QumulusAI Q3 Earnings Call Highlights

Key Points

• Revenue more than doubled sequentially to $14.8 million, up 376% year over year, as newly activated AI computing capacity began generating customer billings. QumulusAI posted its first positive adjusted EBITDA quarter at $3.7 million, despite a $14.4 million net loss.
• Direct-contract remaining performance obligations climbed to $515 million from $173.1 million in June, and the company signed $352 million in new contracts during the quarter. QumulusAI expects $29.1 million of backlog to be recognized in Q4, with most of the remainder scheduled for 2027–2030.
• Data-center capacity remains the main expansion constraint. The company had 8 megawatts active and billing, plus 5.45 megawatts secured but not yet billing, and reaffirmed year-end targets of 18 megawatts contracted capacity and $300 million in forward annual recurring revenue.

QumulusAI NASDAQ: QMLS reported third-quarter revenue of $14.8 million, more than double its second-quarter result, as newly energized computing capacity began billing customers. The company also posted its first quarter of positive adjusted EBITDA while recording a net loss.

CEO Mike Maniscalco said customer demand for AI computing remains strong, but access to data center capacity is limiting the pace of expansion. QumulusAI ended the quarter with 8 megawatts of capacity active and billing across five colocation sites and another 5.45 megawatts secured but not yet billing.

Revenue Growth and Contract Backlog

Third-quarter revenue rose 376% from a year earlier and increased from $6.7 million in the second quarter. CFO Scott Krosnowski said much of the capacity energized during the quarter billed for only part of the period. Compute accounted for 96% of revenue, compared with 47% a year earlier.

Adjusted EBITDA was $3.7 million, or 25% of revenue, compared with an adjusted EBITDA loss of $0.8 million in the second quarter. The company reported an operating loss of $41.6 million and a net loss of $14.4 million, or $0.43 per share.

The operating loss included a $26.4 million non-cash charge for shares issued as a professional services fee related to the company’s public listing. Net loss also reflected $32.3 million of non-cash gains from fair-value changes in convertible notes, a related option and warrants, Krosnowski said.

Remaining performance obligations from direct customer compute contracts reached $515 million at quarter-end, up from $173.1 million at June 30. Krosnowski said $29.1 million was scheduled to be recognized in the fourth quarter, with the remaining $486 million scheduled from 2027 through 2030. Maniscalco said the schedule points to another sequential doubling of quarterly revenue.

QumulusAI signed $352 million of new direct contracts in the third quarter, bringing 2026 signings to $529 million, Maniscalco said. Its latest three-year contracts are drawing approximately $19 million to $21 million in annualized revenue per megawatt, up from the $18 million to $20 million range the company reported in August.

Capacity Expansion and Financing

The 5.45 megawatts of secured capacity not yet billing includes 3.75 megawatts in metro Atlanta, expected to be ready for service in the second quarter of 2027; 0.5 megawatts in Columbia, South Carolina, expected in December; and 1.2 megawatts in Denver, expected in the first quarter of 2027.

During the question-and-answer session, Krosnowski said the Denver capacity was being marketed and had not yet been sold to a customer. Maniscalco said he believed all of the newly secured capacity except Columbia was being marketed; the Columbia capacity had already been sold and announced.

QumulusAI also has a right of first offer on another 7 megawatts at an Atlanta site. Maniscalco said the data center owner’s planned expansion was earmarked for 2027, but the timing was at the owner’s discretion. Separately, the company has a long-term lease at an Oklahoma site with 10 megawatts of active power running its legacy mining operation and up to 7 megawatts of additional grid power available.

Maniscalco said QumulusAI uses customer deposits and equipment financing tied to individual contracts to fund deployments. Major multiyear contracts typically require prepayments of 15% to 20% of total contract value, according to Krosnowski.

Third-quarter operating cash flow was positive $32 million. Capital expenditures were approximately $100 million, primarily for GPU servers, while the company added approximately $94 million of equipment and facilities under new leases. Debt and finance lease obligations totaled $248 million at September 30, and cash and restricted cash totaled $31.1 million, including $9.2 million restricted under loan agreements.

Outlook and Supply Constraints

QumulusAI reaffirmed its year-end targets of 18 megawatts of contracted capacity and $300 million in forward annual recurring revenue. Krosnowski said the company had 13.45 megawatts active or secured under agreements, with the balance in late-stage negotiations. About $209 million of forward ARR was under signed customer contracts, he said.

Asked about financing for two B300 agreements signed in September, Krosnowski said one order was fully paid for and the other had a 20% prepayment, with the balance in process. Maniscalco said the ready-for-service date on the latter had slipped slightly and the company was aligning final financing with the timing of data center capacity coming online.

Maniscalco said he had not seen a major shift in demand during the past month and that existing customers appeared more urgent about securing capacity. He also said some smaller colocation operators had become less willing to sell to customers without investment-grade credit, prompting QumulusAI to consider higher upfront prepayments and credit instruments as potential ways to secure supply.

About QumulusAI (NASDAQ:QMLS)

QumulusAI is a cloud infrastructure company specializing in rapid deployment of graphics processing unit (“GPU”)-powered solutions for artificial intelligence (“AI”) applications, serving a critical market that is often overlooked by large-scale cloud providers (“hyperscalers”), which operate massive, standardized computing infrastructures primarily serving the largest enterprises. Our platform delivers flexible, competitively priced, and customizable solutions for underserved small and mid-market customers—including machine learning teams, AI infrastructure startups, and research institutions—while also supporting the scale and complexity requirements of large enterprises, such as long-term deployments or supplemental on-demand compute capacity.

This instant news alert was generated by narrative science technology and financial data from MarketBeat in order to provide readers with the fastest reporting and unbiased coverage. Please send any questions or comments about this story to contact@marketbeat.com.