The Number CPG Finance Chiefs Need to See Now
Bain & Company released analysis this week projecting that a typical consumer packaged goods company with $10 billion in revenue could see its IT costs climb 75% by 2035, driven almost entirely by the demands of scaling AI. The projection lands alongside a Gartner forecast that spending on AI models and platforms specifically will jump 63% to reach $64 billion, inside a broader global IT spending increase of 14.2% year over year in 2026. Together, the two forecasts describe an AI cost curve that is steepening faster than most CPG technology budgets were built to absorb.
For CPG companies already running lean technology organizations relative to pure retailers, a 75% cost increase over a decade is a number that should be showing up in five-year capital plans being drafted right now, not a distant planning problem. Bain's analysis names the specific cost drivers behind the increase: infrastructure requirements, security investments, talent acquisition and retention, architectural complexity, data governance requirements, and AI platform obsolescence, a list broad enough to touch nearly every line of a technology budget.
Where the Money Actually Goes
The cost drivers Bain identifies are not primarily about buying more AI licenses. Infrastructure and security spending typically dwarf software licensing costs once a company moves AI workloads from pilot to production, because production systems require redundancy, monitoring and compliance controls that pilots never needed. Talent costs compound the problem, since the specialists who can build and govern production AI systems remain scarce and expensive relative to the broader technology labor market.
Architectural complexity and platform obsolescence round out the list, and both point to a less obvious cost, rework. As AI platforms evolve quickly, CPG companies that commit early to a specific vendor or architecture risk having to rebuild significant portions of their AI stack within a few years, a cost that rarely shows up in the initial business case but shows up clearly in Bain's decade-long projection.
The Waste Problem Hiding Inside the Growth Number
Bain's research found that roughly one-quarter of current AI spending is being wasted on pilots that get powered down before ever reaching production, and fewer than 25% of enterprises have successfully scaled an AI initiative past the pilot stage. That means the 75% cost increase Bain projects is happening even as a large share of current AI spending produces no lasting output, a combination that should worry any CPG board approving AI budgets without clear scaling criteria attached.
The waste figure also suggests that simply spending more will not solve the underlying problem. A CPG company that increases its AI budget without fixing the reasons pilots stall, unclear ownership, weak data foundations, insufficient production infrastructure, will likely see its waste percentage hold steady even as absolute spending climbs, compounding the cost increase Bain is warning about rather than justifying it.
The Double Bind Bain Describes
Danielle Burgs Escobar, who leads Bain & Company's U.K. enterprise technology practice, described the risk in blunt terms: "If you underinvest, you get left behind. Your competitive advantage erodes." But she was equally direct about the opposite failure mode: "When you overspend, you lose credibility, become just a cost center." Escobar added that the specific severity of the cost increase will vary by company, saying it is "probably more or less extreme, but the trend is going to be pretty universal."
That framing puts CPG technology leaders in a genuinely difficult position, since the penalty for getting the AI budget wrong exists in both directions rather than just one. A CTO who under-funds AI risks watching competitors pull ahead on personalization, supply chain optimization and demand forecasting. A CTO who over-funds it risks a board conversation about why the technology function's costs keep rising without a matching revenue or efficiency story to show for it.
What Separates the Companies That Scale AI Successfully
The fewer than 25% of enterprises that have successfully scaled AI, per Bain's research, appear to share a common trait: they treat production infrastructure and governance as part of the initial AI investment rather than an afterthought bolted on once a pilot proves promising. That is a more expensive way to start, but it avoids the rework costs that Bain identifies as one of the biggest drivers of the long-term cost increase.
It also means the CPG companies most likely to avoid the worst of Bain's 75% projection are the ones willing to spend more upfront on the boring parts, security architecture, data governance, talent retention, rather than racing to ship the most visible AI feature first. That is a harder case to make to a board excited about AI's top-line potential, but Bain's research suggests it is the case CPG technology leaders need to be making anyway.
Building a Budget That Survives Either Direction
The practical takeaway for CPG and retail technology leaders is to build AI budget requests that explicitly name which failure mode they are managing against, the cost of falling behind competitors or the cost of an AI function that cannot show a return. Bain's research gives you the language to make that trade-off explicit to a board, rather than letting the budget conversation default to a simple up-or-down vote on total spending.
Given that Gartner's 63% jump in AI platform spending is already underway in 2026, the window to build these governance structures before costs accelerate further is narrowing. CPG technology leaders who wait until the 75% increase Bain projects has already materialized will be retrofitting governance onto a much larger, harder-to-unwind budget than the one they have the chance to design deliberately today.
