Analysis: churn was fixed, expansion was not

Annual recurring revenue is the number that decides this company’s trajectory, and it moved from EUR 38.1 million to EUR 37.9 million over twelve months. That near flat line is the sum of two large and opposite forces. Cancellations dropped by more than three quarters in euro terms, from EUR 8.1 million to EUR 1.9 million, and eleven new logos were added. If nothing else had changed, ARR would have risen materially. It did not, which means the shortfall in upselling existing accounts was close in size to the improvement in churn.

That is a different problem from the one Exasol had a year ago, and arguably a harder one. Churn is addressed by service, contract terms and product stability, and the numbers say the company has done that work. Expansion revenue depends on customers deciding to put more workload on the platform, and the two reasons the company gives, hesitancy about new on-premises infrastructure and a partner rollout that ran late, are both outside its direct control. The partner delay is the more tractable of the two and the one a reader can track, because a rollout that slips can also arrive.

The guidance history matters here and is easy to miss. The outlook restated on 19 August is not the outlook the company carried into the summer. Exasol revised its guidance on 12 August 2026, one week earlier, and the August 19 release states the full year expectations in line with that revision: ARR to change by between minus 2 percent and plus 2 percent, revenue to decline in the upper-single-digit percentage range, and EBITDA between EUR 3.0 million and EUR 3.5 million. A confirmation of preliminary figures a week after a guidance revision is a much weaker signal than a confirmation of the guidance itself, and the two should not be read as the same event.

Set the EBITDA guidance against the half year. EBITDA of EUR 1.6 million in the first six months against a full year range of EUR 3.0 million to EUR 3.5 million implies a second half broadly similar to the first, with the upper end requiring some improvement. Exasol is not being asked to produce a step change in the remaining months, which is consistent with a cost base that has already absorbed the revenue decline. The ARR range of minus 2 percent to plus 2 percent is wide enough to accommodate the current trajectory without further deterioration, and at EUR 37.9 million the company is inside it.

What the disclosure does not establish is the composition of the eleven new customers. New logo count without contract value tells a reader that the funnel is working and nothing about whether the additions are large enough to replace the expansion revenue that did not arrive. The half-year report published alongside the release, rather than the release itself, is where that detail would sit.

What the documents say

Exasol AG (ETR: EXL) confirmed its preliminary first half figures and published its half-year report on 19 August 2026 at 07:30 CET/CEST. The Nuremberg company sells high-performance database technology for on-premises and hybrid environments, and its shares trade on the Regulated Unofficial Market in Frankfurt, in the Scale segment, and in Dusseldorf, Hamburg, Munich, Stuttgart and on Tradegate, under ISIN DE000A0LR9G9.

Annual recurring revenue stood at EUR 37.9 million at 30 June 2026, against EUR 38.1 million a year earlier. Recurring revenue for the six months rose to EUR 18.9 million from EUR 18.6 million. Total revenue fell to EUR 19.0 million from EUR 21.5 million, and EBITDA came in at EUR 1.6 million after EUR 2.0 million in the first half of 2025.

Where the revenue went

The revenue decline has a single stated cause. Non-recurring hardware and service revenue totalled EUR 0.1 million in the first half of 2026, against EUR 2.9 million in the prior-year period. That swing in one-off business is larger than the fall in total revenue, which means the recurring part of the business grew while the reported top line shrank.

Retention improved on both measures the company reports. The volume of contract cancellations and adjustments fell to EUR 1.9 million from EUR 8.1 million a year earlier. The rolling ARR churn rate declined to 10 percent at 30 June 2026 from 24 percent at the end of financial year 2025. Exasol added eleven new customers in the six months and attributed that to its focus on data sovereignty and support for agentic AI workloads.

Against those gains, the company said the expansion of existing customer relationships developed below expectations, citing reduced willingness to invest in new on-premises infrastructures and a delayed sales rollout with a strategic cooperation partner.

The sovereignty argument

Exasol frames itself around digital sovereignty and compliance with the EU General Data Protection Regulation, and positions its database for customers in highly regulated industries and the public sector who want to control data, execution and governance across the environments they choose. That framing has a concrete legal anchor. Regulation (EU) 2016/679 sets out the conditions under which personal data may be processed and transferred, and it is the compliance obligation that pushes some German and European buyers toward on-premises and hybrid deployment rather than public cloud.

The company’s own newsroom carries a run of material on the same theme, including a study it commissioned from BARC on European demand for technical and legal control over data. Demand described in a study is not demand booked as ARR, and the first half figures are the test of the connection between the two. So far the sovereignty argument shows up in the new customer line, where eleven additions were recorded, and not yet in the expansion line.

What comes next

The half-year report is available through the company’s investor relations site, and management, chief executive Jörg Tewes and chief financial officer Jan-Dirk Henrich, presented the figures in an English language webcast on 19 August 2026 at 2:00 p.m. CEST. Three items will settle the questions this release leaves open: whether the rolling churn rate holds near 10 percent through the second half rather than reflecting a favourable renewal calendar, whether the delayed partner rollout reaches customers before the year ends, and whether ARR finishes inside the guided band. The first two determine the third.