Gartner: Half of Gen AI Projects Might Exceed Budget Plan by 2028
Organizations may be undervaluing the actual expense of generative AI as they move from experimentation to production, according to Gartner’s “10 Best Practices for Optimizing Generative and Agentic AI Costs” report.
“Organizations transitioning from GenAI pilots to production experience an impolite awakening when it concerns expenses,” Gartner researchers found. “Producing a production-ready GenAI system can be orders of magnitude more expensive than running a pilot.”
The market watchers forecast that a minimum of 50 percent of GenAI initiatives will surpass their prepared spending plans by 2028 due to poor architectural choices and a lack of functional knowledge.
The warning reflects a growing challenge dealing with the AI industry. While much of the conversation has focused on design capabilities, Gartner argues that the genuine test for business will be operating AI systems effectively at scale.
< img src="https://pubads.g.doubleclick.net/gampad/ad?iu=/5978/eof.cam&t=item%253db7113d0e_2c4c_4fc3_bf6d_b913d27ec1bb%26pos%253dbox_c1%26Topic%253dArtificial_Intelligence%252cGenAI%252cResearch%252cBreaking_News%252cCentral_IT%252cIT_Leadership%252cARTICLE_TYPE%252cAUDIENCE&sz=300x250|640x481 & tile = 4 & c = 123456789"alt =" "/ > A major driver of those expenses is inference, the process of utilizing a qualified AI design to react to triggers, generate content, examine data, or carry out other jobs in production. Unlike training, which is usually a large upfront expense, inference costs repeat each time users or applications call the model. Gartner expects inference to account for at least 70 percent of a model’s lifetime expenses, shifting attention away from training and toward the day-to-day realities of serving AI work at scale.
The obstacle becomes even greater with agentic AI. Unlike standard chatbots that create a single response, AI agents can activate several model calls, retrieve data, gain access to external tools, and perform multi-step workflows.
As organizations deploy more self-governing systems, AI use and associated expenses can rise considerably.
The message is that success in the AI period will depend on more than model efficiency. Gartner declares that companies must focus on expense governance, architectural performance, model choice, and usage tracking to scale generative and agentic AI without sustaining unsustainable costs.
“Through 2028, at least 50% of GenAI jobs will overrun their budgeted costs due to bad architectural options and absence of functional know-how,” the report kept in mind.