The Artificial Intelligence (AI) market continues to heat up, with tech juggernauts and emergent startups alike constantly innovating to accommodate new enterprise demands. AMD is one company that is privy to evolving market trends. At its Advancing AI 2025 event in San Jose, California, AMD unveiled new chips, system designs, and partnerships aimed at grabbing more of the AI compute market. Beyond the hardware and software upgrades, a recurring theme at the event was the expanding client list of neocloud providers. While neoclouds have enjoyed recent traction in the AI space, several roadblocks lie ahead before long-term success is a sure thing.
AMD Prepares for the Next Era of Cloud AI
AMD introduced the Instinct MI350 Series GPUs, set the stage for its future MI400 line, and expanded its ROCm 7 software stack. These announcements align with the company’s vision for supporting AI developers with faster, more efficient compute power—something in high demand as enterprises race to build and train large-scale AI models.
But what stood out to our analysts was who is leveraging the technology: neocloud providers. While large cloud providers like Oracle Cloud Infrastructure (OCI) will continue to be important customers for AMD and other chipset manufacturers, the growing relevance of neoclouds expands the overall market opportunity.
Unlike hyperscalers (e.g., Amazon Web Services (AWS), Google Cloud, and Azure), neocloud providers are smaller, more specialized cloud platforms. Their key differentiators include competitive pricing, AI centricity, compliance with regulations, and local points-of-presence. For organizations seeking better pricing models and requiring niche features that hyperscalers don’t always support, neoclouds make a compelling alternative.
At Advancing AI, neocloud providers Vultr and Crusoe announced strategic partnerships with AMD. Vultr will be one of the first to champion AMD’s new MI355X Graphics Processing Units (GPUs) later this year, adding to its expanding AI cloud stack. Meanwhile, Crusoe is also investing US$400 million in MI355X units to power a new liquid-cooled AI data center. AMD’s latest GPUs are also being leveraged by AI server vendors/Original Equipment Manufacturers (OEMs). Aivres, another San Jose-based firm, is purchasing the MI350 Series to improve server designs for High Performance Computing (HPC).
Can AMD’s Open Ecosystem Approach Power the Next Generation of Cloud Providers?
The neocloud momentum is real and growing fast, presenting potentially lucrative opportunities for chipset providers. While the recent prominence of neocloud providers is undeniable, their long-term relevance to the cloud industry remains uncertain. Neoclouds face several structural challenges, including uncertain ROI, smaller go-to-market budgets, and a more limited ecosystem footprint compared to hyperscalers.
To date, neoclouds have thrived by filling critical market gaps left by larger cloud providers, particularly in areas requiring high-speed GPU availability, transparent pricing, and flexible deployment models. However, unless these providers evolve into foundational pillars of the broader cloud infrastructure landscape, they risk becoming transient players used by enterprises for low-cost training before workloads are migrated back to more integrated Virtual Private Cloud (VPC) environments. This dynamic could limit their long-term strategic relevance.
AMD’s Advancing AI event in June 2025 underscored just how seriously the company is leaning into the neocloud opportunity. Through a sweeping showcase of its latest AI portfolio, AMD demonstrated a clear commitment to enabling neocloud providers with scalable, open infrastructure, positioning itself as a key enabler of a more distributed and diversified AI compute landscape.
Still, this strategy comes with inherent risk: it depends, at least in part, on the long-term viability and growth of neocloud providers. AMD’s open ecosystem approach (centered around hardware interoperability and choice) is a clear strength here, offering neoclouds the flexibility they need to innovate and differentiate. Together with a strong and reliable upgrade path of its software stack, AMD’s open architecture approach can be an impactful value proposition—particularly as neocloud providers break up the traditional roles and diversify the established AI ecosystem considerably.