AI is forcing technology companies to rethink the foundations of cloud computing, as rising demand for artificial intelligence tools drives a surge in spending on infrastructure.
Rather than focusing primarily on software, companies are now directing significant investment toward the physical systems needed to run AI workloads. This includes advanced chips, high-speed networking equipment, power systems, and large-scale data centres designed to handle intensive computing tasks.
The scale of this investment is expanding rapidly. Major US technology firms—including Alphabet, Amazon, Meta, and Microsoft—are projected to spend around US$650 billion on AI-related infrastructure in 2026, according to a Reuters-cited analysis. This marks a sharp increase from an estimated US$410 billion in 2025, highlighting how quickly demand is growing.
These investments are reshaping cloud platforms. The challenge is no longer just about building better software, but also about creating the physical capacity required to run AI systems efficiently at scale.
AI workloads drive infrastructure demand
Running large AI models requires enormous computing power. Training and deploying these systems often involves thousands of graphics processing units (GPUs) working together across distributed data centres.
This demand is accelerating innovation in networking and data transfer technologies. For example, Nvidia has announced plans to invest US$2 billion each in photonics companies Lumentum and Coherent to improve communication within AI data centres, according to Reuters.
Photonics technology uses light instead of electrical signals to transmit data, allowing faster speeds and lower power consumption. As AI systems grow in size and complexity, faster and more efficient data movement between processors is becoming critical.
This shift highlights a new bottleneck in AI development. For many organisations, the primary constraint is no longer software capabilities but the infrastructure required to support large-scale AI workloads.
Enterprise adoption fuels cloud growth
The rise in infrastructure demand is also being driven by increased enterprise adoption of AI. Businesses are integrating AI tools into operations for tasks such as data analysis, customer service automation, and internal productivity.
Most organisations lack the in-house resources to support these workloads, leading them to rely on cloud providers that offer access to large GPU clusters and specialised AI hardware.
According to Reuters, companies across the tech sector are entering long-term agreements to secure computing capacity. These multi-year deals, often worth billions, reflect growing competition for access to AI infrastructure.
As a result, cloud providers and hardware manufacturers are ramping up investments in data centres and supply chains to meet this demand.
The rising cost of AI infrastructure
Building AI-ready infrastructure is complex and expensive. Data centres designed for AI consume vast amounts of electricity and require advanced cooling systems to maintain safe operating conditions. They must also support high-speed networking capable of handling massive volumes of data.
The broader investment trend mirrors the rapid expansion of the AI sector. The Stanford AI Index Report estimates that global private investment in generative AI reached US$33.9 billion in 2024, up 18.7 percent from the previous year.
While that figure reflects startup and private funding, spending by large technology companies on infrastructure is significantly higher. Much of this capital is being directed toward building new data centres, securing energy resources, and developing advanced networking technologies.
Large-scale projects are also emerging to support future demand. One example is the Stargate initiative, backed by OpenAI, SoftBank, and Oracle, which plans to invest up to US$500 billion in AI infrastructure in the United States over several years.
Implications for enterprise cloud strategy
For enterprises, this surge in infrastructure investment signals a shift in cloud computing priorities. Providers are increasingly focusing on specialised hardware, large GPU clusters, and high-speed networks tailored for AI workloads.
Access to these resources is likely to become a key factor for organisations planning to deploy AI at scale. Companies may need to consider factors such as data centre location, availability of AI hardware, and long-term computing costs when choosing cloud providers.
As AI continues to evolve, the future of cloud computing will depend not only on software innovation but also on the ability to build and scale the physical infrastructure that powers it.
.webp)