Why Tritax Big Box Sold New Shares Below What Its Buildings Are Worth

In early August, Tritax Big Box REIT — the £4.6 billion London-listed landlord of giant distribution warehouses that is now trying to become a data-centre developer — sold about 213 million new shares at 164 pence each and raised £350 million, a bit under $470 million. That part is routine. The strange part hides in the headline, which read like nothing at all: "Issue of Equity."

Here is the weird bit. At the end of June, Tritax reported its net tangible assets at 185.9 pence per share — essentially the value of its buildings minus debt, divided by the shares outstanding, the way U.K. REITs measure their net asset value. New investors paid 164 pence for a share stamped with 185.9 pence of net assets. They paid roughly 88 pence for a pound's worth of property. The placement was struck at an 11.8% discount to net asset value, which you get when you price new stock below the number on the balance sheet.

That matters because one of the oldest rules of property investing says a real estate company that issues shares at a discount to net asset value is doing something bad for the people who already own it. There are two reasons. The first is arithmetic. Each new share brings in 164 pence of cash but collects a claim on 185.9 pence of net assets, and the 21.9-pence gap comes out of other shareholders' pockets. Because the new shares are only about 8% of the company's capital, the total damage is small — rough maths puts net asset value per share at about 184 pence afterward, a bit under a 1% ding to the existing holders. The second reason is darker and is the one academics have been pointing at for decades: if a company needs money and the best it can do is sell equity below book value, that is usually a confession that the book value is not all it appears, or that the cash position is worse than it looks. When equity issuance bears that signal, the stock tends to fall.

Tritax's shares did fall — about 3.6% to 165.5 pence when trading resumed — but that was still above the 164 pence placement price, a detail that matters. A market that thought the company had just sold itself for a song would have left the stock below where the new buyers paid. Instead the raise was, in the words of chief executive Colin Godfrey, "massively oversubscribed", and the new shares were merely money that had to be absorbed.

The reason the discount did not spook anyone is the destination of the money. Tritax is not raising capital to buy more of the same warehouses. It has spent the last year converting its land platform into a data-centre pipeline — the strategy is, basically, hunting for sheds that come with power attached. The new funds pay for two data-centre schemes in Greater London, for which the company secured 235 megawatts of grid connections, nearly doubling its secured power supply to 507 megawatts. That is the scarce input: in Britain, a grid connection is the long-lead, hard-to-get item, and a developer holding land plus power plus planning permission owns the thing other people cannot buy. The company got planning permission in July for its Manor Farm data centre near Heathrow, and in June it announced a 147-megawatt project it aims to complete in phases from late 2027.

The pitch for the London projects: yields on cost of 9-11%, £50-60 million of additional rent in time, and "capital profits" of £300-400 million — development profit of more than 50% on the cost of building. That is the mechanism. A big-box warehouse portfolio throws off a running yield of roughly 5-6%; a data centre is pitched at nearly double that, and when the finished building is let and revalued, the gap between cost and value is where the development profit lives. Management called the raise "materially accretive" to earnings and raised its adjusted earnings-per-share growth target from about 50% by 2030 to about 65% by 2030/31. That, in one sentence, is the strategy: give us 164 pence, and instead of just adding it to 185.9 pence of buildings, we will spend it converting our own land — with power nobody else can get for years — into assets the market currently prices at much higher multiples. The ~1% dilution is the price of the conversion, and the claimed upside is an order of magnitude bigger. The whole trade is a bet that the data-centre economics are real.

It is also, notably, a bet about what kind of company Tritax is, because the plumbing of the raise betrays it. At about 8% of share capital, the issue could be done as a quick, non-pre-emptive placing — within roughly the tenth of a FTSE-listed company's shares that a board is authorized to allot to whoever it chooses between shareholder votes — instead of a rights issue giving every existing shareholder first refusal at the discounted price. Institutions took the bulk, about £338 million of it. Ordinary investors got a slice through a small "RetailBook" tranche, about £10.6 million, and even a request from Blackstone, which holds an 8.6% stake, for £30 million was cut to £3 million because the book was that deep. Retail holders who wanted 164 pence shares mostly did not get them. That is how the dilution this issue created was allocated: to the institutions who bid fastest, at a price most of the people who took the hit never saw.

For a U.S. investor the first lesson is the correct reaction to the phrase "equity issue." It is not automatically doom. The useful move is per-share arithmetic: what is the discount, how much of the company is being issued, and what does the money earn. Tritax's raise fails the old "never issue below net asset value" test loudly and passes the modern version of it — small dilution, high claimed deployment yield — which is why the stock barely blinked.

The second lesson is that this is now a specific long-cycle bet, and the things to watch are concrete. Tritax is a FTSE 100 name now (it joined the index in March), reachable from the U.S. through the London listing under the ticker BBOX or over the counter as TTBXF, with currency risk and the usual U.K. tax quirks on REIT payouts attached. Even now the stock trades around 160 pence, still roughly 14% below the 185.9 pence of net asset value per share. What makes the raise look clever, later: the two London data centres coming in near the 9-11% yield-on-cost pitch by 2030-31, the £300-400 million of development profits showing up in revaluations, and the market paying for the company as something other than a sheds landlord once the sheds discount recedes. What makes it look like the classic below-NAV warn-off: missed dates, letting disappointments, yields that land lower than pitched, or the cost of carrying undeveloped land eating the spread before the revaluation arrives.

The announcement itself — generic headline, no drama — is a fairly honest document of that trade. Existing shareholders gave up a bit under 1% of net asset value per share, and the company handed the discount to whoever wanted it most, on the promise that the money could be put to work in the one asset class the market is paying up for. For a landlord whose shares have traded below its buildings for years, that is a rational trade. It is only a good one if the data centres actually get built.