Artificial intelligence is set to trigger a vast expansion in global data centre investment, with cumulative capital expenditure forecast to reach $1.7 trillion by 2030, according to fresh industry projections that point to an infrastructure supercycle reshaping the technology landscape.Research firm Dell’Oro Group says the scale and speed of AI deployment are pushing hyperscale cloud providers and enterprises to accelerate spending on servers, networking equipment and power-intensive facilities at levels not seen in previous cloud buildouts. The firm’s latest outlook suggests that the global data centre industry has entered a new phase of structural growth, underpinned by generative AI workloads that require far greater computing density and advanced silicon.
The projected $1.7 trillion figure represents aggregate spending on data centre IT equipment over the remainder of the decade. Analysts say the bulk of this investment will be concentrated among large cloud providers in North America, Europe and parts of Asia, with the United States retaining a dominant share as companies race to deploy AI infrastructure at scale.
Hyperscalers such as Microsoft, Amazon and Alphabet have already lifted their capital expenditure guidance over the past year, citing demand for AI services. Microsoft disclosed plans to spend tens of billions of dollars on infrastructure, much of it directed towards expanding Azure’s AI capabilities. Amazon Web Services and Google Cloud have made similar commitments, while Meta Platforms has sharply increased spending on AI-optimised servers and data centre capacity to support its generative AI initiatives.
The spending wave is closely tied to the rapid adoption of large language models and other generative AI systems that require specialised graphics processing units and high-bandwidth networking. Nvidia, which has become the central supplier of AI accelerators, has seen revenue surge as cloud providers compete to secure its latest chips. Advanced Micro Devices and Intel are also seeking a greater share of the AI silicon market, while networking vendors are benefiting from demand for high-speed interconnects linking thousands of GPUs inside single facilities.
Dell’Oro Group’s assessment aligns with broader market expectations that AI-related infrastructure will drive multi-year growth across the semiconductor and data centre supply chain. Industry executives describe a shift from incremental capacity upgrades to wholesale architectural redesigns, with racks configured around AI clusters rather than traditional enterprise workloads.
Energy consumption and power availability have emerged as critical constraints in this expansion. AI data centres draw significantly more electricity than conventional facilities, prompting operators to secure long-term power agreements and explore alternative energy sources. In the United States, utilities in states such as Virginia and Texas are facing mounting requests for new grid connections as data centre developers seek sites with reliable power and fibre connectivity.
European governments are also grappling with the implications of large-scale AI infrastructure. Policymakers have voiced concerns about energy usage, water consumption for cooling and the environmental footprint of sprawling data centre campuses. At the same time, officials recognise the strategic importance of hosting AI infrastructure domestically to support innovation and digital sovereignty.
Asia is expected to remain a key growth region, with China continuing to invest heavily in domestic data centre capacity despite export controls affecting access to the most advanced chips. Southeast Asian markets such as Singapore and Malaysia have attracted new projects, although land and power constraints have led some governments to impose tighter planning rules.
The $1.7 trillion spending forecast underscores a broader reordering of capital allocation within the technology sector. Over the past decade, cloud computing drove a steady rise in data centre investment, but the advent of generative AI has accelerated that trajectory. Analysts describe this as a supercycle because it combines new application demand, rapid hardware innovation and strategic competition among global technology leaders.
Financial markets have responded accordingly. Shares of chipmakers and data centre equipment suppliers have outperformed broader indices, reflecting investor expectations of sustained earnings growth. At the same time, some analysts caution that the pace of spending may fluctuate depending on macroeconomic conditions, supply chain bottlenecks and the commercial returns generated by AI applications.
Questions remain over whether revenue from AI services will justify the scale of capital deployed. While enterprises are experimenting with generative AI tools for customer service, software development and data analysis, monetisation models are still evolving. Technology executives argue that early investment is essential to secure leadership in a market they believe will transform industries ranging from healthcare to finance.
Infrastructure providers are already adapting to new requirements. Liquid cooling systems are being introduced to manage the heat generated by dense AI clusters. Data centre designs are evolving to accommodate higher rack power densities, sometimes exceeding 100 kilowatts per rack, far above historical norms. Supply chains for transformers, switchgear and backup power systems are under strain as demand intensifies.
Topics
Technology