According to Cloudian CEO Michael Tso, the artificial intelligence (AI) market is “frothy” rather than in a bubble, emphasizing that “there’s a mountain underneath this thing. ” Most investments in AI to date have focused on training large language models (LLMs). Training a foundation model costs around $10 billion, he noted, and this is an ongoing process. Nevertheless, the real profit lies in inference. “There are probably just 25 companies in the world that can afford the cost of training, but inference fits within most organisations’ budgets, ” Tso explained. He added that applying LLMs to an organisation’s own data has been a complex challenge because foundation models lack specific domain knowledge. To address this, organisations need an easy-to-deploy solution that integrates both generic and proprietary data. Cloudian’s new AI data platform aims to fill this gap. Early versions have already been delivered to select customers, and according to Tso, “people are understanding it very quickly. ” By embedding AI directly into the storage platform, the models can analyze and comprehend the stored data. For instance, Cloudian uploaded its own user manual to the system, which then correctly answered queries such as, “How do I set up a firewall?” Tso highlighted that this method also preserves existing access controls on the data. Cloudian’s collaboration with Nvidia is “very natural, ” he said.

While Nvidia has open-sourced some models, it does not want to sell to or directly support enterprise customers who wish to use them. Cloudian claims to be the first to market with this on-premises integration concept. The platform offers a solution for organisations unwilling to run AI workloads on public clouds. Developers of large AI models typically keep tight control over their training data sources, so “people have to take care of their data, or there’s no limit to where it will go. ” “No company wants to volunteer their data, ” Tso stated, adding that many enterprises hesitate to invest in cloud-based AI due to fears of exposing proprietary information. Some companies, like Samsung, have even blocked employee access to all public AI chatbots. In the Asia-Pacific region, Cloudian partners with sovereign AI provider ResetData to deliver AI infrastructure as a service with physically isolated servers, preparing for anticipated growth in Australian AI spending in 2026. “The key to our approach is that it is your data, ” stored and processed in an environment that ensures sovereignty and security, ” said James Wright, Cloudian’s managing director for Asia-Pacific and Japan, describing it as a compelling offer for potential customers. Tso believes that security and sovereignty concerns will likely drive renewed datacentre activity, whether in private facilities or spaces offered by major datacentre co-location providers. Despite significant enterprise interest in AI, Tso suggested the technology is awaiting its “Lotus 1-2-3 moment” — a breakthrough application that firmly establishes its value. “Once you have a use case like that, it kills everything, ” he remarked. He added that this moment could occur when all of an organisation’s data is securely accessible to an AI—a capability Cloudian’s platform claims to enable.