QumulusAI (NASDAQ: QMLS) reported second quarter 2026 revenue of $6.7 million, a 118% year-over-year increase from $3.1 million. The neocloud infrastructure provider posted a quarterly loss of $(0.72) per share.

This marks the company's first quarterly report since beginning trading on the Nasdaq Global Market on July 16, 2026. Management highlighted that contracted GPU capacity is now generating revenue, with direct customer relationships accounting for more than 96% of the recurring revenue base.

Financial Performance

Revenue growth was driven primarily by the compute power segment, which reached $5.6 million in the quarter, representing approximately 84% of total revenue compared to 43% in the same period last year. Gross profit increased 163% to $4.5 million.

The operating loss widened to $7.7 million from $2.2 million, reflecting a $5.8 million rise in depreciation and amortization tied to expanded HPC infrastructure. General and administrative costs also increased due to public company readiness and headcount growth.

What the Numbers Show

The net loss of $22.8 million was primarily driven by non-operational factors rather than core business performance. Specifically, the period included a $19.2 million non-cash loss on the issuance of convertible notes. In contrast, the prior-year period benefited from a $14.5 million non-cash gain on the remeasurement of an investment in The Cloud Minders. Excluding these items, Adjusted EBITDA loss narrowed slightly to $0.8 million from $0.3 million, indicating stable underlying operational cash generation despite higher investment in infrastructure.

Operational Highlights

The company grew its deployed GPU fleet by approximately 224%, increasing from 952 to 3,088 units during the quarter. It ended the period with 8 MW of HPC capacity under executed lease and colocation agreements.

Key operational developments include:

• Signed 21 new direct customer AI compute contracts with an aggregate expected take-or-pay value of $169.7 million.
• Secured a colocation agreement in metropolitan Atlanta for up to 3.75 MW, with rights to expand by up to 7 MW at the same site.
• Purchased 1,632 NVIDIA Blackwell B300 GPUs to meet accelerating demand.
• Became an NVIDIA Cloud Partner on July 17, 2026.

Balance Sheet and Cash Flow

Operating cash flow for the first half of 2026 stood at $22.3 million, supported by a $30.5 million rise in deferred revenue as customers committed and paid ahead of delivery. Total cash and restricted cash increased to approximately $39.9 million at June 30, 2026, from $11.7 million at the start of the period. Total assets more than doubled to $215.0 million from $91.7 million year-end 2025, largely due to increases in property and equipment and finance right-of-use assets.

QumulusAI (NASDAQ: QMLS) has secured a seven-year colocation agreement for up to 3.75 MW of data center capacity in metropolitan Atlanta, reinforcing its infrastructure build-out for AI computing workloads. The agreement includes a right of first offer for up to 7 MW of additional contiguous capacity, potentially expanding the site’s total footprint to 10.75 MW. The contracted capacity is expected to support up to 2,048 NVIDIA Blackwell B300-class GPUs, with delivery scheduled for Q4 2026, subject to the transition of the existing occupant.

Strategic Alignment With Customer Demand

The Atlanta acquisition addresses the supply-side constraints for demand QumulusAI has publicly aggregated since early June. The company has announced customer agreements valued at more than $246 million during this period. Key components of this pipeline include:

• A three-year inference agreement valued at more than $71 million, announced on July 28.
• A GPU-as-a-Service agreement with DRW, announced on August 11.

To meet this contracted demand, QumulusAI purchased 1,632 NVIDIA Blackwell B300 GPUs in July. The company established its corporate headquarters in Georgia Tech’s Tech Square in June, positioning Atlanta as its home market for these deployments.

What the Numbers Show

The relationship between QumulusAI’s hardware procurement and its infrastructure leasing highlights a tightly coupled deployment strategy. The company purchased 1,632 GPUs in July to service agreements totaling more than $246 million. Simultaneously, it secured physical capacity for up to 2,048 GPUs in Atlanta. This near-parity between procured hardware and reserved power capacity suggests that the Atlanta site is intended to absorb the majority of the immediate hardware inventory, minimizing idle asset risk while ensuring power availability matches compute supply.