Summary

  • Digital Realty’s $7.8bn figure values the three Northern Virginia assets at 100% share and includes assumed debt plus estimated capital still needed to complete development; it is not the cash paid to Blackstone.
  • Consideration for Blackstone’s disclosed blended 64% interests was about $3.5bn: $1.231bn in cash and 12,310,249 non-voting shares calculated from $2.346bn of stock value.
  • The three sites total 288MW of expected critical IT capacity and are 100% leased, but the filings tie completion, rent commencement and Core FFO accretion to 2027 and 2028.
  • Digital Realty already held the remaining blended 36%. The 10-Q treats the step-up as an asset acquisition and includes that prior interest’s net book value in purchase consideration without recognising a remeasurement gain or loss.

A $7.8bn data-centre transaction appears to invite one question: how did the buyer pay? Digital Realty’s filings require a different first question: what exactly does each number measure?

The largest figure is not a cheque. The smallest percentage is not a simple average. The 288MW is not a meter reading. “100% leased” is not the same state as rent commencement. Once those boundaries are restored, the acquisition becomes more intelligible—and more dependent on delivery after closing.

Four ledgers sit behind one announcement

On 29 June 2026, Digital Realty agreed to acquire all of Blackstone affiliates’ interests in the Digital Carver Dulles 9 and Digital Carver Brickyard joint ventures. The deal covered Blackstone’s 80% positions in two 96MW data centres in Manassas and its 50% position in a 96MW facility on Digital Realty’s Dulles campus in Sterling. The filing calls the acquired package a blended 64% partnership interest. Digital Realty already owned the remaining blended 36%; closing made the ventures wholly owned subsidiaries of its Operating Partnership.

The first ledger is the $7.8bn gross value at 100% share. Digital Realty explicitly says that figure includes assumed debt and estimated remaining capital expenditure required to complete the ongoing development. It therefore describes the whole asset perimeter under a completed-development frame. It is not restricted to Blackstone’s interest, and it is not restricted to money transferred at closing.

The second ledger is consideration for the seller’s interest. The exact 8-K terms were $1.231bn in cash plus non-voting stock whose quantity was calculated by dividing $2,346,087,437.83 by Digital Realty’s 29 June closing share price and rounding down. The subsequent closing filing identifies 12,310,249 shares. The 10-Q presents the package in rounded form: approximately $3.5bn of consideration, including $1.2bn cash and 12.3m shares.

The third ledger is capital structure. The shares initially issued to Blackstone were non-voting but otherwise carried substantially the same economic rights as ordinary common stock. They converted one-for-one when transferred outside Blackstone’s affiliate group. Blackstone then sold 12,310,249 common shares to public investors at $185 each on 1 July. Digital Realty received no proceeds from that secondary sale. The $185 resale price is consequently not a second payment to Digital Realty, nor is it the cash purchase price of the data centres.

The fourth ledger is capacity and income timing. The three facilities contain 288MW of expected critical IT capacity—96MW each. The word “expected” matters because the filing also says development remained under way. Two facilities were expected to stabilise in the first half of 2027 and the third in the first half of 2028. Digital Realty linked anticipated Core FFO accretion to those years, “as development is completed and rents commence”.

Put together, the ledgers explain why subtracting $3.5bn from $7.8bn does not reveal a bargain, hidden payment or instant gain. The two figures have different denominators and include different obligations.

The 64% is a disclosed blend, not mental arithmetic

The facility-level interests—80%, 80% and 50%—do not average to 64%. The filings nevertheless consistently identify the acquired partnership interests as a blended 64% and Digital Realty’s prior interests as a blended 36%. That is a warning against reconstructing the blend from capacity alone. Equal megawatts do not prove equal venture value, equal capital accounts or equal economic weighting.

The public documents do not provide the full asset-by-asset weighting required to reproduce the blend. The disciplined reading is therefore to preserve both layers: the specific ownership positions at each facility and the company’s disclosed blended partnership percentage. Replacing either with an improvised average would manufacture precision the filings do not supply.

An asset acquisition changes the accounting bridge

Digital Realty’s Q2 10-Q classifies the transaction as an asset acquisition. Because the company already held the remaining blended 36%, the accounting purchase-consideration frame includes the net book value of that previously held interest. No gain or loss was recognised on it.

That detail matters. A reader looking only at $1.231bn cash and 12.31m shares sees what moved to the seller, not the entire accounting basis brought into the now wholly owned ventures. A reader looking only at $7.8bn sees a 100%-share gross-value frame that also contains assumed debt and future completion spending. Neither view, on its own, is the purchase accounting.

The filing also records $201m of promote income and $14m of promote expense at closing. Those amounts reflect incentive economics achieved when the acquisition closed. They are not extra cash purchase consideration, and they should not be used to force the headline values into a false one-line reconciliation.

The missing line items are as important as the disclosed ones. The three filings do not separately state the assumed debt embedded in $7.8bn, the remaining completion capex, the book value of Digital Realty’s old 36% interest, or a valuation for each facility. Without those pieces, the public record supports a perimeter map, not an exact bridge from $7.8bn to $3.5bn.

Fully leased is not yet fully stabilised

All three data centres were described as 100% leased to three distinct investment-grade hyperscale customers under 15-year leases, with weighted average Aa3/AA- credit and 3.6% annual rent escalators. Those are material protections. They reduce speculative lease-up risk and give Digital Realty a contractual path to long-duration income.

They do not collapse the construction clock. The 8-K says the tenants are subsidiaries of investment-grade parents, while the parent entities are not parties to the leases and may not be legally required to cure a subsidiary default. It also says the development must be completed and rents must commence before the expected accretion appears.

“100% leased” therefore describes contractual allocation of future space and power. “288MW” describes expected critical IT capacity. “Stabilised” requires the projects to reach their operating and income state. These conditions can eventually coincide, but the filing does not say they already had.

The same distinction limits the advertised initial stabilised capitalisation rate of more than 6.5%. It is an expectation attached to a completed, stabilised asset frame. It is not evidence of current cash yield on the $1.231bn cash outlay, nor proof that every megawatt was already energised and billable.

What the transaction actually changed

The closing transferred the remaining venture interests to Digital Realty, consolidated ownership and replaced a third-party capital structure with wholly owned operating subsidiaries. It also shifted part of the seller consideration into Digital Realty equity rather than an all-cash purchase. Those are immediate control and financing changes.

The operating transformation is slower. Construction completion, commissioning, rent commencement and stabilisation remain dated into 2027 and 2028. The transaction acquired the right to those economics and the obligation to finish the path toward them. It did not move future capacity backward in time.

That is the correct reading of the headline: $7.8bn describes the whole developed-value perimeter; roughly $3.5bn describes what Blackstone received for its disclosed blended interest; $1.231bn and 12.31m shares describe the payment instruments; and 288MW describes an expected capacity base whose income clock was still ahead.

Sources