Market forecasts for the technology sector are rarely so clear-cut. According to the latest data from analyst firm Gartner, global IT spending will reach $6.31 trillion in 2026. This is evidence of a shift in the centre of gravity of global business. The 13.5 per cent year-on-year increase, significantly higher than previous estimates, is a direct result of the artificial intelligence infrastructure arms race.
A foundation of concrete and silicon: Exploding the data centre sector
The most glaring point in the report is the dynamics of investment in data centres. Gartner predicts that spending in this segment will grow by 55.8% in 2026, surpassing the $788 billion barrier. To understand the scale of this phenomenon, it is important to look at it through the lens of technological change: we are not dealing with a simple expansion of existing resources, but with a complete reconfiguration of computing architecture.
Traditional data centres, optimised for data storage and standard business applications, are giving way to HPC facilities. These are designed for the specific requirements of graphics processing units (GPUs) and TPUs, which are at the heart of modern AI. The surge in investment extends not only to the servers themselves, but also to advanced liquid cooling systems, high-density power infrastructure and enabling technologies, without which scaling large-scale language models (LLMs) would be impossible.
In parallel, the IT services segment, infrastructure deployments and the IaaS model will generate a turnover of $1.87 trillion. This suggests that the market is ripe for consuming computing power in a hybrid model, where physical infrastructure goes hand in hand with specialised management.
The dominance of hyperscalers: The computing oligopoly
A phenomenon of a structural nature is the increasing concentration of computing power in the hands of a few players. By 2031, hyperscalers – mainly Microsoft, Google (Alphabet) and AWS (Amazon) – are forecast to control as much as 67% of global data centre capacity.
This year alone, these three giants plan to spend more than $500 billion on capital expenditure related to AI infrastructure. Such gigantic outlays create a barrier to entry almost impossible for new players to overcome. For businesses, this means that they have to strategically choose a cloud provider that de facto becomes a partner in delivering a data-driven competitive advantage.
We are also seeing a new geopolitical map of IT investment. Microsoft’s $25 billion investment in Australia or Meta’s construction of the world’s 32nd data centre show that the availability of stable energy sources and space is becoming more important than proximity to traditional business clusters.
Strategic alliances and supply chain
Analysis of recent market deals sheds light on the direction in which the industry is heading. Anthropic’s agreements with Google and Broadcom to supply TPU (Tensor Processing Unit) power from 2027 onwards point to the growing importance of proprietary chips to make the giants independent of the dominance of third-party processor suppliers.
Even the biggest players need flexibility and specialised GPU cloud providers to cope with surges in computing power demand, as evidenced by Meta’s $21 billion partnership with CoreWeave. The biggest profits will be generated not by the AI developers themselves, but by the companies supplying the ‘components’ of this revolution – from accelerator manufacturers to power suppliers.
Market insights for business
In the context of the upcoming 2026 Investment Summit, business leaders should consider three key lessons:
- Infrastructure as a bottleneck: A 55.8% increase in spending on data centres suggests that access to computing power may become a scarce commodity. Companies planning large-scale AI deployments need to secure infrastructure resources in advance to avoid product development downtime.
- The need for cost optimisation: With IT spending reaching $6 trillion, efficiency becomes key. The shift from generic cloud solutions to AI-optimised infrastructure (such as IaaS supported by TPUs/GPUs) will determine the margins of digital projects.
- A new ecosystem of suppliers: Companies such as Broadcom and CoreWeave are worth watching. They represent a new category of technology partners who, through specialisation, are able to provide the components needed to scale AI faster and cheaper than traditional hardware suppliers.
