Why AI Infrastructure Costs Are Quietly Inflating Your Software Budget

AI Is Becoming a Line Item on Every Invoice

AI infrastructure costs are the capital and operating expenses required to build, power, and maintain the data centers, models, and cloud systems that deliver AI-driven features inside modern software, and those costs are increasingly being recovered through software subscription price increases and new cloud usage billing charges rather than absorbed by the vendors themselves. This is not a theoretical concern: a major tech research firm now warns customers to brace for bigger software bills as AI and software vendors raise prices and add usage-based fees to pass their AI costs directly to customers. Working from a survey of more than 2,600 business and technology decision-makers, it found software budgets were expected to rise for most organizations as these pricing tactics spread across the market.

Usage-Based AI Pricing: From Flat Subscriptions to Metered Spend

The pivot from flat subscriptions to metered consumption is the clearest sign that AI infrastructure costs are being pushed downstream. In the last six months, Anthropic, OpenAI, and GitHub have shifted some services away from flat-rate subscriptions toward usage-based billing, raising cost concerns among users who now have to track tokens, calls, and credits instead of a predictable monthly fee. According to the tech research firm, software budgets are expected to rise "as vendors increase prices or add usage charges to pass their AI costs to customers". Bain & Company has estimated that the build cost for AI datacenters will hit around USD 2 trillion (approx. RM9.2 trillion) by 2030, and vendors clearly do not intend to carry that burden alone. As usage-based pricing models become more common, many organizations are still scrambling to forecast, monitor, and manage AI spending effectively.

Squarespace’s 26% Hike Shows How AI Features Become a Tax

If you want a concrete example of software subscription price increases driven by AI, look at Squarespace. Users have reported emails detailing impending price hikes of up to 26 percent, depending on plan and payment option. The Core plan on annual billing is jumping from the equivalent of USD 23 (approx. RM106) per month to USD 29 (approx. RM134), a 26 percent increase that adds USD 72 (approx. RM333) a year. The Plus plan is also climbing 26 percent, while the Basic plan rises from USD 16 (approx. RM74) to USD 19 (approx. RM88) per month on annual terms, a 19 percent increase. Photographers, designers, and other visual artists are now commiserating online and asking for alternative hosting options as the value equation shifts. The consensus among many of these users is that they are being asked to fund more AI tools on the platform—tools they did not ask for and do not particularly want.

Enterprise Budgets: AI Costs Creep While Staffing Stays High

The uncomfortable truth for enterprise IT leaders is that AI is not replacing their people; it is stacking new costs on top of existing ones. Staffing already accounted for 35 percent of IT budgets in 2025, and for 2027, 67 percent of tech decision-makers expect staffing budgets to increase, 23 percent expect them to stay flat, and only 10 percent expect them to decline. At the same time, AI is driving increases in data and software spending, with 80 percent of decision-makers expecting those budgets to rise. So the narrative that AI will “pay for itself” through layoffs is at odds with where the money is actually going. The tech research firm notes that the “AI washing of layoffs” will continue, but warns organizations to guard against inflated promises that AI can replace employees across the board. In practice, data and analytics staffing is expected to grow, not shrink.

What IT Teams Should Do Now: Treat AI Like a Cost, Not Magic

Enterprise IT teams cannot stop vendors from chasing AI-driven revenue, but they can stop AI from quietly consuming the entire enterprise software budget. The starting point is a rigorous audit of AI service contracts and usage patterns, especially where cloud usage billing charges and token-based models have replaced flat fees. The tech research firm argues that organizations should adapt their FinOps practices to handle unpredictable AI costs. "Traditional FinOps wasn't built for token-based, usage-driven AI costs, but that team is certainly best positioned to build these new capabilities and must make this leap in 2027". Its report recommends funding runtime cost controls such as model routing, semantic caching, and usage guardrails to prevent runaway spend. The organizations that outperform in 2027, it concludes, will not be those spending the most on AI, but those investing in trusted data, strong governance, and the ability to adapt as technology and customer behavior evolve.