DALLAS, July 22, 2026 Csquare gave the IPO market a clean read on what public buyers will and will not fund in AI-adjacent infrastructure. The pricing got 50 million shares sold and onto the NYSE, but only after the company cut the deal to $21 a share from the $23 to $27 range it marketed on July 6. That still left Csquare with a $1.05 billion IPO, large enough to matter on any calendar, yet the discount is the part worth watching.
Csquare is selling a timely asset. On its website the company says it operates 80 data centers, 3.5 million square feet and more than 500 megawatts across 30 markets, while its IPO materials describe a carrier-neutral colocation and interconnection platform serving enterprise, network, cloud and technology customers in the U.S., Canada and the U.K. The site also leans openly into high-density workloads and edge AI inference. Brookfield spent the last two years building that footprint, starting with the combination of Evoque and Cyxtera under the Centersquare brand and then adding 10 more data centers in October 2025. The pitch is easy to understand.
The public-market math is harder. In the final prospectus, Csquare says revenue for the three months ended March 31, 2026 rose 16% from a year earlier to about $270.5 million. The same filing shows interest expense of $88.4 million, cash of roughly $67.6 million, total liabilities of about $6.3 billion and an accumulated deficit of roughly $1.3 billion. IPOGrid reads that as a financing story as much as a growth story. Investors were being asked to buy into digital-infrastructure demand while underwriting a balance sheet that still needed repair.
The company was explicit about that repair job. A pricing free-writing prospectus says Csquare expects to use $921.0 million of net proceeds to repay in full borrowings under its revolving credit facility, promissory note and Series 2024-1 variable funding notes, with the remainder going to other debt, fees and general corporate purposes. That is sensible capital allocation, but it also narrows the immediate upside for new shareholders. The reviewer’s concern is straightforward: when most of a billion-dollar IPO is earmarked for deleveraging, equity holders are funding cleanup first and expansion second.
Brookfield made the ask easier to place, but it did not make the governance question disappear. The same FWP says Brookfield-related buyers and other investors introduced by Brookfield agreed to buy $250 million of stock in the offering without an underwriting discount to the company, and that Brookfield and Brookfield Wealth Solutions affiliates will still own about 69.0% of the outstanding shares after the deal, or 65.8% if the shoe is fully exercised. In an earlier S-1/A, Csquare said it expects to qualify as a controlled company under NYSE rules and to rely on exemptions from certain governance requirements. Our interpretation is that the IPO came with meaningful sponsor support, but also with a plainly limited governance hand for new public holders.
The underwriting bench helps explain why the deal still cleared. Csquare’s pricing release names Morgan Stanley and TD Securities as representatives, with Wells Fargo, BofA, BMO and Scotiabank as joint lead book-running managers and another eight firms spread across the bookrunner and co-manager lines. That is a heavy syndicate for a first-time issuer, and it fits a company trying to place more than $1 billion of paper while still educating investors on an AI-linked, real-asset thesis.
What matters now is what the cut price signals. The company got public, rang the bell and kept its AI-infrastructure pitch intact. But the $21 print, below range, says buyers wanted compensation for leverage, sponsor control and the fact that this equity raise doubles as a debt refinancing. IPOGrid would frame Csquare as a useful read-through for the rest of the infrastructure calendar: scaled AI-adjacent issuers can still get public, but the market is forcing more discipline on terms, ownership economics and balance-sheet cleanup than private sellers may prefer.