12 Jul

AI is changing server sales but paying off for enterprises

    AI is revolutionizing the sales of servers and proving to be a profitable investment for enterprises, according to two reports from Omdia. The reports indicate a decline in overall server shipments as businesses prioritize systems driven by graphics processing units (GPUs) in line with the adoption of artificial intelligence.

    Omdia’s upcoming cloud and data center market report predicts a decrease in server shipments for the first time since 2007. However, unlike the server drop in 2007 caused by a global economic crisis, the current shift in server purchasing trends has a more positive outlook.

    While the demand for compute resources remains high, there is a growing preference for more expensive servers with specialized hardware for AI model training, particularly GPUs, over the traditional enterprise servers equipped only with a central processing unit (CPU).

    During the first quarter of 2023, Omdia observed that 2.8 million servers were sold, significantly below the expected 3.2 million. This decline in shipments from the fourth quarter of 2022 to the first quarter of 2023 marks the largest on record, as reported by Omdia.

    The fear of missing out on the thriving AI market is driving both cloud service providers (CSPs) and enterprises to increase their investments in AI hardware. However, AI hardware comes with a considerably higher price tag compared to traditional server hardware. To mitigate the substantial costs associated with AI server investments, CSPs and enterprises are deferring the refreshment of their existing servers.

    Vladimir Galabov, head of the cloud and data center research practice at Omdia, highlighted the substantial investments in AI clusters, describing them as colossal projects. Each server can cost up to half a million dollars, which is equivalent to 50-60 general-purpose servers.

    There has been a trend in recent years to extend the deployment lifespan of hardware. Previously, servers were typically refreshed every three to four years, but now tier-one CSPs use servers for six years on average, while tier-two providers report lifespans of up to 10 years, according to Omdia.

    Several factors contribute to the slowdown in server hardware sales. Omdia points to macroeconomic headwinds and the increasing cost of and limited access to capital as drivers for investment tradeoffs. Additionally, most organizations have already optimized their operations with a hybrid capex/opex model, combining on-premises computing with cloud services.

    To meet immediate computing needs while carefully evaluating long-term demand, businesses are adopting the lower-risk strategy of utilizing cloud services. This approach leads to fewer on-premises hardware sales and greater adoption of cloud services.

    The quick returns on AI investments are evident, according to Omdia. Increased topline revenue, reduced bottom line costs, improved efficiency, and enhanced customer experiences are among the benefits. Although AI deployments are relatively new, they are already generating positive results. Approximately 54% of respondents reported measuring positive results of 1% or more, while 14% achieved a return on investment (ROI) of 11% or higher, depending on the category.

    Notably, Omdia conducted its survey of 369 enterprises in February 2023, well before any generative AI projects could have made an impact. Therefore, the research assesses early AI deployments.

    Omdia projects that as more companies witness the benefits of AI, the early ROI will likely accelerate AI innovation. Generative AI initiatives, in particular, are expected to gain more opportunities to demonstrate their value.

    As AI projects mature within organizations and their benefits become apparent, companies will expand AI deployments to other use cases and business areas. With tangible ROI justifying these programs, senior management will gain confidence in expanding AI initiatives, consequently driving increased spending.

    Share this