JLL’s Record Q3: Leasing Up 27%, Data Centers Named, AI Not

On October 28, 2026, Jones Lang LaSalle Incorporated furnished its third-quarter results on a Form 8-K (Item 2.02) and filed its Form 10-Q the same day. Revenue was $7,281.2 million, up 12 percent, against $6,510.4 million a year earlier. Adjusted EBITDA was $460.2 million, up 33 percent from $347.3 million, and diluted earnings per share was $6.85, up 49 percent in dollars. Leasing Advisory revenue reached $941.8 million, up 27 percent in local currency, and the company says U.S. growth of 39 percent was led by industrial, office and data center assets. Capital Markets Services revenue was $736.8 million, up 21 percent in local currency. Real Estate Management Services grew 9 percent, driven by Workplace Management and Project Management.

On capital, JLL says it intends to execute a new $500 million accelerated share repurchase imminently, bringing full-year repurchases to more than $1 billion, with $2.5 billion remaining authorized at September 30. Net debt was $858.7 million, a net leverage ratio of 0.5 times, and CEO Christian Ulbrich says JLL raised its full-year adjusted EPS target range. These are company statements in a company release, not PSV findings. The 10-Q attributes the U.S. leasing gain to a meaningful uptick in average deal size, and says U.S. office leasing revenue rose 18 percent against market volumes up 2 percent, while U.S. industrial revenue rose 56 percent against market volumes up 38 percent, comparisons that rely on JLL Research volume data.

The quarter matters to operators for what it says about deal size and where the large tenants are. JLL names data centers as one of three U.S. leasing drivers, in a quarter when its own leasing revenue growth outran the market it cites. It does not give a data center revenue figure, a megawatt count or a client count, so the size of the data center contribution cannot be read from the filings. For the AI infrastructure story, this is corroboration from the largest listed brokerage that compute demand is showing up in leasing fees, not a measurement of how much.

The filings are equally informative about what they leave out. A search of the earnings release and the 10-Q finds no use of the words artificial intelligence, generative AI or machine learning. JLL’s Software and Technology Solutions line, which sits inside Real Estate Management Services, reported $58.2 million of third-quarter revenue against $58.6 million a year earlier, down 1 percent, and $173.1 million for nine months, up 1 percent. JLL attributes margin expansion to higher revenue and incremental platform leverage, and it describes higher nine-month capital expenditures as primarily technology infrastructure and workspace optimization. Its Proptech Investments, which it reports outside its segments, carried at $369.0 million at September 30 against $353.0 million at year end. PSV reads that as a company still investing in technology while crediting scale, not AI, for the margin, and that is PSV analysis, not a JLL statement.

The tool is a short claims check run on any earnings release or investor deck that a CRE firm, vendor or advisor cites as proof that AI is working. Inputs are the release, the 10-Q or 10-K, and the claim as repeated by the person selling or presenting it. The output is a one-page table that marks each claim as stated in the filing, stated only in a deck or on a call, or absent, with the page where the filing says it. For this quarter the table would show that leasing growth, deal size, platform leverage and data center demand are stated, and that AI-driven margin is absent.

The reviewer is the analyst or principal who intends to rely on the claim, and the approval gate is simple: a claim that is not in the filing does not enter an investment memo, a vendor selection or a client pitch as a fact. Arithmetic, such as margin change or the growth in a line item, is computed from the filed figures and labeled as computed. PSV has not tested any JLL product, and nothing here is a statement about JLL’s tools or their performance.