How to think about your AI investments

Tech giants recently announced some AI capex spending plans so large they’re blowing our feeble human minds.

While the vast majority of companies aren’t planning to spend $200 billion on AI infrastructure like Amazon, many expect to increase spending on the technology this year.

AI capex budgets at this stage are strategic statements as much as actual spending plans, according to experts, who also viewed AI investments as a long-term production booster versus an operational line item.

For some companies, notably those producing information rather than a physical product, AI is “their biggest capital allocation” right now, according to Aamer Baig, a McKinsey senior partner who leads the firm’s technology practice. “We know that is the province of the CFO, so applying the same sort of discipline of capital allocation to AI has been very helpful” to organizations, Baig told CFO Brew.

Making AI investment decisions requires “a collaborative approach across our leadership team and transparent communication,” Chris Miorin, CFO and COO of supply chain risk platform developer apexanalytix, told us.

Miorin said his firm will have more capital expenses this year on AI, particularly around “AI-specific hardware capacity” and associated software, which it will build in-house. It will hire people to develop this software, which Miorin described as a “capitalized R&D expense.”

Big money. Report after report signals 2026 will see a big uptick in AI spending across organizations.

A McKinsey survey of 600-plus tech and business leaders found that half the respondents planned to increase their technology budgets this year by more than 4%, and half identified AI as a priority investment—putting it ahead of all other tech priorities including cybersecurity and modernizing infrastructure. More than half (54%) of top performers (orgs with at least 10% revenue and EBIT growth over three years) said AI was an investment priority this year.

Tech giants announced huge AI-fueled capex plans in recent earnings calls. Amazon’s $200 billion capex budget is up significantly from $128.3 billion in 2025, and Google’s $185 billion spending plan is double the $91.4 billion it spent last year.

“What we’re facing is an arms race, with companies trying to make the biggest statements possible and to use the largest numbers possible,” James Gellert, cofounder and executive chair of financial health data and analytics company RapidRatings, told CFO Brew. “But these numbers are truly extraordinary.”

Gellert said he suspects these companies may not actually spend these vast sums of money, at least right away. Rather, “they’re making a statement of intent” and want to show they’re serious about investing in AI over the long haul.

Get your mind right. Baig said that top-performing organizations view AI investments as an input to the products they make, rather than an operational expense—even if that’s not always how they’re classified on financial statements. These companies view AI as “not a budget line item,” but an investment whose benefit comes from more efficient production over time.

Echoing Baig, Miorin said apexanalytix views AI investment through “the productivity gains that we’re forecasting, the product roadmap acceleration that we’re expecting to see from that investment, that’s where we’re really underwriting the return on capital.”

Miorin said that when his firm’s technology business unit brings him a proposed capital investment, he first needs to “thoughtfully model out the expected returns” of that proposal. Then he approaches other departments to see if a different, similarly sized investment would yield a better return.

“We have to constantly weigh all these projects and their expected returns, and whoever gets the highest expected risk-adjusted returns—which is a very important piece of that calculation—gets the dollars,” he said.