Independent · Source-driven · Last verified Sept. 4, 2026
AI agent infrastructure, documented layer by layer: protocols, orchestration, memory, identity, payments, observability, reasoning and compute. Twenty-seven reference articles and five opinions, each opening with a bold claim and settling into dated figures with sources readers can check.
A December 2025 arXiv paper from OpenServ Labs argues that bounded reasoning graphs, written in Mermaid and handed to nano-class models, deliver flagship accuracy at a fraction of the price.
“Reasoning cost is one of the biggest hidden blockers to real autonomy.”Armağan Amcalar · Entrepreneur UK
8 min read · 14 sources · 5 figures
The stack, drawn
Hover a layer to see the articles that document it. Each layer carries a Pantone Color of the Year; the year is part of the label.
Two planes touch every layer: identity and security decide what an agent may do; observability and evaluation record what it did. Foundations and Opinion sit outside the diagram as atlas and editorial.
Contents
j k to move, ↵ to open
Opinion · by Ryan Elliott Dennis
Reasoning cost is the hidden lease every autonomous agent pays, and the December 2025 BRAID paper, the ARC Prize price curves and Anthropic's cache economics all point to a 2027 in which the cheapest correct answer wins.
7 min · 14 sources · opinion
Agent protocols such as MCP, A2A, x402, AP2, MPP and Mastercard's Agent Pay for Machines are the toll roads of the agent economy, and the booths will hold their margin longer than the cars that pass through them.
7 min · 14 sources · opinion
Agent payments on stablecoin rails are real and tiny, the data say machines pick rails by fee, finality and permission, and that makes BitMine's ether treasury a bet on the wrong variable.
7 min · 14 sources · opinion
AI agent identity, signed mandates and audit trails become the moat of the agent stack after the summer 2026 evaluation escapes, and by 2027 an agent carries a credential or gets blocked.
7 min · 14 sources · opinion
Small language models wrapped in bounded reasoning will carry most production agent traffic by 2027, and the margin moves from the vendors who train models to whoever owns the reasoning structure and the routing.
7 min · 14 sources · opinion
How to read this journal
The pillar article defines the field and fixes the vocabulary every other piece uses.
Each layer of the stack has its own articles, its own figures and its own Pantone year.
Every article ends with an AP-style source list. Machines can read the same text at /llms.txt.