70.4 million was classified within one year and $325.6 million beyond one year. Those amounts equal approximately 17.8% and 82.2% of the total, respectively. The percentages are simple editorial calculations, not a revenue schedule: timing can still move with capacity availability, usage, order terms and other contract conditions. They do show that most of the visibility is long-dated.

The comparison base also needs care. Backblaze changed its RPO disclosure in 2026 to include contracts with original durations of one year or less, then recast the December 2025 comparative number upward by $4 million. The $70.2 million year-end figure is usable because it has been conformed, but it is not the number originally published on precisely the same basis.

Capacity controls the revenue clock

The third ledger begins with a physical act. Under committed-capacity arrangements, Backblaze treats reserved storage capacity over the contract term as a single performance obligation. It recognises revenue according to the proportion of contracted capacity made available. Under managed-storage arrangements, revenue begins when the applicable capacity has been deployed and the service is available to the customer.

That distinction prevents a contract announcement from becoming revenue by declaration. A signed order can enter RPO. Only deployed or available service can enter the income statement. Cash collection has its own timing again.

The new mechanism also changes how Backblaze is sold. Its traditional B2 consumption business recognises fees as customers use the platform; subscription services are recognised over their service periods. The CoreWeave arrangement adds a committed-capacity model large enough to make deployment schedules material to the company’s reported growth. Backblaze itself says the pace of future B2 growth will depend on customer deployment schedules, usage levels and the commencement of contracted services.

The June quarter offers a clean baseline. B2 revenue was $26.6 million, up 34% year on year, and total company revenue was $42.7 million. The filing attributes the quarterly B2 increase to $4.5 million from new customers, $1.5 million from May’s pay-as-you-go price increase and $0.7 million from higher use by existing customers. It simultaneously says that no revenue from the new committed and managed arrangements had yet been recognised. The strong reported growth and the large new backlog are adjacent facts, not the same fact.

The fourth ledger pays for the promise

Storage capacity requires hardware, space, power, network and people before it becomes an accounting performance obligation satisfied. Backblaze’s quarter already shows a heavier physical base. Property and equipment rose to $70.1 million from $57.3 million at December. Current and non-current finance-lease liabilities together reached $46.7 million, up from $36.2 million.

The company generally uses finance leases to obtain hard drives and other infrastructure equipment. At 30 June, those leases carried a weighted-average discount rate of 12.9%. On the same day, Backblaze amended its credit agreement to raise the permitted amount of capital-lease indebtedness to $150 million. That ceiling is not borrowing already incurred, and the asset increase cannot be assigned wholly to CoreWeave. Both disclosures show the financing channel available for a larger deployment programme.

Cost of revenue has begun to reflect expansion even before the new contract’s revenue. In Q2, it rose by $2.6 million year on year. Backblaze attributes $1.0 million to additional depreciation on infrastructure placed in service and $0.9 million to extra rent and facility costs for storage-capacity growth. Gross margin remained 63%. These costs cover the wider business, not a disclosed CoreWeave profit-and-loss account. They nevertheless show why a capacity contract cannot be valued only by its top line.

The warrant channel is part of this fourth ledger too. The initial warrant covers 3,053,314 shares and vests in twenty equal quarterly instalments over five years while the related order form continues. The additional warrant covers 1,141,562 shares; its tranches depend on managed-storage purchase orders and then vest over time. If an order form is terminated, future vesting stops, although already vested portions can remain exercisable. Both warrants can be exercised in cash or on a cashless basis.

At June-end, the full warrant share count represented about 6.8% of outstanding Class A shares. That is a maximum exposure, not present dilution: all shares were still shown as unvested at the reporting date. Yet it is economically real consideration. If warrants vest and are exercised, existing owners may be diluted; if exercised cashlessly, Backblaze may receive no cash proceeds. The customer’s incentive, the vendor’s reported revenue and the shareholder’s ownership percentage are linked but do not move at the same time.

A large agreement, with the proof still ahead

Backblaze raised its 2026 revenue outlook to $172 million–$174 million and presented a directional view of B2 revenue growth above 40% in 2027. The CoreWeave agreement plainly supports that confidence. It does not supply a disclosed quarterly bridge from the $313.4 million RPO contribution to either forecast.

The agreement matters because it moves Backblaze up the infrastructure stack. It is no longer only selling self-service object storage or ordinary enterprise subscriptions. One part of the arrangement places committed capacity on its own platform; the other has Backblaze operate storage inside a major AI cloud provider’s facilities. This can create a durable, embedded relationship and a stronger planning base.

It can also create concentration, execution and capital risk. A customer large enough to transform backlog can become large enough to shape capacity decisions. A multi-year order can stabilise planning, yet its value still depends on deployments, capacity use and continuation terms. Warrants align incentives, but they ask shareholders to share some of the upside before the income statement has recorded the service.

The market should therefore resist choosing between two crude stories: “$335 million is already won” and “RPO is meaningless until it becomes cash”. The contract is meaningful precisely because it has passed the commitment test. It remains unfinished because commitment, deployment, revenue, cash and ownership are separate receipts.

Sources