However, a Jefferies report has pointed out that decelerating growth among software companies suggests challenges in translating this surge into immediate returns.

Semiconductor stocks have risen 43 per cent year-to-date, fuelled by AI optimism, while the IGV software index has gained 30 per cent.

 Despite its high-profile investment in OpenAI, Microsoft’s stock has underperformed the IGV by 19 per cent since the release of ChatGPT.

 According to Jefferies, Microsoft’s AI revenue accounts for just three per cent of total revenue.

Other major players like Snowflake and Salesforce expect negligible AI contributions to their fiscal 2025 performance, while Adobe’s Firefly AI, launched in March 2023, has not boosted revenue.

The Jefferies report highlights widespread challenges across the software sector. Over 60 per cent of companies report that their current architecture cannot support AI workloads without significant modifications. Data engineers face inefficiencies, with half spending most of their time resolving data source connections. Real-time processing and data pipeline constraints remain primary hurdles.

A survey by Jefferies found that large enterprises lead in AI adoption, benefiting from existing data infrastructure. However, an AWS partner consultant noted, “a very low percentage of POCs are making it into production.” Training costs are another major barrier, with costs for frontier AI models reaching $100 million.

Physical infrastructure limitations compound the problem. Meta has cited power consumption during model training as stretching the limits of the power grid, while Microsoft’s Corporate VP of Azure Hardware acknowledged that AI infrastructure “can hardly meet the needs of AI model development.”

Scale AI CEO Alexandr Wang said the industry has “hit a wall on pre-training.”

Meanwhile, companies are pivoting towards purchased AI solutions. A Salesforce partner highlighted how a customer replaced a team of “50 people trying to build agents” with Salesforce’s AI solution in mere hours.

Despite this shift, infrastructure spending is accelerating while software revenue growth lags, with the IGV index trailing semiconductor stocks by 13 percentage points in 2024.