Cousins Ran Out of Office Buildings Worth Buying
The Sun Belt REIT posted its best rent spreads in five years, then quietly switched from buyer to builder — committing $50 million of land to a Dallas tower whose anchor tenant it still won't name.
CUZ · Earnings Call · 2026-10-30
An inflection the equity doesn't believe
Cousins Properties has spent the better part of three years telling investors that Sun Belt trophy office was heading into a supply cliff. In the third quarter it finally produced the receipts. The company leased 546,000 square feet, including 220,000 of new and expansion deals, and completed 326,000 square feet of renewals at a pace 25% above its three-year run rate. Second-generation cash rents rolled up 19.8%. “Obviously, 19.8% is a high number. I think it's the highest in 5 years” — Richard Hickson, Executive Vice President of Operations · 2026-10-30, Richard Hickson conceded — before noting that all but one market with second-gen activity grew rents at a double-digit clip. That is 50 consecutive quarters of positive roll-ups, spanning twelve years. The income statement followed. Same-property cash NOI rose 7.4% year over year — 8.6% excluding the under-construction 201 North Tryon — the strongest quarter since the first three months of 2020. Occupancy reached 89.5% against 93% leased, and management reiterated a 90% occupancy target by year-end. The late-stage pipeline, which stood at one million square feet in the September investor deck, had grown to roughly 1.4 million within a month. And only 853,000 square feet expires through the end of 2027 — half the average of the prior two years, with nothing over 100,000 square feet. So why is the stock up just 1% over 90 days, 8.5% off its July peak, and still 30.9% below its 2020 high? Because the operating recovery has been visible for a while. What changed this quarter was something else.From buyer to builder
On the fourth-quarter 2025 call, Colin Connolly was still shopping. “Given the number of opportunities that we're looking at across our footprint, that does make me hopeful by year-end, we will have identified one” — Michael Connolly, President and Chief Executive Officer · 2026-02-06, he said of a prospective ground-up development. It took longer than a year-end, but it has now happened: a 2.3-acre site at 2401 McKinney in Uptown Dallas, acquired through a joint venture with Trammell Crow in which Cousins holds 97.5%. Initial outlay was roughly $50 million — $42.2 million for land, $7.8 million of predevelopment. The plan is a trophy tower of approximately 700,000 square feet, with a construction start contemplated for the first half of 2027. On per-square-foot economics, Connolly was explicit: “overall construction costs in the Dallas Uptown market on a per square foot basis, depending on what you build range somewhere between $900 to $1,000 a foot” — Michael Connolly, President and Chief Executive Officer · 2026-10-30. Seven hundred thousand feet at that range implies a $630–700 million project for a company with a $4.9 billion market cap. This is a strategy shift disguised as a continuation. Kennedy Hicks framed the acquisition market bluntly: core office assets still trade around a 7% cap rate. Cousins wants 150 to 200 basis points above that — an 8.5% to 9% development yield. A year ago Connolly described the alternative: “we've been buying at GAAP cap rates in the 8s” — Michael Connolly, President and Chief Executive Officer · 2026-02-06. Those purchases are gone. Build-to-suit is the new conduit to that return, at least on paper. The underwriting bar set previously was unambiguous:Connect the dots and the real question becomes whether that pre-leasing bar is met before the first shovel moves.I would say we're targeting plus or minus 50% on a pre-lease basis. And I would expect cap rates -- or excuse me, development yields to be at least 150 basis points, if not 200 basis points higher than stabilized cap rates today. So that would put you in the, call it, 8.5% to 9% range today on a development yield.
The funding stack quietly changed too
For several quarters, the answer to "how will you pay for this?" was dispositions. That has been retired. “The recent increase in interest rates and limited equity in the office sector will likely constrain overall transaction volume in the near term. We don't need to sell assets and will remain selective on dispositions” — Jane Hicks, Executive Vice President and Chief Investment Officer · 2026-10-30, Hicks said. Tellingly, noncore asset is among the company's sharpest keyword decliners this quarter — a phrase that was load-bearing in the February script. What replaced it is more elegant: an ATM program that has already sold 3.9 million forward shares at an average of $30.66, none settled. “We've got, give or take, $120 million of unsettled ATM shares that we've issued. So that could be a good starting point” — Gregg D. Adzema, Executive Vice President and Chief Financial Officer · 2026-10-30, CFO Gregg Adzema said, adding that settlement can be extended again. It may be the single most under-appreciated line in the call: a pre-funded equity option that costs nothing until exercised. The cost of that flexibility shows up elsewhere. Interest coverage has compressed from a peak of 11.0x in 2016 to 3.9x. Liabilities have climbed from roughly 32% of assets in 2017 to 50.3% today. And the numerator is unforgiving: interest expense has gone from about 10% of operating cash flow in 2016 to roughly 30%. None of this is alarming for a REIT with investment-grade ratings and bonds trading at the tightest spread to Treasuries among traditional office peers. But it reframes the development pivot: this is not a company with infinite slack.The tenant nobody will name
Which brings us to the most interesting forty seconds of the call. Asked why the anchor tenant at 2401 McKinney hasn't been announced, Connolly demurred about finalizing building plans and costs, then offered this:Read that again: public statements plus incentives. That is the language of a state-and-local economic development package, i.e. a large-format corporate user. Cousins has not signed a lease; it has procured the land, the partner, the design work, and — apparently — a live negotiation with a party whose counterparty has already gone public with its intentions. The macro claim supporting all of this is that AI is a demand tailwind, not a headwind, for office. “to date, AI is proving to be more of a friend than a foe to the office sector” — Michael Connolly, President and Chief Executive Officer · 2026-10-30, Connolly said, arguing that while AI-native firms cluster in San Francisco and New York, the hyperscalers distributing headcount are landing in Austin, Nashville and Atlanta. Austin activity this quarter was 77% technology, with hyperscaler demand the largest component. It is a coherent story — and one that runs against a tape where AI data center equities have de-rated sharply over the last 90 days.I think a lot of the media speculation regarding a particular group has been focused on public statements that they've made as well as incentives that they have received. And I think that should show some pretty significant commitment.