AI capital spending surge sweeps across the globe! Goldman Sachs reassesses AI computing power landscape, shouting that the AI server market is heading towards $1.24 trillion.

In the context of the almost endless growth in demand for artificial intelligence-related computing power infrastructure, financial giant Goldman Sachs currently predicts that by 2030, the market size of AI-related servers will reach $1.24 trillion, while the traditional server market size is also expected to exceed $160 billion. Goldman Sachs' latest optimistic outlook for the AI server market highlights that AI servers are no longer just a hardware enterprise product market, but are evolving into a "new infrastructure market" in the global digital economy era. Dell and Foxconn's recent strong performance announcements have actually become the most direct validation of the continuous large-scale explosion in demand for AI computing power infrastructure. In the latest quarter, Dell's AI server-related revenue increased by 757% year-on-year, with AI orders reaching $24.4 billion, and AI backlog exceeding $51.3 billion, while the revenue outlook for AI servers was further significantly increased. More importantly, Dell's AI server customer base has expanded from traditional cloud computing companies to large enterprise customers, Neocloud operators, and AI-native companies, indicating that the demand for AI computing power is accelerating its spread from a few technology giants to global enterprises. From the perspective of the industry chain, in addition to Dell, Foxconn is actually one of the core beneficiaries of this AI server super cycle; in fact, it is even a more underlying beneficiary. Dell is responsible for sales and system integration, while Foxconn is deeply involved in the manufacturing and assembly of NVIDIA's GB200, GB300, Blackwell, and future Rubin platform AI servers and AI cabinets. Previously, Foxconn has clearly stated that its AI server business has surpassed its smartphone business, becoming one of the group's largest revenue sources. Its clients include not only NVIDIA, but also Microsoft, Amazon AWS, Google, Meta, Oracle, and a large number of global cloud computing service providers. This Taiwan-based contract manufacturing giant also generates a significant portion of its revenue from assembling Apple's iPhones and MacBooks. However, in the past year, its AI server-related contract manufacturing and manufacturing business has surpassed its smartphone business, becoming the largest contributor to Foxconn's overall revenue, reflecting that the global AI data center construction boom is still in full swing. Investors regard Foxconn as the "core barometer of the AI computing power industry chain," more accurately as the "core barometer of AI server/AI cabinet manufacturing and delivery". It is involved in the critical delivery link between NVIDIA GPUs, CPUs, DRAM/HBM/NAND, data center optical interconnect systems, Ethernet switching networks, liquid cooling systems, AI PCBs, server cabinet power supply integration and the large-scale cloud CapEx of major cloud providers; when Dell's server business surpasses smartphones as the largest revenue contributor, it means that global AI data center construction has moved from chip procurement to the deployment of complete server cabinets, clusters, and data center setups. Morgan Stanley previously stated that the AI computing arms race has entered a stage of systemic expansion, with the institution's capital expenditure expectations for US tech giants in 2026 revised up significantly from $433 billion a year ago to $805 billion, with expectations that capital expenditure in 2027 will reach $1.1 trillion, further revised up from the previous forecast of $950 billion. Morgan Stanley's latest expectations highlight that the supply chain bottleneck in AI computing power infrastructure has expanded from "large-scale acquisition of GPUs/ASICs" to "striving to simultaneously address the entire chain of AI data center delivery processes, including data center power equipment, liquid cooling, data center CPUs, DRAM/NAND/HBM, optical communication/optical interconnect, high-performance network interconnection, transformers, gas turbines, and more". From GPUs to whole clusters: Goldman Sachs reassesses AI computing landscape, trillion-dollar server super cycle accelerating realization "We have raised our average AI server forecast for the period from 2026 to 2030 by about 18%, and we currently expect the market size to reach approximately $1.24 trillion by 2030 (compared to our previous forecast of around $961 billion in the fourth quarter of 2025), compared to the AI server market size of about $300 billion at the end of 2025; of which, the shipment forecast has been raised by about 3%, and the average selling price (ASP) forecast has been raised by about 15%," Goldman Sachs analysts wrote in a research report to clients. "We have raised our average forecast for traditional servers from 2026 to 2030 by about 31%, and we currently expect the market size to reach approximately $164 billion by 2030 (compared to $105 billion in the fourth quarter of 2025); of which, the shipment forecast has been raised by about 19%, and the average selling price forecast has been raised by about 10%." Goldman Sachs updated the above forecast, coinciding with the powerful quarterly performance announcements from global AI server manufacturing leader Dell Technologies (DELL) and Hewlett Packard Enterprise (HPE), and the strong monthly sales data from Hon Hai Precision Industry Co, a manufacturer of AI server/AI cabinet computing power infrastructure clusters such as NVIDIA Blackwell/Vera Rubin. According to data from research firm 650 Group, Dell has significantly increased its market share in both the traditional server market and the AI server market, with particularly strong performance in the enterprise (enterprise) and emerging cloud computing service providers such as Coreweave. Goldman Sachs analysts explained in the report, "650 Group had previously forecast that Dell's traditional server revenue in the first quarter of the 2026 calendar year (C1Q26) would increase by 85% year-on-year (compared to Dell's actual reported value of 92%), driven mainly by a 24% increase in shipments and a 49% increase in average selling price. By customer type, 650 Group expects Dell's traditional server revenue growth in the first quarter of the 2026 calendar year to be relatively balanced between emerging cloud service providers and enterprise customers: revenue related to emerging cloud service providers is expected to increase by 102% year-on-year (with shipments increasing by 34% and average selling prices increasing by 50%); while revenue from enterprise customers is expected to increase by 91% year-on-year (ships increasing by 29% and average selling prices increasing by 48%). Dell's traditional server hardware revenue in the first quarter of the 2026 calendar year was approximately $8 billion, with enterprise customers accounting for 54% of the customer base, emerging cloud service providers for 40%, and hyperscalers and traditional computing service providers for 6% (compared to 52%, 36%, and 12% respectively in the first quarter of 2025). Dell has achieved a significant increase in market share based on traditional server revenue (30% in the first quarter of the 2026 calendar year compared to 20% in the first quarter of 2025), with market share from emerging cloud computing service providers increasing from 26% to 35%, and from enterprise customers increasing from 25% to 38%. In the first quarter of the 2026 calendar year, Dell's AI server revenue increased by $11.3 billion year-on-year (representing a significant increase of 622% year-on-year), driven mainly by a 300% increase in shipments and an 81% increase in average selling price. By specific customer types, 650 Group expects that Dell's revenue growth related to AI servers in the first quarter of the 2026 calendar year comes from a wide range of sources, with emerging cloud service providers contributing about $9 billion in incremental revenue and enterprise customers contributing about $2 billion in incremental revenue. Dell's market share based on AI server revenue calculation has also significantly increased (17% in the first quarter of the 2026 fiscal year compared to 5% in the first quarter of the 2025 fiscal year), with market share from emerging cloud service providers increasing from 24% to 48%, and market share from enterprise customers increasing from 17% to 47%." On the other hand, Goldman Sachs added that HPE "likely maintained its share in the traditional server market, but ceded some share in the AI server market". At the same time, Dell's strong market position in the AI server market means that another AI server manufacturer, Super Micro Computer (SMCI), is likely losing some market share in the AI server market, but gaining about 1 percentage point of market share in the traditional server market. "Barometer of the AI Computing Power Industry Chain" Hon Hai's Strong Monthly Data During the April-May period, Hon Hai Precision Industry Co, the global largest electronic product contract manufacturer and manufacturer of AI server/AI cabinet computing power infrastructure clusters such as NVIDIA Blackwell/Vera Rubin, achieved a significant overall revenue increase of 34%, highlighting the continuous strong demand for NVIDIA's AI server product line from major cloud computing companies and AI application leaders such as Anthropic and SpaceXAI. The AI servers manufactured by Hon Hai are crucial for meeting the near-quantum-level AI inference edge computing power resource needs of major cloud platforms like Amazon AWS, Microsoft Azure, as well as high efficiency operation of AI application product lines such as Claude, ChatGPT, and Gemini. With Broadcom's latest disclosure of AI semiconductor performance growth prospects falling short of Wall Street's expectations, there are signs of a cooling trend in the investment frenzy in the AI computing power industry chain, leading to a rotation towards value and defensive stocks. Recently, AI-related hot tech stocks have significantly retraced, especially AI semiconductor stocks such as Broadcom, ARM, and Micron Technology, causing a decline in the Nasdaq 100 index. The latest set of sales data released by Hon Hai has greatly strengthened the financial market AI investment theme closely related to strong demand for AI computing power infrastructure. The physical orders at Hon Hai are still validating that "AI capital expenditures are real and accelerating in implementation." Especially amidst the recent pullback trajectory in the AI computing power industry chain in the US stock market due to Broadcom's less impressive AI semiconductor performance guidance and the sharp drop in the stock prices of South Korea's two leading storage giants - SK Hynix and Samsung Electronics due to massive profit-taking sell-off, Hon Hai's data release holds even more positive signal significance. At a time when the AI super bull market is in a pullback trajectory, Hon Hai, the largest scale AI computing power cluster assembly partner for NVIDIA, continues to devour the "AI infrastructure dividend." Based on the monthly sales data released by Hon Hai, sales have accumulated to NT$1.69 trillion (approximately $53.6 billion) over the past two months (April and May). The latest expectations from Wall Street analysts show that Hon Hai's overall revenue for the quarter ending in June is expected to achieve a significant year-on-year increase of 32%. Calculated on a monthly basis, Hon Hai's revenue in May saw a year-on-year increase of 40%. The detailed data released by Hon Hai on Friday showed that the unaudited consolidated revenue in May reached NT$859.4 billion, representing a significant year-on-year increase of 39.57% and a month-on-month increase of 3.28%, setting a new high for the same period in previous years. In USD terms, Hon Hai's revenue in May increased by approximately 38.5% year-on-year and by 4.2% month-on-month. The company simultaneously disclosed that its cumulative revenue for the first five months of 2026 reached NT$3.82 trillion, representing a significant year-on-year increase of 31.79%, also setting a record high for this period in its history. Both the single-month revenue in May and the cumulative revenue for the first five months of 2026 have set new highs for Hon Hai in the same period in previous years. In terms of performance outlook, Hon Hai has qualitatively described the second quarter as "significantly better than previously expected growth," with a much more optimistic tone than previous guidance, but at the same time, it cautions that uncertainties arising from global political and economic fluctuations need to be monitored. Hon Hai also stated that the AI server rack business is expected to continue its growth momentum, providing ongoing support for overall revenue, and emphasized that the global AI server contract manufacturing and manufacturing demand around NVIDIA Blackwell and Vera Rubin's AI computing power clusters remains extremely strong, further strengthening the validation of "AI computing power demand frenzy" at the entity AI order level, highlighting that the massive AI server orders brought in by Hon Hai's largest scale AI computing power infrastructure manufacturing and manufacturing customer NVIDIA are strongly holding up the "AI super bull market," indicating that the bull market driven by AI is not merely a narrative in the capital markets. Hon Hai is the most crucial assembler of server clusters with NVIDIA AI accelerators, and the company expects its AI server business to expand significantly in 2026. With the four major US tech giants Google parent Alphabet Inc., Amazon, Facebook parent Meta Platforms, and Microsoft collectively earmarking at least $725 billion for AI computing power infrastructure expenditures this year, global demand for AI computing power continues to soar, despite ongoing concerns about AI bubble related to excess capacity and how to monetize this cutting-edge global technology on a large scale.