AI spending rises to $2.7 trillion

Research firm Gartner predicts that global spending on artificial intelligence (AI) will rise to around $2.7 trillion in 2023. This is a significant increase of 49.5 percent compared to the previous year. The growth is primarily driven by investments in AI infrastructure and the integration of AI features into existing software. According to Gartner analyst John-David Lovelock, the demand for AI infrastructure remains strong, despite rising memory chip costs. He describes the construction of AI data centers as the largest infrastructure project undertaken by mankind.

The share of AI infrastructure in total AI spending is remarkable; in 2026, this segment is expected to reach nearly $1.5 trillion, compared to more than $980 billion a year earlier. This year, infrastructure already accounts for more than half of all AI spending. Large hyperscalers and service providers purchasing AI-optimized servers are the biggest cost contributors, with companies like Nscale, Nebius, CoreWeave, Lambda, Crusoe, and Core Scientific benefiting from this demand.

In addition to infrastructure, the market for AI software is also growing significantly, with an expected value of over $461 billion in 2023. Within this, AI agents, assistants, and generative AI models are the main growth engines. There is also a strong increase in AI cybersecurity, which is expected to rise from nearly $26 billion in 2025 to over $51 billion this year. This can be explained by organizations wanting to secure their existing systems faster through the integration of AI features that vendors are already incorporating.

However, Lovelock notes a shift in how companies approach their AI strategies. Fewer organizations are opting for large transformation projects with service providers. Instead, they are investing in smaller, indirect projects for better utilization of AI features in their existing software. For the coming year, Gartner expects spending on AI infrastructure to further rise to nearly $2 trillion.

Read the full article from AI Insider.