Neocloud infrastructure operators gathered at the RAISE Summit to discuss market consolidation, funding, and competitive positioning.

Last week I attended the RAISE Summit in Paris, where 9,000 participants gathered to discuss enterprise AI, coding tools, and the evolving neocloud sector. The event underscored the booming AI infrastructure market, with content heavily focused on AI labs and neoclouds—sectors at the epicenter of AI innovation and risk. Several key themes emerged throughout the conference.

**Data Sovereignty and Model Agnosticism**

A dominant theme was enterprises' rising demand for data and AI sovereignty. Organizations are increasingly seeking model-agnostic architectures and adopting open-source AI models to avoid vendor lock-in and protect proprietary data, signaling a broad shift toward custom, secure deployments. French President Emmanuel Macron urged European AI independence and data sovereignty, while investor Mark Cuban emphasized AI's role in accelerating innovation and fueling a startup boom.

Cuban, speaking with Lovable Cofounder Anton Osika, described how the explosion of AI tools is enabling people to unleash creativity and translate ideas into code faster. "Your biggest limitation is your time and your imagination," he said. "The number of startups is accelerating dramatically. As things change, there is more opportunity. France is becoming more entrepreneurial."

**Enterprise Challenges and Model Strategy**

A panel titled "AI Pilot to Production: The Valley of Death and How to Cross It" examined real-world implementation challenges, including token costs and data protections. Barak Kaufman, Chief Strategy Officer of Wonderful, identified return on investment as the chief challenge for enterprises. "For many enterprise organizations, they are starting to calculate the ROI. It has to come from revenue as well as cost savings," he said. "Tokenmaxxing is definitely the wrong approach. You need to figure out in which situations you want to have an open-source model, but for certain use cases. The biggest execution gap is the change management."

Kaufman noted that boardroom discussions differ markedly from vendor hype. A key trend is avoiding lock-in to high-cost models and experimenting with open-source alternatives. "What I'm seeing in closed doors is how often the CxOs are looking at how to have model-agnostic architectures. The answer is not to go all-in on one model. There's not a single enterprise that I've met anywhere that's not thinking about how to avoid model lock-in," he said.

Andrew Feldman, Cofounder and CEO of AI chip maker Cerebras, challenged the industry's focus on token counting. "I don't think we should count tokens," he said. "We should measure work. We're doing work. Tokens are a very rough measure."

Jonathan Corbin, Founder and CEO of Maven AGI, highlighted another trend: enterprises protecting proprietary data and moving away from freely sharing it with large frontier LLMs. "We are seeing more and more people moving to an on-prem deployment. I don't want people accessing data without controlling that," he said. This reflects findings in the Enterprise AI Index, which shows most enterprises are seeking proprietary AI implementations.

**Neoclouds as Mini Hyperscalers**

Neoclouds are emerging as future "mini hyperscalers," expanding beyond GPU rentals to offer global infrastructure, application services, and enhanced data privacy. The category is diverse, ranging from sophisticated altscalers such as Vultr and Akamai to newly converted datacenter operations from cryptocurrency assets, such as IREN and Nebius.

Kevin Cochrane, Chief Marketing Officer of Vultr, framed AI infrastructure buildout as a long-term cycle. "Are we overhyped? We are just at the start of the journey," he said. "The journey that we are on is a supercycle. It's a 30-year era of innovation through the entire compute stack. We are at the dawn of a new compute cycle. We are just getting to the new age where enterprises can deploy AI-native applications."

However, even in a 30-year cycle, risks remain. While the Internet has achieved compounded growth of over 40% per year for three decades, it has experienced rocky periods. Widely reported challenges in GPU cloud reliability mean operational excellence will be a key differentiator. Initially, neoclouds competed on availability or cost—hyperscalers charged substantially more for services like S3 storage and GPU rental, if available at all. But as demand for data sovereignty, privacy, and on-prem integration increases, enterprises will expect more sophisticated services: hybrid data portability and regional data sovereignty zones.

Suresh Vasudevan, CEO of Clockwork Systems, emphasized the steep quality gradient among neoclouds. "The ones at the top of the pyramid, they are making investments in infrastructure and making it robust," he said. Top neoclouds are focusing heavily on observability. "You need to measure everything," Vasudevan added. Clockwork recently introduced a guarantee for its TorchPass GPU fault tolerance software, offering credits to customers who don't achieve promised results.

Roman Chernin, Cofounder and Chief Business Officer of Nebius, stressed that neocloud sophistication must rise as customers grapple with token costs and agent management. "We control the infrastructure and go down stack and up stack. You want to deploy the agent that works. You don't want to be an expert on which token is better. What you want is the next layer of abstraction, the platform will run the agent and optimize it," he said.

**Infrastructure Orchestration and Automation**

Neoclouds are building GPU and cloud compute capacity rapidly, but lack the software resources and expertise of hyperscalers like AWS and Microsoft. The summit attracted numerous infrastructure orchestration vendors such as Rafay, emma, and Netris; data management companies including Backblaze, DDN, and Bright Data; and enterprise AI tools like Cognition and Lovable. Beyond GPU failure, challenges include automating services at scale, partitioning resources for multi-tenancy, and ensuring security and data connectivity.

Orchestration software is becoming critical to the neocloud ecosystem. Netris, which automates GPU cluster network buildout and enforces hard multi-tenancy in hardware, has largely become the industry standard, labeling its approach NAAM—Network Automation, Abstraction, and Multi-Tenancy. Rafay provides GPU configuration and automation middleware, partnering with Netris. Mohan Atreya, Chief Product Officer of Rafay, described the trajectory: "The pattern is the customers want regular compute and AI in the neoclouds. They will evolve into mini hyperscalers."

Emma, a Luxembourg-based cloud orchestration platform, provides infrastructure automation from physical to application layers with integrated agentic AI operations, which it calls "Infrastructure-as-Prompt." According to Emma Founder Dmitry Panenkov, "Neoclouds have to evolve from real estate to value add."

The RAISE Summit demonstrated the scale and sophistication of the AI infrastructure market as it rapidly matures toward supporting enterprise production workloads.