Subscribe
DBTech45 Newsletter Desk

A-Portfolio

One stock a day that deserves a deeper look. Research, never a buy signal.

From the DBTech45 desk, weekday morningsBeta 2 · 2026-07-20The Vault · 17 filed
SPY778
+0.69% today
QQQ733
+1.32% today
IWM304
+0.32% today

A-Portfolio Daily, July 20, 2026

Csquare (NYSE: CSQR), the assets are real; the balance sheet still owns the story

Bias: Watch, not buy. The operating momentum is credible, but leverage and sponsor governance keep it below A-Portfolio quality.

Risk: Debt, interest expense, retrofit capital, and Brookfield control can absorb the upside before common shareholders see it.

Action: No position. Put CSQR on the verification list and wait for audited first-half results plus the first post-IPO quarter.

Why this matters, in plain English

Csquare rents secure space, power, cooling, and network connections to companies that need somewhere reliable to run their computers. Demand is rising because cloud systems and AI need more power and better cooling, and Csquare's new customer commitments are growing faster than its sales. If those commitments become paying contracts without another debt binge, the company can produce more profit from buildings it already owns. Derek should care because the operating business may be better than the weak IPO reception suggests, but the debt must be beaten first.

Why this name now

Csquare completed its NYSE IPO on July 17 after selling 50 million primary shares at $21, below the marketed $23-$27 range. The stock closed July 17 at $20.40, or 2.9% below the offer price. That is not momentum; it is a rejected deal looking for proof.

The proof is not imaginary. Preliminary first-half 2026 revenue is $546.7-$552.2 million, up 14%-16% year over year. Management's non-GAAP adjusted EBITDA estimate is $221.1-$231.2 million, up 19%-25%. Bookings are estimated at $126.3-$128.9 million, versus $93.1 million a year ago; the midpoint increase is 37.1%. Csquare also says 86% of 2025 revenue came from recurring enterprise colocation and interconnection services, no customer exceeded 7% of revenue, and its average remaining contract term was roughly 33 months at March 31.

The knife is the capital structure. At June 30, debt and finance leases were approximately $5.40 billion. The IPO delivered $1.01 billion of net proceeds, almost all earmarked for debt repayment and fees. Even after applying the full net proceeds and subtracting disclosed cash plus restricted cash, pro forma net debt is roughly $4.06 billion. First-half interest expense is expected to be $181.2 million, up 70.7% year over year. GAAP net loss is preliminarily estimated at $111.7-$122.7 million.

Numbers that matter

  • Price: $20.40 at the July 17 close; 2.81 million shares traded.
  • Post-IPO shares: approximately 155.0 million, excluding the underwriters' option.
  • Implied equity value: approximately $3.16 billion at $20.40.
  • Pro forma net debt: approximately $4.06 billion, assuming all $1.01 billion of net IPO proceeds reduce debt and fees do not require incremental cash.
  • Implied enterprise value: approximately $7.22 billion.
  • First-half 2026 preliminary revenue: $546.7-$552.2 million; midpoint growth 15.1%.
  • First-half 2026 preliminary adjusted EBITDA: $221.1-$231.2 million. Annualizing the midpoint gives $452.3 million, putting the stock near 16.0x enterprise value / annualized adjusted EBITDA. This is a rough run-rate, not guidance.
  • First-half 2026 preliminary FFO: $51.3-$62.4 million. FFO is management's non-GAAP funds-from-operations measure, not free cash flow.
  • 2025 GAAP results: $987.0 million revenue, $100.4 million operating income, $119.9 million net loss, and $172.0 million operating cash flow.
  • Capacity: 389 MW of sellable power at March 31, with management identifying 670 MW of potential expansion. Management estimates brownfield expansion at $4-$8 million per MW.
  • AI mix: management says approximately 15% of monthly recurring revenue in Q1 was AI/HPC-related and 80% of its 10 largest 2025 deals were AI/HPC-related. These are company-defined metrics, not audited segment revenue.

A-Portfolio score: 62/100, PASS FOR NOW

  • Category leadership / moat: 14/20, 64 sites, 21 metros, 1,700-plus customers, sticky contracts, and 36,600-plus interconnection products create real switching costs. Equinix and Digital Realty remain stronger ecosystems.
  • Growth inflection: 16/20, bookings growth is running ahead of revenue and adjusted EBITDA is expanding faster than sales.
  • Financial quality / cash durability: 7/20, recurring revenue is strong; debt, rising interest expense, falling near-term operating cash flow, and GAAP losses are not.
  • Valuation asymmetry: 7/15, below-range pricing helps, but roughly 16x annualized adjusted EBITDA is not distressed when leverage is this high.
  • Management / execution: 7/10, the Cyxtera/Evoque integration and booking acceleration deserve credit; the 2025 owner distribution and leverage deserve scrutiny.
  • Catalysts: 8/10, audited first-half results, debt repayment, booking conversion, and high-density retrofits can change the verdict.
  • Entry quality: 3/5, the deal broke issue price, but two trading days are not price discovery.

Bull / base / bear

Bull: Bookings convert quickly, churn stays low, adjusted EBITDA compounds above management's medium-term 14% opportunity, and debt falls without new equity. Existing powered buildings let Csquare add capacity faster and cheaper than greenfield competitors. The equity becomes a credible enterprise-infrastructure compounder rather than a sponsor refinancing.

Base: Revenue grows in the mid-teens and adjusted EBITDA improves, but interest, maintenance spending, and selective retrofits consume most of the cash. The company survives and delevers slowly; the stock remains a levered data-center landlord rather than an A-Portfolio leader.

Bear: Legacy facilities require more cooling and electrical work than promoted, first-half cash flow lands at the weak end, and interest stays punitive. Brookfield retains control while public shareholders absorb capital needs. The cheap-looking IPO becomes a debt stub with an AI label.

What the market may be missing

The market may be treating CSQR as another speculative AI infrastructure listing. It is different: customers bring their own servers and GPUs, while Csquare sells the scarce physical layer, space, power, cooling, and connectivity. That avoids the rapid GPU-obsolescence risk carried by neoclouds.

But the market may also be correctly discounting a sponsor-led refinancing. Brookfield-affiliated entities own approximately 103.9 million shares and retain about 69% of voting power. The prospectus also gives Brookfield broad board, information, consent, and registration rights. Public shareholders do not control this company.

The decisive metric is not headline AI demand. It is whether bookings growth converts into recurring revenue and cash faster than interest and retrofit spending consume it.

Kill criteria

Drop CSQR from A-Portfolio consideration if any of these hit:

  • Bookings fail to convert into sustained double-digit organic revenue growth over the next two reported quarters.
  • Pro forma leverage fails to trend toward management's long-term 6x-7x target.
  • Operating cash flow remains structurally below interest expense plus required capital spending.
  • Net revenue churn rises materially above the current low-single-digit range.
  • AI/HPC deployments require broad, expensive retrofits that invalidate the $4-$8 million-per-MW brownfield economics.
  • Brookfield monetizes aggressively after lockups or directs capital toward related-party priorities at minority shareholders' expense.

Entry discipline

No position. CSQR does not qualify for A-Portfolio sizing at 62/100. Revisit only after:

1. audited first-half results confirm the preliminary ranges;

2. the balance sheet reflects the IPO debt repayment;

3. one full post-IPO quarter shows booking-to-revenue conversion; and

4. management provides a credible bridge from adjusted EBITDA to cash available after interest and capital spending.

A broken IPO can get cheaper. Do not confuse a 2.9% discount to issue price with a margin of safety.

Sources and as-of date

As of: July 20, 2026. Market price is the July 17 close.

  • SEC S-1/A, operating data, preliminary first-half results, debt, ownership and risk factors: https://www.sec.gov/Archives/edgar/data/2105398/000110465926080556/tm264837-13_s1a.htm
  • SEC S-1/A full submission text: https://www.sec.gov/Archives/edgar/data/2105398/000110465926080556/0001104659-26-080556.txt
  • SEC 8-K confirming IPO close and $1.01 billion net proceeds: https://www.sec.gov/Archives/edgar/data/2105398/000110465926084616/tm264837d22_8k.htm
  • SEC stockholders agreement detailing Brookfield governance rights: https://www.sec.gov/Archives/edgar/data/2105398/000110465926084616/tm264837d22_ex10-2.htm
  • Csquare official IPO pricing release: https://investor.csquare.com/2026-07-16-Csquare,-Inc-Announces-Pricing-of-Initial-Public-Offering
  • Csquare investor overview: https://investor.csquare.com/overview
  • Morningstar/PitchBook on weak reception, cash flow, debt, and retrofit needs: https://www.morningstar.com/stocks/csquare-ipo-gets-weak-market-reception
  • Data Center Dynamics on pricing and first-day reception: https://www.datacenterdynamics.com/en/news/csquare-goes-public-but-falls-short-of-share-price-and-valuation-target/
  • July 17 market data: https://exa.ai/library/markets/stock/CSQR?date=2026-07-17

Calculations: market capitalization = $20.40 × 155.020702 million shares. Pro forma net debt = $5.3997 billion debt and finance leases − $1.0100 billion net IPO proceeds − $0.3303 billion cash and restricted cash. Enterprise value = equity value + pro forma net debt. All calculations were performed with decimal arithmetic. Preliminary first-half figures are management estimates and have not been audited or reviewed by Deloitte.

Research only, not personalized financial advice. Derek makes the allocation decision.

Signal and Noise →86'd →