What's Happening?
Rackspace has announced two significant multi-year customer agreements for compute capacity, bolstering its position in the AI infrastructure market. The first agreement is with Together AI, a neocloud startup, for which Rackspace will provide and operate
dedicated Nvidia GB300 NVL72 infrastructure. This deployment is scheduled for 2027 and includes associated power, liquid cooling, networking, and storage. Additionally, Rackspace and Together AI have signed a non-binding letter of intent to explore a joint go-to-market strategy for enterprise-focused AI solutions, with Together AI offering inference capabilities and Rackspace providing the underlying AI infrastructure. The second agreement is with Blacksmith, an infrastructure platform for code building and testing. Rackspace will supply Blacksmith with dedicated compute powered by AMD Epyc CPUs and storage, totaling approximately 2MW of capacity in Northern Virginia, with deployment commencing in early 2027. These agreements highlight Rackspace's commitment to expanding its AI infrastructure offerings and leveraging its operational expertise in complex enterprise environments.
Why It's Important?
These agreements are crucial for Rackspace as they solidify its role as a key provider of specialized AI infrastructure, a rapidly growing sector. By securing contracts with innovative AI companies like Together AI and Blacksmith, Rackspace is positioning itself at the forefront of the enterprise AI market. The partnership with Together AI, particularly the exploration of a joint go-to-market approach, could create a powerful synergy, offering comprehensive AI solutions to enterprises. This move allows Rackspace to capitalize on the increasing demand for robust and scalable compute resources required for AI development and deployment. For Together AI and Blacksmith, these agreements ensure access to high-performance, reliable infrastructure, which is essential for scaling their operations and delivering on customer expectations. The investment in Nvidia GB300 NVL72 systems and AMD Epyc CPUs also demonstrates Rackspace's commitment to utilizing cutting-edge technology to meet the demanding requirements of AI workloads, potentially attracting more clients in the future.
What's Next?
Rackspace plans to deploy the Nvidia GB300 NVL72 infrastructure for Together AI by 2027, and the AMD Epyc CPU-powered compute for Blacksmith will begin deployment in early 2027. The company is actively seeking separate financing for the purchase of the Nvidia GB300 NVL72 systems, indicating ongoing financial discussions with potential lenders. The non-binding letter of intent with Together AI suggests future collaboration on enterprise AI solutions, which could lead to new product offerings and market expansion. Rackspace also recently secured a $66 million private placement of common stock from an investor group led by Voss Capital, LP, which will be used for general corporate purposes, including funding GPU-related initiatives. This capital injection will support Rackspace's growth and its ability to invest in the necessary hardware and infrastructure to fulfill these new contracts and pursue further opportunities in the AI sector.
Beyond the Headlines
These developments underscore a broader trend in the technology industry: the increasing specialization and demand for dedicated AI infrastructure. As AI models become more complex and data-intensive, the need for purpose-built, high-performance computing environments, often with advanced cooling solutions like liquid cooling, is growing. Rackspace's strategy to provide and operate such specialized infrastructure reflects a shift from generic cloud services to more tailored solutions for AI workloads. This could lead to a more fragmented but highly specialized infrastructure market, where providers differentiate themselves not just by scale but by their ability to handle specific, demanding computational tasks. The emphasis on operational expertise, as highlighted by Rackspace's CEO, suggests that managing these complex AI environments will be a critical differentiator, moving beyond mere hardware provision to comprehensive service delivery. This trend also points to the continued dominance of specialized hardware like Nvidia GPUs and AMD CPUs in the AI compute landscape.
