OVHcloud acquires Gladia to build sovereign voice AI stack
OVHcloud, Europe's largest cloud provider by server count, has completed its acquisition of Gladia, a Paris-founded voice AI startup, paying for the deal entirely in newly issued shares rather than cash. The all-equity structure, valued at roughly €28.3m at closing prices, reflects OVHcloud's ambition to internalise the technical building blocks of voice AI rather than simply resell third-party models, placing the French group in more direct competition with the hyperscalers that dominate AI-services revenue globally.
Gladia, incorporated in 2022, built a speech-to-text (STT) platform delivered through a single API that transcribes audio in real time and in batch across more than 100 languages. The company says it now serves more than 300,000 developers and 2,000 enterprise customers. OVHcloud says it will fold Gladia's technology into OVHcloud and OVHai, its managed AI services layer, and strengthen what it calls its AI Lab, the internal research unit aimed at developing next-generation agentic and multimodal AI.
Why voice AI sits at the centre of the agentic shift
The timing is not incidental. Across the AI industry, the pivot from text-only large language models to multimodal, voice-enabled agents is accelerating. Contact-centre automation, real-time meeting intelligence, and autonomous customer-service pipelines all depend on accurate, low-latency STT as a foundational layer. By internalising Gladia's capabilities, OVHcloud can offer customers a vertically integrated agentic stack without routing audio data through US-controlled infrastructure, a significant differentiator in European regulated sectors such as banking, insurance, and public administration.
The deal structure reflects the complexity of the transaction: 1,807,186 new OVH ordinary shares were issued at €15.66 apiece, representing approximately 1.18% of post-deal share capital. Warrant mechanisms tied to Gladia's performance targets in 2026 and 2027 could add a further 3,361,404 shares, lifting theoretical maximum dilution to roughly 3.3%. Lock-up and orderly-disposal commitments from Gladia's founders cap near-term overhang risk. Naolys Audit, appointed as contribution appraiser by the Commercial Court of Lille-Métropole, found no qualification in its fairness assessment.
Digital sovereignty as a capital allocation thesis
For cross-sector investors, OVHcloud's Gladia deal is best read as a chapter in a larger sovereign-cloud story rather than a standalone M&A event. European regulatory pressure around data residency and AI model provenance is intensifying: the EU AI Act's obligations for high-risk AI systems, combined with DORA requirements in financial services and NIS2 in critical infrastructure, are creating structural demand for cloud providers that can certify the entire AI pipeline as European-domiciled. OVHcloud, which already operates more than 500,000 servers across 46 data centres on four continents, is positioning itself as the only provider that can match hyperscaler capability while guaranteeing data sovereignty end-to-end.
That positioning has implications beyond Europe. Gulf sovereign wealth programmes and Southeast Asian digital-economy strategies have each identified trusted, non-US AI infrastructure as a strategic gap. OVHcloud has historically targeted these markets through its global data centre footprint, and a vertically integrated voice AI capability now gives the group a more differentiated product to bring to sovereign-AI conversations in those geographies. For investors watching capital flows into the sovereign-cloud and AI-infrastructure intersection, OVHcloud's string of AI Lab acquisitions represents a distinct model: absorb specialist AI capabilities via equity rather than cash, avoid diluting liquidity, and compound technical depth rather than chasing revenue via model-resale margins. The next question is whether that model generates sufficient developer adoption velocity to close the gap on hyperscaler AI-services market share before the window of regulatory advantage narrows.