Did you know AI Adoption Enters New Phase Driven by Infrastructure Overhauls and Embedded Features

 

According to Gartner forecasts, global spending on generative AI is going to increase a lot in 2025 with $644 billion in spending, about a 76.4% increase from 2024. Many industries are adopting AI and many other analyses like research conducted by AI at Wharton have predicted a 130% growth in AI spending. The report by Gartner also shows where most of the AI spending will go, with 80% of the $644 billion going to hardware investments. $398.3 billion with an increase of 99.5% will be for devices and with a growth of 33.1%, servers will see $180.6 billion in spending. Software spending will account for $37.2 billion with a growth of 93.9% while services will account for $27.8 billion with 162.6% growth.

Gartner says that AI spending for hardware is going to grow over time and the share of services and software spending is going to decrease as opposed to the expectations that it's going to increase. Many software are going to incorporate GenAI features, so less money will be needed for GenAI software spending. Many organizations will have to adjust their infrastructure planning and technology budgets accordingly as hardware continues to drive AI adoption. Discrete AI projects may become less common because AI features will be available in everyday tools and AI software will change how AI adoption is approached.

The report by Gartner also highlights a critical issue about GenAI proof-of-concept (PoC) projects failing to deliver the outcomes expected of them, which is creating a PoC graveyard. Three major barriers are contributing to the failure of these projects: data inadequacy, resistance to change, and RoI shortfalls. These barriers show that challenges to the adoption of GenAI are mainly because of organizational readiness rather than technical limitations. Gartner also predicts that enterprises are going to see a shift from developing internal generative AI solutions to using commercial off-the-shelf products in 2025. This change will come as companies have started realizing that custom AI projects come with more challenges and instead, they will use solutions that integrate AI into existing software, offering clearer ROI and easier implementation.