This is investment research, not personal financial advice.

OVH GROUPE (EURONEXT-PARIS:OVH) closed 13.6% lower at €13.91 on 26 August after chief financial officer Stéphanie Besnier said she would leave at month-end. The fall erased roughly one-sixth of the post-event equity value in one session. There was no profit warning, no downgrade to the FY2026 outlook and no disclosed balance-sheet surprise. The announcement named Deputy CFO Julien Rabette as interim finance chief and added two go-to-market leaders while the search for a permanent CFO continues (OVH 26 Aug 2026; Google Finance 2026).

The size of the reaction therefore rests on an inference. Besnier helped build the five-year plan and the FY2027 budget, and she departs just as OVHcloud is meant to prove that years of heavy infrastructure spending can produce positive cash after interest and leases. That makes the market's concern intelligible. It does not make every euro of the fall fundamental.

Five annual filings and the latest interim accounts point to a split verdict. OVHcloud has built a credible European cloud platform, lifted annual revenue from €663 million to €1.085 billion since FY2021 and restored adjusted EBITDA margin above 40%. Yet FY2025 cash from operations exceeded cash capital expenditure by only €50 million. After €59.7 million of interest and roughly €22 million of lease principal, the owner-cash bridge remained about €32 million negative. At €13.91, the shares already reflect some execution damage, but they still assume a large change in cash economics. The reaction looks too severe for the announcement alone and roughly proportionate only if the departure foreshadows slippage in that conversion.

The announcement carried no earnings warning

The 26 August release was unusually compact for an event that removed so much market value. Besnier, CFO for three years, will remain until 31 August. Rabette, the deputy CFO, takes the role on an interim basis. Founder, chair and chief executive Octave Klaba said the pair had spent the past year building the next five-year plan and the FY2027 budget; he also said refinancing was secured through 2030. Pierre Byramjee joined the executive committee as Chief Revenue Officer Digital Cloud, while Sylvain Siou became VP Customer Solutions & Success within the Corporate route to market (OVH 26 Aug 2026).

Independent coverage confirmed the departure and appointments but added no financial deterioration or regulatory event (Telecompaper 2026). This matters because the stock's €2.19 decline is not the present value of a disclosed earnings cut. It is the market charging for three softer risks: loss of continuity before a new strategy period, a founder-led concentration of authority, and doubt about the timetable for cash conversion.

The first risk is concrete. A CFO who helped frame the FY2027 budget leaves five days before that year begins, and there is no named permanent successor. The second is older. Klaba returned as chair and CEO in October 2025 after Benjamin Revcolevschi's tenure, bringing strategy and execution back under the founder. The third dominates valuation. OVHcloud reaffirmed in June that FY2026 should produce positive levered free cash flow, even after H1 reported a €14.2 million outflow. October's annual result now has to prove that second-half working capital and lower capital expenditure closed the gap (OVH H1 2026; OVH Q3 2026).

None of those facts proves that the budget is impaired. The release says the budget exists, the deputy finance chief supplies continuity and funding is in place. A 13.6% fall would be easier to call proportionate if it followed weaker revenue, a covenant issue or revised cash expectations. It did not. The event changes confidence in delivery more than it changes the disclosed plan.

One-sixth of the equity value prices a governance inference

The closing price implies a €2,134.61 million market capitalisation. Using 153.46 million issued shares produces the same result. Adding €1.126 billion of net debt excluding leases at February produces an enterprise value near €3.26 billion before lease liabilities. That is about 7.4 times FY2025 adjusted EBITDA and approximately 7.2 times a simple annualisation of H1 FY2026 adjusted EBITDA. These multiples look modest for a recurring cloud business, but adjusted EBITDA is not the scarce metric here. Cash is.

Before the announcement, the same issued-share count valued the equity near €2.47 billion. The one-day repricing was therefore equivalent to almost six times FY2025 company-defined unlevered free cash flow, or 6.7 times the simple operating-cash-flow less capex bridge. Put differently, the market treated the CFO change as if it removed several years of the current cash run-rate. That is a large penalty for a release without new financial numbers.

There is still a reason the equity is volatile. Net debt excluding leases is more than half the post-fall equity value. Small changes in enterprise value land on a thinner equity base, so a 10% enterprise-value adjustment can translate into a much larger percentage change for shareholders. The €350 million share tender completed at €9 in FY2025 increased this sensitivity. It reduced the share count at an attractive price relative to €13.91, but it also helped lift net debt excluding leases from €667 million to €1.103 billion in one year (OVH FY2024; OVH FY2025).

That capital decision creates the central asymmetry. If capex falls and owner cash rises, the tender concentrated future cash flows into fewer shares. If the conversion stalls, debt and leases get paid before owners. The CFO departure arrived at the point where that fork was due to narrow.

The business sells control, not hyperscale breadth

OVHcloud earns revenue from three infrastructure families. Private Cloud, which includes dedicated bare-metal servers and hosted private environments, supplied 60% of revenue in the first nine months of FY2026. Public Cloud contributed 22% in the third quarter, while Webcloud, including domains, hosting and collaboration products, supplied the balance. Most revenue is recurring because customers rent compute, storage, networking and managed environments rather than purchase hardware outright (OVH Q3 2026).

Its differentiation is not product breadth against Amazon Web Services, Microsoft Azure or Google Cloud. Those groups can spend more in a quarter than OVHcloud can over several years. OVHcloud instead combines European legal jurisdiction, data-location control, transparent bandwidth economics and its own industrial stack. It designs and assembles servers, develops water-cooling systems, operates 46 data centres and runs a 100 Tb/s network. For public bodies and regulated enterprises, that combination can matter more than access to the longest software catalogue.

The model has two reinforcement loops. The first is industrial: higher utilisation spreads data-centre and network costs, while in-house server design can improve power density and hardware economics. The second is trust-based: security certifications and sovereign references make the platform eligible for workloads that create further references. SecNumCloud annual recurring revenue reached €24 million in FY2025, up 63%, and a consortium including OVHcloud received the European Commission's SEAL-3 sovereignty classification in a six-year cloud framework (OVH FY2025; OVH Q3 2026).

But the loop is narrow. EU enterprise cloud adoption rose to 52.7% in 2025, so the addressable market is growing (Eurostat 2026). Local providers have not converted that growth into broad infrastructure share. The three US hyperscalers still account for roughly 70% of European cloud infrastructure revenue, while all European providers together account for about 15%. Sovereignty supplies an entry ticket to selected workloads, not immunity from price, product or service competition.

The mix also restrains group growth. Public Cloud grew 20.2% like for like in Q3 FY2026, but Private Cloud grew 4.0% and Webcloud 2.0%. With Private Cloud still the majority of revenue, strong growth in the smaller engine lifted total quarterly growth to 6.9%, not 20%. Group net revenue retention eased from 107% in FY2024 to 105% in FY2025, 104% in H1 FY2026 and 102% in Q3. Existing customers are still expanding, though by progressively less (OVH FY2024; OVH FY2025; OVH H1 2026; OVH Q3 2026).

Five years of growth bought with €1.87 billion of capex

The history shows real operating progress and equally real capital intensity. Figures below are reported in euros. EBITDA is OVHcloud's adjusted measure. Operating cash flow and cash capex come from the consolidated cash-flow statements; their difference is an author calculation, not company-reported free cash flow.

Fiscal period Revenue (€m) Adjusted EBITDA (€m) Margin Operating cash flow (€m) Cash capex (€m) OCF less capex (€m, computed) Net debt ex leases (€m)
FY2021 663.3 262.0 39.5% 268.2 343.2 -75.0 655.7
FY2022 788.0 308.1 39.1% 274.1 453.4 -179.3 525.3
FY2023 897.3 325.0 36.3% 329.8 357.8 -28.0 607.6
FY2024 993.1 381.5 38.4% 368.2 343.1 25.1 667.2
FY2025 1,084.6 437.8 40.4% 419.0 368.9 50.1 1,102.8
H1 FY2026 555.3 227.2 40.9% 271.1 238.5 32.6 1,125.5

Sources: OVH FY2021, OVH FY2022, OVH FY2023, OVH FY2024, OVH FY2025 and OVH H1 2026. Interim figures are six-month values and are not directly comparable with full years. FY2025 cash capex of €368.9 million differs from the company's €361.4 million capex performance measure because the latter uses its stated alternative-measure definition.

Revenue compounded while margin first compressed and then recovered. FY2022 and FY2023 absorbed the build-out: cash capex reached €453 million in FY2022 and adjusted EBITDA margin fell to 36.3% the next year. By FY2025, revenue was 64% above FY2021 and adjusted EBITDA was 67% higher. Margin reached 40.4%, above the IPO-period level (OVH FY2021; OVH FY2025).

The capex trend changed later than the earnings trend. Across FY2021 to FY2025, OVHcloud spent about €1.87 billion of cash capex to add €421 million of annual revenue. That is not a clean incremental-return calculation because much of the spending maintains existing servers, supports future capacity or replaces assets with short useful lives. It does show why EBITDA growth alone cannot establish economic compounding.

OVHcloud's own return on capital employed improved by 2.9 percentage points between FY2021 and FY2025, helped by margin recovery and lower capital intensity. The filing defines ROCE as adjusted EBITDA after depreciation, amortisation, impairment and tax, divided by prior-year capital employed. The charted improvement is evidence of better use of the asset base, but the disclosed level remains low relative to the 40% EBITDA margin because depreciation is substantial (OVH FY2025).

H1 FY2026 complicates the apparent improvement. Cash capex rose to €238.5 million, or 42.9% of revenue on the company's capex measure, as management brought purchases forward to secure components. Operating cash flow benefited from €54.7 million of working-capital inflow. That timing support is why the half-year OCF less capex bridge looks positive while reported levered free cash flow remained €14.2 million negative (OVH H1 2026).

A 40% EBITDA margin is not owner cash

The FY2025 owner-cash bridge starts with €419.0 million of operating cash flow. Subtract €368.9 million of cash purchases of property, equipment and intangible assets, and €50.1 million remains. OVHcloud reported €57.6 million of unlevered free cash flow using its own definitions. Both figures show improvement. Neither is cash available to equity after financing the full operating footprint.

The next claims matter. Cash interest paid was about €59.7 million in FY2025, up sharply as debt and rates increased. Lease principal was roughly €22.3 million. Subtracting both from the €50.1 million simple bridge leaves about negative €31.9 million. This is an author-computed owner-cash estimate; it is deliberately stricter than the company's unlevered measure and before any discretionary distribution (OVH FY2025).

The result does not mean the assets have no value. Growth capex can create future revenue, and treating every euro as maintenance understates cash economics during an expansion. The opposite treatment is also unsafe. Servers depreciate, customers expect current hardware and data centres need continual investment. A useful valuation therefore cannot assume either that all growth capex disappears or that current capex intensity persists forever.

The H1 FY2026 filing provides a near-term test. Adjusted EBITDA rose 5.9% to €227.2 million and margin reached 40.9%, yet unlevered free cash flow fell 12.2% to €32.3 million and levered free cash flow was negative €14.2 million. Management attributed the capex spike to front-loading and repeated its expectation of positive levered cash for the full year. Q3 revenue then accelerated to 6.9% like for like, led by 20.2% Public Cloud growth, and the company repeated that expectation again (OVH H1 2026; OVH Q3 2026).

This is why the CFO event landed so hard. The market had already been asked to trust timing: second-half capex normalisation, working-capital reversal and the start of a lower-capital-intensity plan. The person responsible for reporting and financing that bridge is leaving before the proof arrives.

The moat has procurement evidence and share limits

European sovereignty has moved from marketing language into procurement criteria. The Commission's framework assessed legal, operational and supply-chain control, security, openness and resilience. The consortium containing OVHcloud reached SEAL-3, above the minimum level. That creates a reference for defence, government and regulated-enterprise conversations that hyperscalers cannot replicate simply by cutting price (OVH Q3 2026).

OVHcloud also controls more of its physical stack than many regional providers. Its server design, water cooling and owned network can lower bandwidth and energy costs while letting it state where data and operations sit. Customers seeking predictable network bills or EU jurisdiction have a concrete reason to include it in tenders.

Financial evidence supports a stable, specialised advantage rather than a widening general moat. Corporate revenue exceeded €200 million in FY2025, Public Cloud passed €100 million and the group had almost 1,200 customers each generating more than €100,000 of annual recurring revenue. SecNumCloud revenue grew quickly from a small base. Against that, retention has weakened and Private Cloud growth has slowed. The EU Data Act will remove switching and egress charges from January 2027, which can help migrations into OVHcloud but also lowers contractual friction for migrations out.

IONOS is a useful European comparison, though not a pure one because web presence and productivity products make up much of its group. IONOS generated €701 million of H1 2026 revenue, €245 million of adjusted EBITDA and a 35.0% margin. Its FY2025 free cash flow was far larger than OVHcloud's despite a lower EBITDA margin. The comparison shows that OVHcloud's 40% margin is not, by itself, evidence of superior distributable economics (IONOS H1 2026).

Management quality must therefore be judged through capital allocation and delivery, not only product credibility. The founder's return brought a clearer strategic owner into the chief executive role. It also increased key-person and governance concentration. A permanent CFO with authority over the Step Ahead plan would reduce that tension; a prolonged interim period would leave it open.

The balance sheet turned cash conversion into the crux

Net debt excluding leases was €1.125 billion at February 2026, up from €667 million eighteen months earlier. Including leases, net debt was €1.269 billion. Management reported leverage of 2.6 times and said all debt was hedged at an average 4.4% rate, with refinancing secured through 2030 (OVH H1 2026; OVH 26 Aug 2026).

That is not an immediate solvency problem. The group has positive EBITDA, recurring revenue, funding headroom and no disclosed near-term maturity wall. The risk is opportunity cost and equity sensitivity. Interest expense consumes cash that could fund capacity, and leverage reduces room for another debt-funded capital return before owner cash is established.

The €350 million tender at €9 per share is the largest allocation decision in the five-year history. At the post-fall price it looks economically favourable on a per-share basis. At the same time, the transaction transferred balance-sheet capacity into equity concentration before the cash model had fully matured. Net debt ex leases increased by €436 million during FY2025. If cash conversion arrives, that was accretive. If it does not, a low purchase price cannot erase the financing burden.

The anti-thesis to the market's concern is strong enough to state plainly. Funding is secured, the interim CFO was already deputy CFO, the FY2027 budget has been prepared, Q3 growth accelerated and no outlook changed. A finance-chief transition need not disturb operations. The thesis against that comfort is equally specific: the CEO changed in FY2025, the CFO now leaves, retention is easing and the capital plan depends on a second-half cash swing that has not yet been reported. October can resolve much of the disagreement.

What €13.91 assumes

A conventional earnings multiple obscures too much depreciation and capex. The valuation below uses a ten-year enterprise DCF based on owner cash flow to the firm, then cross-checks it against forward EV to adjusted EBITDA. Starting revenue is an author estimate of €1.15 billion for FY2026. Net debt excluding leases is €1.125 billion, and the per-share calculation uses 153.46 million issued shares. Lease claims are considered in the scenario ranges rather than subtracted mechanically a second time.

At €13.91, the equity plus net debt implies enterprise value of about €3.26 billion. A reverse DCF using 8.75% cost of capital, 2% terminal growth, revenue growth fading from 6.5% to 3%, and an 8% starting owner-cash margin requires owner cash to reach roughly 17.3% of revenue in year ten. FY2025's strict owner-cash margin was negative 2.9%. The market is still pricing a large capex conversion after the fall.

The sensitivity is unforgiving. With a 14% terminal owner-cash margin, the model produces about €10.4 per share at an 8.75% discount rate. Raising the terminal margin to 18% lifts it to roughly €14.7. At a 9.5% discount rate, even a 16% terminal margin gives about €10.3. These are author estimates, not company forecasts, and small assumption changes have large equity effects because net debt ranks first.

The model's central range is wide by necessity. OVHcloud can plausibly sustain a 40% adjusted EBITDA margin while producing very different cash results depending on server cycles, capacity utilisation, growth capex and lease renewals. The post-event price is not obviously low on owner cash; it is low only if the FY2026 inflection develops into a durable reduction in capital intensity.

Four paths from the same data-centre estate

Scenario Operating assumptions DCF cross-check Forward EBITDA cross-check Value range per share
Severe downside Growth near 1%; owner cash stays below 4%; debt absorbs most value €0 €6.29 €0 to €7
Bear Growth begins near 4%; owner cash reaches 11% only gradually; 5.5x EBITDA €2.20 €10.24 €3 to €11
Base Growth begins near 6.5%; owner cash rises from 8% to 14%; 7.5x EBITDA €10.38 €18.11 €9 to €19
Bull Public Cloud sustains high-teens growth; owner cash reaches 18%; 9x EBITDA €28.27 €25.90 €24 to €30

The severe case is not a prediction of insolvency. It captures an equity structure in which weak growth and continued capital intensity leave enterprise value close to debt and lease claims. The bear case gives the platform credit for survival and some conversion, but not enough to support the current price on the DCF method.

The base case straddles €13.91. That is not a calibrated conclusion built around the screen price; the two methods diverge because EBITDA capitalisation gives far more credit to future capex reduction than near-term cash does. The gap itself is the finding. Investors are debating not the existence of EBITDA but how much of it will become owner cash.

The bull case requires two things together: Public Cloud continues to grow at a high-teens rate and total capital intensity declines despite AI and sovereign-cloud demand. Faster revenue without lower capital intensity is not enough. Lower capex accompanied by weaker product capacity is not enough either.

Measured against these paths, the 13.6% fall removed some execution premium but did not price a cash-stall outcome. The announcement contained too little new financial information to justify the whole move on a stand-alone basis. Yet the resulting price still depends on management delivering the very transition that Besnier's exit made less certain. That is why the reaction can be both optically severe and economically coherent.

October decides whether timing was the whole story

Three disclosures now carry more information than the resignation itself. First, FY2026 results should show whether levered free cash flow turned positive after the H1 outflow. A negative full-year result would indicate slippage, not merely front-loading. Second, the capex ratio should fall from H1's 42.9%; a full-year level above 35% would make the implied owner-cash ramp harder to defend. Third, the permanent CFO appointment and FY2027 outlook should show whether the prepared budget survived the transition.

Operating data matter alongside cash. Private Cloud growth below 3% or net revenue retention below 100% would show contraction in the large installed base before Public Cloud is big enough to carry the group. Net debt ex leases above three times adjusted EBITDA would narrow strategic flexibility. Conversely, positive levered cash, lower capex intensity and stable retention would make the management change look more like succession friction than a break in the plan.

The market's verdict on 26 August was harsher than the disclosed facts but consistent with OVHcloud's leveraged cash sensitivity. At €13.91, it is pricing neither a broken platform nor effortless conversion. It is pricing a functioning European cloud whose next finance chief must turn a 40% EBITDA margin into cash after servers, leases and interest. October will show whether the one-session shock anticipated a missed inflection or merely charged early for uncertainty.

Source notes and confidence

Verification is full for the fetched company filings, triggering company release, Euronext identity record and market snapshot used here. Valuation, strict owner cash and scenario outputs are author calculations reconstructed from filed inputs. The biggest missing item is the permanent CFO choice; the second is FY2026's audited split between recurring and growth capex. Those absences are the story, not gaps filled with estimates.

References

  • Euronext 2026, Euronext Paris product page for OVH GROUPE (OVH), 26 August 2026.

  • Google Finance 2026, OVH GROUPE market snapshot, 26 August 2026.

  • OVH 26 Aug 2026, OVHcloud announces changes to its organisation and Executive Committee, 26 August 2026.

  • OVH FY2021, OVH GROUPE consolidated financial statements for FY2021, 16 November 2021.

  • OVH FY2022, OVH GROUPE consolidated financial statements for FY2022, 26 October 2022.

  • OVH FY2023, OVH GROUPE consolidated financial statements for FY2023, 25 October 2023.

  • OVH FY2024, OVH GROUPE consolidated financial statements for FY2024, 24 October 2024.

  • OVH FY2025, OVH GROUPE consolidated financial statements for FY2025, 21 October 2025.

  • OVH H1 2026, OVH GROUPE interim consolidated financial statements for H1 FY2026, 9 April 2026.

  • OVH Q3 2026, OVHcloud Q3 FY2026 revenue release, 25 June 2026.

  • IONOS H1 2026, IONOS Group SE H1 2026 financial report, 6 August 2026.

  • Eurostat 2026, Cloud computing services used by enterprises dataset, updated 3 February 2026.

  • Telecompaper 2026, OVHcloud loses CFO and announces new appointments, 26 August 2026.