/
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

Current AI market dynamics point to frontier models becoming commodity infrastructure as the token crunch eases, with value shifting to products built on top

AI is in a supply crunch today, but what happens when we come out of it? …

Benedict Evans

Context & Ripple Effects

Related coverage has tracked a tension between falling per-token prices and rising application costs as reasoning models consume more tokens. It also documented mounting competitive and financial pressure on frontier-model providers as models became harder to differentiate.

More recent analysis locates durable AI moats in private data, judgment, and verifiable workflows rather than benchmarkable model capabilities. The easing of token constraints makes that distinction more consequential for companies building on top of models.

First-order effects

  • Greater model capacity and less constrained token supply reduce the scarcity premium for frontier-model access, increasing pressure on providers to compete on price, availability, and service rather than raw model capability alone.
  • Application builders gain more flexibility to use and switch among model providers, while the relative importance of product design, workflow integration, and proprietary inputs rises.

Second-order effects

  • Model vendors are pushed toward bundling, distribution, enterprise relationships, and infrastructure economics to defend margins as base-model capability becomes less differentiated.
  • Developers may face lower unit pricing but not necessarily lower total AI spending: token-intensive reasoning workloads can expand as supply loosens, shifting cost management toward application-level efficiency and product ROI.

Third-order effects

  • If model access continues to standardize, industry value is likely to concentrate less in broadly benchmarkable intelligence and more in products that embed private data, trusted judgment, and operational workflows.
  • The market may increasingly resemble other infrastructure layers: a competitive supply base underneath a fragmented application layer, though frontier providers could retain leverage where capacity, reliability, or specialized capabilities remain scarce.

The trend: AI is moving from a supply-constrained frontier-model market toward a more commoditized infrastructure layer, with defensibility shifting to the applications and proprietary contexts built above it.

Discussion

  • Shivakumar Krishnamurthy Shivakumar Krishnamurthy on linkedin
    Two interesting essays from respected thinkers on how AI can escape the commodity trap and where is value captured.  One thing is certain, it is still early days. …
  • Benedict Evans Benedict Evans on linkedin
    A new essay - ways to think about token pricing.  —  AI is in a supply crunch today, but what happens when we come out of it? …
  • @akhilrao @akhilrao on bluesky
    also unclear to me: in whatever long run future we get to, will uncertainty over when AI lab revenues land remain large relative to what they can hedge? this seems to determine the premia the big labs will pay for compute and the kinds of power infrastructure we'll have in say a …
  • @davidcrespo @davidcrespo on bluesky
    I always liked Ben Evans. this is an especially good example of what a good version of Zitron's analysis would look like: axes of variation, how they could play out, this or that is more likely. as opposed to “I'm certain what will happen next despite having been wrong every day …
  • @claeshs Claes Holtzmann on bluesky
    Increasingly expensive frontier models... something's gotta give.  —  “every path to foundation models having market dominance, strategic leverage, value capture, winner-takes-all effects, or anything else other than becoming commodity infrastructure, requires something to change…