Premiums Don't Raise Capital. Sponsors Do. — two data centre REITs trade above NAV, and only one premium can actually be spent

Two data centre trusts trade above the value of the assets they hold. One of those premiums sits on a deep secondary market that has supported equity fundraising. The other is a print so sharp that exchange trading-control measures suspended trading the next morning.

Keppel DC REIT closed 2025 at S$2.25 per unit against a reported net asset value of S$1.71 — a premium of roughly 1.32x, struck in a deep, continuously traded market of about 2.44 billion units. The GDS C-REIT (the Southern GDS Data Centre REIT, SSE: 508060) also trades above NAV — at a premium that has swung between roughly 1.28x and 1.49x across three official closes from August 2025 to July 2026 — but the top of that range came from a different kind of event: an early-January 2026 spike to RMB 4.515, about 49% above the audited NAV of RMB 3.0291 per unit, followed the next morning by a temporary trading suspension after the exchange's trading-control measures were triggered.

Edition 17 asked whether the scarcity premium on the GDS C-REIT's first listed asset could fund the second — the Langfang Shucheng expansion it filed for on 11 July 2026. That verdict is still gated: as of writing, the required independent valuation report and issue documents have not yet been disclosed. So this edition takes the question one layer down. Before you ask whether a premium travels to the next asset, you have to ask what the premium actually is — and whether the two most-cited data centre premiums in the region are even the same kind of number.

Keppel DC REIT (SGX: AJBU) and the GDS C-REIT — the Southern GDS Data Centre REIT, SSE: 508060 — both trade above NAV, but they are built differently. Keppel DC REIT is a conventional listed trust that holds its data centres directly and raises equity in a deep secondary market. The GDS C-REIT is a China public infrastructure REIT (公募REIT) structured over an asset-backed securities (ABS) plan — the fund subscribes to the plan that holds the asset rather than owning it outright — and its move to a second asset runs through a formal, rule-priced placement mechanism (定向扩募). (GDS here is GDS Holdings, the data-centre operator and originator-sponsor; the listed vehicle is the C-REIT, managed by Southern Fund Management.)

One clarification, because "ABS" is doing two different jobs here. In the US, data-centre ABS is a mature debt market: issuers securitise data-centre lease cash flows into rated term notes. It is large — about US$25 billion of data-centre securitisation was issued in 2025. Vantage Data Centers, for example, raised €640 million in June 2025 against four fully-leased German facilities, following earlier US and UK deals. Europe entered in 2024; APAC has not priced one yet. Australia is reported as likely to bring the region's first data-centre securitisation in 2026 — market participants have flagged a deal within 6–12 months, with forecasts of A$2–7 billion at maturity — but as of writing nothing has issued.

That US and European market is debt. The GDS C-REIT is different in kind: it is equity, wrapped over an ABS plan but listed and trading at a NAV premium. It is the closest live example anywhere of a listed data-centre vehicle trying to convert its own market premium into acquisition capital for the next asset — which is why the comparison is drawn against a China C-REIT rather than a debt securitisation.

Keppel DC REIT — a sustained premium to NAV. ~1.32x is price over reported NAV per unit — S$2.25 over S$1.71, both as at the same date, 31 December 2025, so numerator and denominator agree on when they were taken. This is the one date-matched observation; over the past year the price has moved (roughly S$2.15–2.44), so the premium has fluctuated too, but within an orderly band in a deep, continuously traded market. [REPORTED] for the date-matched figure; the wider band is [DIRECTIONAL], using the 31 December 2025 audited NAV purely as a common reference point.

The GDS C-REIT — a premium against its own NAV too, but a volatile one. Across three official closing prices, the GDS C-REIT's premium to its latest audited NAV available at publication (RMB 3.0291, as at 31 December 2025) has swung from about 1.35x (RMB 4.095, 11 August 2025) to about 1.49x (RMB 4.515, 8 January 2026 — the spike that triggered a same-morning trading suspension) to about 1.28x (RMB 3.862, 29 July 2026, the most recent official close). Each price is compared against that same audited NAV — the most recent published, since no later audited figure exists yet — which has moved little from the fund's RMB 3.00 inception value, so the multiples are close approximations rather than to-the-day figures. [REPORTED]

The interesting observation is not which trust has the larger premium — on its most recent official close the GDS C-REIT's ~1.28x sits just below Keppel DC REIT's ~1.32x — but the character of each. Keppel DC REIT has traded at a sustained premium formed in a deep, continuously traded market; the GDS C-REIT's official observations include a spike sharp enough to trigger an exchange trading suspension. Both are validly measured against their own NAV; what differs is what each premium is made of.

A premium is not itself a financing mechanism. It is a price signal — it becomes capital only when a sponsor issues into it, and which route is even open depends on market depth, price stability, and issuance method.

Keppel DC REIT funds acquisitions through conventional listed REIT capital raising. Premium → cheaper equity → accretive issuance → asset. To help fund its 2025 acquisitions, Keppel DC REIT ran a pro-rata, non-renounceable preferential offering — about 168% subscribed — that raised gross proceeds of S$404.5 million and listed 180.6 million new units on 22 October 2025, alongside debt. The premium is not priced into the acquisition directly; it lowers Keppel DC REIT's cost of equity, which makes issuing new units accretive, which in turn funds the purchase. A decade-old trust with deep float can raise this way because the market is there to raise into. It has a market, so it raises the conventional way.

The GDS C-REIT reaches asset two by a different route. Premium → directed-placement pricing → acquisition consideration. A fund whose secondary market remains relatively shallow — without the depth of trading liquidity enjoyed by mature Singapore REITs — reaches its second asset through a directed placement (定向扩募) of new units, priced off a formula floor rather than a live clearing price. Under SSE Guideline No. 3, such a placement must price at no less than 90% of the fund's average trading price over the 20 trading days before the pricing base date — a floor keyed to a future date, deliberately insulated from any single January spike.

The progression matters. Same-looking premium; different market structure; different playbook. The premium raises nothing on its own. What raises capital is a sponsor converting it through the route its market depth allows — and that act of conversion, not the headline ratio, is where value is made or lost.

Keppel and GDS asset-1 figures are issuer-reported (Keppel FY2025 statements + Oct-2025 offering results; Southern Fund / SSE filings). The Langfang column is deliberately left open: the independent valuation report (资产评估报告) and issue notice (发售公告) have not been disclosed, so its terms cannot be scored yet.

Not "how big is the premium?" but: priced off what base, in how deep a market, and paid to whom? If the directed placement clears at a rule-based floor to aligned holders and the platform can repeat it, the premium travelled. If successful placement relies primarily on mandatory sponsor participation rather than broad third-party demand, then it was sponsor support — not broad market demand — doing the load-bearing.

The comfortable reading is that APAC's data centre trusts all now carry a scarcity premium, and that the premium is the asset. The uncomfortable reading is that "premium" is doing too much work. The premium raises nothing on its own — what raises capital is a sponsor converting it through the route its market depth allows, and that act of conversion, not the headline ratio, is where value is made or lost.