TL;DR
Global corporate spending on AI has surged to nearly $1.5 trillion this year, prompting comparisons to major wartime expenditures even as measurable business returns remain scarce. Analysts and researchers warn of delayed projects, high failure rates for AI agents, and growing scrutiny of continued massive investment.
What happened
This year corporate and infrastructure investment in artificial intelligence ballooned to levels that industry observers compare with major wartime outlays. Gartner forecasts nearly $1.5 trillion in AI spending for the year, a sum similar to estimated U.S. direct costs for the Iraq and Afghanistan wars from 2001–2014. That scale of deployment has not yet produced broad, verifiable improvements in key financial metrics, according to commentary and industry surveys cited alongside the spending figures. Executives across banking and technology have leaned on talk of AI and agents in public communications, while some economists and analysts argue AI spending is propping up capital investment growth. But research and vendor-reported deployments raise doubts: a Carnegie Mellon study reported a roughly 70% failure rate for AI agents, and Forrester found about 25% of surveyed enterprises plan to delay AI spending into 2027. The debate centers on whether today's heavy outlays will translate into durable business value or represent unsustainable hype.
Why it matters
- Large, sustained capital flows into AI affect broader corporate investment decisions and may crowd out other spending.
- If AI projects fail to deliver measurable financial returns, companies could face write-downs and investor backlash.
- High failure rates for agent-style systems could slow adoption and increase caution among enterprise buyers.
- Delays or cutbacks in AI spending would impact vendors, hyperscale datacenter plans and associated supply chains.
Key facts
- Gartner forecasts nearly $1.5 trillion in global AI spending for the year.
- The article compares that figure to estimated U.S. direct spending of $1.5–$1.7 trillion on the Iraq and Afghanistan wars (2001–2014).
- Some economists have suggested AI investment is the primary factor keeping the U.S. out of recession.
- Apollo Global Management economist Torsten Slok said there is essentially no growth in corporate capital spending outside AI.
- Banks and other corporations are heavily promoting AI and 'agents' in public communications to sustain investment momentum.
- A Carnegie Mellon University study cited in the piece reported that AI agents fail about 70% of the time.
- Forrester research indicated roughly 25% of surveyed enterprises planned to delay AI spending into 2027.
- Executives and vendors continue to pitch agent-based systems despite mixed evidence of business impact.
What to watch next
- Whether the portion of enterprises delaying AI projects (Forrester: ~25%) grows or shrinks in 2026–2027.
- Whether AI investments begin to show measurable improvements in core financial metrics such as EBITDA.
- How the high reported failure rate for AI agents (Carnegie Mellon: ~70%) affects enterprise procurement and vendor roadmaps.
Quick glossary
- Artificial Intelligence (AI): Computer systems or software that perform tasks typically requiring human intelligence, such as pattern recognition, language processing, or decision-making.
- AI agent: An automated software entity that performs tasks or makes decisions on behalf of a user, often by combining models, tools and external data sources.
- EBITDA: Earnings before interest, taxes, depreciation and amortization; a common measure of a company's operating performance.
- Capex: Capital expenditures — funds used by a company to acquire, upgrade or maintain physical assets such as buildings, equipment or data centers.
- Datacenter supercluster: A very large, geographically distributed collection of datacenters and compute resources designed to support high-performance workloads at scale.
Reader FAQ
How much did companies spend on AI this year?
Gartner forecasted nearly $1.5 trillion in global AI spending for the year.
Is AI already delivering measurable financial returns for businesses?
The source says investments have so far produced limited visible gains in key financial metrics and that some customers are questioning returns.
Are enterprises delaying AI projects?
Forrester reported about 25% of surveyed enterprises planned to delay AI spending into 2027.
Do AI agents work reliably in production?
A Carnegie Mellon University study cited in the source found AI agents failed about 70% of the time.
AI + ML 3 AI faces closing time at the cash buffet Will businesses continue to invest in something that's shown so little return? O'Ryan Johnson Wed 24 Dec 2025 // 19:23 UTC OPINION…
Sources
- AI faces closing time at the cash buffet
- Something Ominous Is Happening in the AI Economy
- The AI Boom: Bubble Risk Or Durable Cycle?
- From hype to hard returns: AI enters a new phase
Related posts
- Can AI Fix Operating Rooms? Akara Uses Sensors and Models to Try
- How Mill Sealed an Amazon and Whole Foods Deal for Food Waste Bins
- Waymo trials Google’s Gemini as an in-car AI assistant in robotaxis