The server market is experiencing record growth amid the voracious demand for production-ready AI infrastructure.
The market has more than doubled in size since 2020, with revenues of $235.7 billion in 2024, according to a new IDC report. More than half of that revenue came from servers with embedded graphics processing units (GPUs), a segment that grew by nearly 193% year-over-year in Q4 2024.
“The wide availability of GPUs in the market made server spending explode to a level that we have never seen before,” said Lidice Fernandez, IDC group VP for worldwide enterprise infrastructure trackers. “This is the highest level of spending we’ve seen in the market in the last 15 years.”
Non-x86 server revenues grew more than 250%
The server market reached a record $77.3 billion in revenue in Q4 2024, a year-over-year increase of 91% that was the second-highest growth rate since 2019, according to IDC.
Not surprisingly, Nvidia continues to dominate the server GPU space, comprising more than 90% of total shipments with embedded GPUs in Q4 2024. Dell Technologies and Supermicro had top market share in the overall server space, followed by Hewlett Packard Enterprise, IEIT Systems, Lenovo, and ODM Direct.
While x86 servers used to account for the majority of market spend, they are increasingly being outpaced by non-x86 servers. Revenue generated from x86 servers increased by roughly 60% in Q4 2024, to $54.8 billion, while revenues from non-x86 servers increased 262% year-over-year to $22.5 billion.
The non-x86 growth is largely being fueled by high demand for Nvidia Blackwell chips, Fernandez explained, as well as moves by the Chinese government to support local manufacturers such as Huawei, which is now one of the main suppliers of ARM processors for the Chinese market.
“We’re thinking the ARM pipeline will continue growing, because the appetite for adopting Blackwell chips that are ARM-based and have GPUs embedded in them is really high,” said Fernandez.
Longer-lasting infrastructure
GPUs that could be embedded in servers became more broadly available in 2024, contributing to the market growth. Most of that inventory has so far been consumed by cloud service providers (CSPs) and large hyperscalers, which means availability will open up for other enterprises in 2025, Fernandez noted.
“We can probably expect 2025 to be a year of strong investment,” she said, adding that access to supply is “getting much better.”
At the beginning of 2024, the wait time for GPUs was a lengthy 36 weeks. But, she pointed out, “the largest buyers in the world already have their infrastructure, so there’s going to be more availability.”
However, the industry will come to a point where it begins to digest all the infrastructure that’s been acquired, so investment will likely begin to slow in 2026, she said. Enterprises are beginning to move into different phases of AI deployment that don’t require as many resources. Initial model training, for instance, requires a good deal of infrastructure investment up front, but that infrastructure can be repurposed for later inferencing and fine-tuning steps.
Smaller enterprises with fewer resources are also likely to turn to pre-packaged cloud tools rather than on prem deployments, further driving down demand for servers.
“We’re not going to see the crazy double-digit growth that we’ve been seeing in the past two years,” said Fernandez.
Innovation is also extending refresh cycles: Today’s advanced processors and coprocessors are probably good for 4 to 5 years, whereas in the past, the refresh cycle was around 3 years, Fernandez said.
“Companies could have an investment that will last longer and allow them to transform infrastructure into something that over the years gains efficiency and allows them to move from one environment to another,” she said.
Economy could slow spending, drive up prices
Economic concerns, of course, are always looming — particularly now with hefty tariffs being imposed and proposed by the Trump administration.
GDP projections for most regions of the world have been lowered, Fernandez pointed out, and if there is a full-on tariff war, that will certainly slow spending in the Americas, the European Union, and some Asian countries. The resulting increased prices will not be absorbed by manufacturers; they will be passed along to the end users.
“That results in a decline in server shipments, meaning fewer boxes, but probably with more robust configurations, or the same configurations as last year but just a little more expensive,” Fernandez predicted.
Source:: Network World