The global server market is expected to grow 12.8 percent annually this year, with artificial intelligence (AI) servers projected to account for 16.5 percent, driven by continued investment in AI infrastructure by major cloud service providers (CSPs), market researcher TrendForce Corp (集邦科技) said yesterday.
Global AI server shipments this year are expected to increase 28 percent year-on-year to more than 2.7 million units, driven by sustained demand from CSPs and government sovereign cloud projects, TrendForce analyst Frank Kung (龔明德) told the Taipei Times.
Demand for GPU-based AI servers, including Nvidia Corp’s GB and Vera Rubin rack systems, is expected to remain high, while CSPs would invest in in-house application-specific integrated circuit (ASIC) development, which is also expected to drive demand for AI training and inference, Kung said.
Photo: Bloomberg
Capital expenditures by the five largest CSPs — Google, Amazon Web Services, Meta Platforms Inc, Microsoft Corp and Oracle Corp — are expected to rise 40 percent year-on-year this year, TrendForce said in a separate report released on Tuesday last week.
Google and Microsoft are expected to lead the expansion of general-purpose server procurement to handle the surge in daily inference workloads generated by services such as Microsoft’s Copilot and Google’s Gemini, it said.
In addition to large-scale infrastructure expansion, part of the spending is expected to be earmarked for replacement of general-purpose servers purchased during the 2019 to 2021 cloud investment boom, the researcher said.
However, the share of ASIC-based AI servers is expected to rise to 27.8 percent as Google and Meta ramp up in-house chip development, with shipments of ASIC-based systems forecast to grow faster than those of GPU-based servers, TrendForce said.
Google, in particular, is investing more heavily in its own ASICs than most other CSPs and is emerging as a key market player, it said.
Its tensor processing units, which support Google Cloud Platform services, are also increasingly being sold to external customers such as AI start-up Anthropic PBC, it said.
The growth in the global server market this year would be driven by continued momentum in AI servers, with the rapid expansion of AI inference applications in particular fueling rising demand for storage and edge AI servers, Kung said.
Meanwhile, as demand for AI inference continues to grow rapidly across end devices and industrial applications, it is expected to reshape the computing power structure, which is still heavily skewed toward training workloads, he said.
Inference-based AI servers are expected to account for 44 percent of total AI server shipments this year, with the share projected to rise above 50 percent by 2029, he said.
As CSPs and enterprises expand AI applications across endpoints, the market is expected to develop along two main tracks, Kung said.
The first involves broader deployment of GPU-based systems, such as Nvidia’s platforms, as well as in-house ASIC rack solutions to support large language model training and generative AI inference, he said.
The other track is the growing adoption of more distributed, edge-oriented AI server architectures — such as MGX platforms — designed to support real-time inference workloads closer to data sources, he added.
The EU and US are nearing an agreement to coordinate on producing and securing critical minerals, part of a push to break reliance on Chinese supplies. The potential deal would create incentives, such as minimum prices, that could advantage non-Chinese suppliers, according to a draft of an “action plan” seen by Bloomberg. The EU and US would also cooperate on standards, investments and joint projects, as well as coordinate on any supply disruptions by countries like China. The two sides are additionally seeking other “like-minded partners” to join a multicountry accord to help create these new critical mineral supply chains, which feed into
Elon Musk’s lieutenants have reached out to chip industry suppliers, including Applied Materials Inc, Tokyo Electron Ltd and Lam Research Corp, for his envisioned Terafab, early steps in an audacious and likely arduous attempt to break into the production of cutting-edge chips. Staff working for the joint venture between Tesla Inc and Space Exploration Technologies Corp (SpaceX) have sought price quotes and delivery times for an array of chipmaking gear, people familiar with the matter said. In past weeks, they’ve contacted makers of photomasks, substrates, etchers, depositors, cleaning devices, testers and other tools, according to the people, who asked not to
Japan approved ￥631.5 billion (US$3.97 billion) in additional subsidies to hasten Rapidus Corp’s entry into the high-stakes artificial intelligence (AI) chipmaking arena, ramping up support for a project widely regarded as a long shot. The capital is intended to bankroll Rapidus’ work for information technology firm Fujitsu Ltd, one of the initial customers that Tokyo hopes would get the signature endeavor off the ground. The new money raises the fees and investments that the government is injecting into the start-up to ￥2.6 trillion by the end of the current fiscal year to March next year, the Japanese Ministry of Economy, Trade and
The founder of Chinese property giant Evergrande Group (恆大集團) has pleaded guilty to charges of fraud and bribery, a court said yesterday, the latest blow for what was once the country’s leading developer. Evergrande’s rise was propelled by decades of rapid urbanization and rising living standards, but in 2020, its access to credit dramatically narrowed when the government introduced curbs on excessive borrowing and speculation. The company defaulted in 2021 after struggling to repay creditors. Founder Xu Jiayin (許家印), 67, known as Hui Ka Yan in Cantonese, was reportedly held by police in 2023, with Evergrande saying he had been subjected to
RSS