Cloudian Inc

Why does storage decide the ROI of your AI? For a long time, storage was treated as a pure infrastructure question. Today it is clear: storage is the foundation on which AI actually delivers value. Without a solid data base, the potential of your AI stays untapped. This is exactly where Cloudian comes in as the data foundation for AI. Sovereign S3 object storage for AI, with cyber resilience and governance built in. Three reasons this matters for EMEA right now: * Sovereign S3 object storage: AI runs where your data lives, on-prem and under your control * Cyber resilience: immutable data protects against ransomware and data loss * Governance: data sovereignty and readiness for the EU AI Act, whose high-risk requirements take effect on 2 August The AI Act is getting closer, and the question is no longer whether your data foundation is ready, but when. We help you lay the foundation: sovereign, resilient, and governed. Feel free to reach out if you want to set up the data foundation for your AI sovereign and secure. #SovereignStorage #CyberResilience #EUAIAct

Jahrelang ging die Bewegung nur in eine Richtung: alles rauf in die globale Public Cloud. Diese Woche gibt es einen Namen für die Gegenbewegung. Gartner hat “Geopatriation” zum Top Strategic Technology Trend 2026 erklärt: das gezielte Zurückholen von Daten und Anwendungen aus globalen Clouds in lokale oder souveräne Umgebungen. Die Zahl dahinter ist bemerkenswert. Bis 2030 sollen mehr als 75% der Unternehmen in Europa und dem Nahen Osten ihre virtuellen Workloads geopatriieren. 2025 waren es noch unter 5%. Das ist keine graduelle Verschiebung, das ist eine Umkehr. 61% der westeuropäischen CIOs sagen bereits, dass sie aus geopolitischen Gründen stärker auf lokale Anbieter setzen wollen. Der Druck ist real und messbar. Laut Kyndryl, befragt wurden 3.700 IT-Verantwortliche in 21 Ländern, sagen 83%, dass Souveränitäts- und Repatriierungs-Regulierung im letzten Jahr an Bedeutung gewonnen hat. Gleichzeitig bleibt die alte Lücke bestehen: 98% nennen digitale Souveränität laut SUSE strategisch, aber erst 52% tun tatsächlich etwas dafür. Interessant ist, dass diesmal Regulierung und Wirtschaftlichkeit in dieselbe Richtung zeigen. Deloitte hat gerechnet: On-Premises-AI spart über drei Jahre 50% und mehr gegenüber Cloud-API, sobald das Token-Volumen eine bestimmte Schwelle überschreitet. Wer produktiv AI betreibt, kommt also nicht nur aus Compliance-Gründen nach Hause, sondern auch, weil sich die Rechnung dreht. Bleibt die eigentliche Frage: Wenn die Daten zurückkommen, wo landen sie? Geopatriation entscheidet sich nicht auf der Vertragsebene, sondern an der Storage-Ebene. Dort, wo die Daten physisch liegen und wer sie kontrolliert. Eine S3-kompatible Object-Storage-Grundlage im eigenen Rechenzentrum ist der Ort, an dem zurückgeholte Workloads tragfähig ankommen: die Compliance-Grenze verläuft an Deiner eigenen Infrastruktur, und die AI arbeitet dort, wo die Daten schon liegen. Genau darauf ist Cloudian HyperStore ausgelegt: eine souveräne, S3-kompatible Object-Storage-Grundlage on-prem, auf der AI direkt auf Deinen Daten läuft. Kein Kontrollverlust, keine fremde Hoheit über Deine Daten. Ein Rat aus der Praxis: Wenn Geopatriation bei Euch auf der Agenda steht, klär früh, auf welcher Storage-Ebene die Daten ankommen sollen. Der Umzug nach Hause ist nur so souverän wie das Fundament, auf dem er landet. Plant Ihr, Workloads zurückzuholen, und wenn ja, was ist bei Euch der Auslöser: Regulierung, Kosten oder Kontrolle? Ich bin sehr gespannt auf Deine Sicht. → HyperStore Object Storage: https://lnkd.in/gnnFfZzK → Data Sovereignty mit Cloudian: https://lnkd.in/e56dYiQM → HyperScale AIDP: https://lnkd.in/dQ-JMvYS #Geopatriation #SovereignAI #DataSovereignty

For many years the movement of data went one way: into the public cloud. This week the counter-movement got a name. Gartner has named “geopatriation” a Top Strategic Technology Trend for 2026: deliberately bringing data and applications back out of global clouds into local or sovereign environments. The number behind it is striking. By 2030, more than 75% of enterprises in Europe and the Middle East are set to geopatriate their virtual workloads. In 2025 it was under 5%. That is not a gradual shift, it is a reversal. Already 61% of Western European CIOs say geopolitics will push them to rely more on local cloud providers. The pressure is real and measurable. According to Kyndryl, which surveyed 3,700 IT leaders across 21 countries, 83% say sovereignty and repatriation regulation has grown more important over the past year. At the same time the old gap persists: 98% call digital sovereignty strategic, according to SUSE, yet only 52% are actually doing something about it. What makes this moment different is that regulation and economics point the same way. Deloitte has run the numbers: on-premises AI saves 50% or more over three years compared with cloud API, once token volume crosses a certain threshold. So enterprises running AI in production are coming home not only for compliance, but because the math flips. Which leaves the real question: when the data comes back, where does it land? Geopatriation is not settled at the contract level, it is settled at the storage layer. Where the data physically sits and who controls it. An S3-compatible object storage foundation in your own data centre is where repatriated workloads land on solid ground: the compliance boundary runs along your own infrastructure, and the AI works where the data already lives. That is exactly what Cloudian HyperStore is built for: a sovereign, S3-compatible object storage foundation on-prem, where AI runs directly on your data. No loss of control, no foreign hand on your data. One practical note: if geopatriation is on your agenda, decide early which storage layer the data should land on. The move home is only as sovereign as the foundation it lands on. Are you planning to bring workloads back, and if so, what is the trigger for you: regulation, cost, or control? → HyperStore object storage: https://lnkd.in/gnnFfZzK → Data sovereignty with Cloudian: https://lnkd.in/e56dYiQM → HyperScale AIDP: https://lnkd.in/dQ-JMvYS #Geopatriation #SovereignAI #DataSovereignty

Sovereignty doesn't start with new projects. Most data already lives somewhere, often in the public cloud. The practical step is repatriation. Data repatriation means bringing workloads out of the public cloud back onto storage you own, or that sits with a local provider. With S3-compatible object storage the APIs stay the same, your applications keep running, but control, location and cost move back to you. Three reasons this is gaining momentum in 2026: sovereignty is becoming mandatory, cloud bills keep rising, and AI needs the data where the compute is. Repatriation addresses all three at once. See how data repatriation works with Cloudian: https://lnkd.in/dcwm3-U8 What is holding you back from bringing your data home? #DataRepatriation #DataSovereignty #ObjectStorage

Everyone talks about the value of AI, far fewer about its real cost. And that cost only shows up in operations. The biggest budget drivers are rarely the models. Our Enterprise AI Infrastructure Survey shows where they actually sit. ➡️ https://lnkd.in/eR764fjv #AI #AICost #DataInfrastructure

🏎️ What a day at the Goodwood Festival of Speed! On Friday, we had the pleasure of hosting our customers and channel partners at one of the most iconic events on the motorsport calendar. Between the roar of engines on the hillclimb and the incredible display of automotive history and innovation, it was the perfect backdrop for great conversations about where storage and AI infrastructure is headed next. A huge thank you to everyone who joined us! It's moments like these, away from the usual conference rooms, that remind us how much we value the relationships we've built across our partner and customer community. Here's to more collaboration, more innovation, and (hopefully) a few more high-speed thrills together. Graham Martin David Axler Rosa Lenders

Less than 1 month till the EU AI Act becomes more concrete. The real test isn't the model, it's the data path: governance, logging, lineage. Find out more in Cloudian's new report. ➡️ https://lnkd.in/exWM7TEk #EUAIAct #DataSovereignty

✨ AI success starts with the right data foundation. When 🇨🇭 Swiss IT provider Begasoft set out to build a sovereign AI data lake 🌊 for its Brandbot platform, it needed more than scalable storage. Begasoft needed an S3-native foundation that could keep sensitive customer data entirely within Switzerland while supporting enterprise RAG, multi-tenant AI services, and future growth. By choosing Cloudian HyperStore, Begasoft built an AI-ready platform that delivers: ✅ Swiss data sovereignty for regulated industries ✅ Native S3 compatibility for seamless AI integration ✅ Multi-tenant isolation for secure customer environments ✅ Linear scalability from PoCs to enterprise production The result? A single AI data lake powering both AI/RAG workloads and large-scale data integration, enabling Begasoft to move from pilot projects to production deployments with confidence. Read the full customer story to see how Begasoft is helping Swiss organizations embrace AI without compromising security, compliance, or data sovereignty. https://bit.ly/3SvrDVM #AI #EnterpriseAI #DataInfrastructure #ObjectStorage #DataLake #DataSovereignty #SovereignAI #RAG #AIStorage #AIInfrastructure #ITProvider #DataManagement #ObjectStorage #HybridCloud #DigitalTransformation #Cloudian

How fast does your data actually reach the GPU? That is the question deciding, in 2026, whether your most expensive resource works or waits. Most AI projects do not stall on compute. They stall on the data path. The GPU is there, the model is there, but storage cannot feed it fast enough. And every minute a GPU waits for data is expensive. Three things I am taking from the market this week: 1. The bottleneck has moved. GPUs are not the scarce part anymore, the path from disk into GPU memory is. That is where performance is decided today. 2. Direct access changes the math. With NVIDIA RDMA for S3 we deliver over 200 GB/s of sustained throughput straight into GPU memory, three times faster than flash without RDMA, at 45% lower CPU load on the GPU server. And it scales linearly with every node you add. 3. The effect compounds. Indexing into a Milvus vector database, together with NVIDIA we went from two hours to 16 minutes, eight times faster. That is the difference between a pilot and a production setup that keeps up. My takeaway: a bigger GPU cluster does not help if the data cannot keep up. The answer is object storage that talks directly to the GPU. Fast, efficient, under your control. How do you make sure the GPUs in your AI projects are never waiting for data? I would love to hear your view. #AI #ObjectStorage #GPU #Cloudian #DataSovereignty

Lights, camera, AI. 🎥 We recently joined our partners at Supermicro in California to film an upcoming panel for the #OpenStorageSummit, where Cloudian Inc's VP of Worldwide Solutions Architects, Peter Sjoberg, shared insights on managing unstructured data for AI. As enterprises build AI-ready infrastructure, unstructured data has become one of their most valuable assets. Peter joined industry experts to discuss how object storage, converged file and object architectures, and data preparation are helping organizations unlock the full potential of AI. Stay tuned -the session will be available later this summer. Peter Sjoberg Molly Presley Sherry Lin Wendell Wenjen #AI #DataInfrastructure #ObjectStorage #UnstructuredData #Cloudian #Supermicro #OpenStorageSummit