/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

days · browse · Enter similar · o open

KPMG survey: only 26% of companies have a comprehensive view of their AI costs, while 50% have some visibility and 22% have none or only see costs after billing

Only 26% of companies say they have a comprehensive view of their AI costs, according to a new survey

Wall Street Journal

Context & Ripple Effects

This survey lands after related CEO surveys showed AI spending plans rising even as fewer than half of projects were reported to return more than they cost, and only a minority had delivered expected ROI or scaled enterprise-wide.

The cost-visibility finding adds an operational explanation for that gap: many companies are trying to assess AI’s business value without a complete view of the inputs behind it.

First-order effects

  • Companies without comprehensive AI cost visibility have an immediate budgeting and ROI-measurement problem: they cannot reliably compare AI projects, providers, or deployments on total cost.
  • Finance, procurement, and technology teams face pressure to assign and monitor AI spending before charges arrive, rather than treating it as an opaque shared technology expense.

Second-order effects

  • AI vendors and internal platform teams will face more scrutiny over usage reporting, cost allocation, and the ability to distinguish experimentation from production workloads.
  • As AI investment continues to rise, incomplete cost data can delay scale-up decisions and redirect spending toward projects whose costs and returns can be measured more clearly.

Third-order effects

  • If cost governance remains behind adoption, enterprise AI competition may shift from access to models toward the operational tooling needed to control, attribute, and justify AI spend.
  • The pattern suggests that broad AI experimentation will not automatically become enterprise-wide deployment; repeatable financial controls may become a prerequisite for scaling.

The trend: Enterprise AI is moving from an experimentation phase toward a cost-accountability phase, where spending visibility and measurable returns determine which deployments scale.

Discussion

  • @eglyman Eric Glyman on x
    As I wrote this, I saw X go into meltdown over tokens.  You've seen the headlines: “Uber blows yearly AI budget in just one quarter.”  “Meta employee burns 281 billion tokens in April.”  But, the problem isn't spending.  Spending works.  Since 2023, the top quartile of our AI spe…
  • Ritesh Patel Ritesh Patel on linkedin
    74% of companies don't have comprehensive visibility into what they're spending on AI.  —  That's from a KPMG survey cited in today's WSJ. …
  • @whocouldsay Evelyn on bluesky
    Maybe it's just because I had a decent amount of exposure to enterprise cloud billing at my last job, but the idea that any company is paying usage based billing with no controls, not even getting into the ai part of this, is bananas [embedded post]
  • @oisinmcgann Oisín McGann on bluesky
    I want to work in a business where someone's using my services on a regular basis and I can just bill them without having to explain what I'm billing them for.  [embedded post]
  • @edzitron.com Ed Zitron on bluesky
    The managerial class is being scammed in real time by their own egos.  If this was a project introduced by an employee it would've been killed in a month but because the executive sect has AI psychosis they'll tolerate the fact that Anthropic doesn't appear to provide realtime to…