Dell AI Server Sales Soar While Storage Growth Waits for Project Lightning

Dell’s AI Server Sales Surge While Storage Lags Behind

Dell Technologies recently reported a striking increase in its datacentre sales, largely fueled by new server and networking products aimed at artificial intelligence (AI) workloads. However, storage revenues have not kept pace, highlighting a clear divide in infrastructure demand.

The company’s second-quarter results reveal a 69% year-on-year rise in server and networking equipment sales, reaching $16.8 billion. This outstrips its personal computing division, which generated $12.5 billion in revenue. Dell's early entry into the AI market, with servers equipped with the latest Nvidia GPUs and high-throughput networking gear, has driven this surge.

“In the last six months we have delivered $10bn worth of servers for AI,” said Jeff Clarke, Dell’s vice-chairman and COO. “That’s more than was attained in the whole previous year. Demand is strong and sales of the new AI hardware has totalled $20bn for the year.”

Overall, Dell posted record quarterly revenue of $29.8 billion, up 19% from the previous year. Infrastructure-related sales grew 44% to $16.8 billion, with servers and networking contributing $12.9 billion. In contrast, storage revenue declined by 3% to $3.9 billion, indicating that storage hardware is not yet benefiting from the AI spending boom. The client services segment saw modest growth of 1%, with enterprise PCs up 2% and consumer devices down 7%.

Storage Trails Behind Compute in AI Spending

The current AI trend favors compute power over storage capacity or performance. This pattern is echoed in the results of flash storage leaders. NetApp, a top flash storage provider, reported only 1% annual growth with $1.56 billion in quarterly revenue. Pure Storage showed 13% growth to $861 million, largely driven by a unique order of SSD DirectFlash Modules (DFM) for Meta’s hyperscale operations. HPE’s server revenue rose 16% to $4.9 billion, but its storage figures were not detailed.

Why Storage Has Yet to Catch Up

AI workloads require data to move rapidly between storage and compute. However, storage technologies currently lag behind the data throughput capabilities of AI servers. GPUs in these servers can handle data transfers at 400Gbps to 800Gbps, whereas storage systems typically offer throughput closer to 100Gbps.

It appears many enterprises prioritize upgrading their compute infrastructure first, possibly staging investments to address storage enhancements later. Meanwhile, storage vendors are focusing on parallel file system architectures designed specifically for AI needs.

Vendors like Vast Data, Hammerspace, and Weka offer massively parallel storage access. Dell’s response is Project Lightning, which comprises PowerScale (the rebranded Isilon scale-out NAS). However, Project Lightning remains in development with no confirmed release date. NetApp counters with its Ontap Data Platform for AI, while Pure Storage promotes FlashBlade//Exa as a high-performance option.

What This Means for IT and Development Professionals

• Expect continued strong demand in AI server and networking gear as enterprises focus on compute power.
• Storage upgrades may follow as AI workloads mature and require faster, parallel-access storage solutions.
• Keep an eye on emerging products like Dell’s Project Lightning and advances from other vendors addressing AI-optimized storage.

For professionals aiming to stay current with AI infrastructure trends and enhance skills in AI hardware and storage, exploring specialized training can be valuable. Resources such as Complete AI Training’s latest AI courses offer focused knowledge on these evolving technologies.