📊 Full opportunity report: Raw-feed licensing. The contract that doesn’t exist yet. on ThorstenMeyerAI.com — validation score, market gap, and execution plan.

TL;DR

The industry lacks a standard contract for raw-feed licensing for AI downstream rewrites, creating a significant legal and economic gap. This gap mirrors historical issues in music licensing and is driven by conflicting interests among key stakeholders.

Industry experts confirm that a formal, industry-standard contract for raw-feed licensing for downstream AI content rewriting does not currently exist, despite the clear economic parallels with music streaming royalties. This gap has significant implications for licensing, pricing, and legal clarity in the evolving AI content ecosystem.

Training-data licensing and display licensing are well-established, with contracted deals and clear frameworks. However, the third category—raw-feed licensing for downstream per-audience rewrites—lacks a standardized contract, creating a legal and economic void. This missing contract is critical because the unit economics of AI rewriting are comparable to music streaming royalties, which have a long-standing statutory framework dating back to 1909.

According to Thorsten Meyer, a researcher in this field, the core issue is that the industry has yet to develop a comprehensive licensing structure for raw-feed content used in AI downstream rewriting. The absence of this contract means that parties like AI labs, publishers, wire cooperatives, and search engines operate without clear legal or pricing standards, leading to potential conflicts and misaligned incentives. The gap resembles the situation in early 20th-century music licensing, when legal frameworks were still being established after landmark cases and legislative actions.

Raw-Feed Licensing: The Contract That Doesn’t Exist Yet — Thorsten Meyer AI
FEED
● DISPATCH / MAY 2026
THORSTEN MEYER AI · POST-WIRE · § 02
POST-WIRE · 02
NEWS / LICENSING ECONOMICS
Essay · Contract-Forensic Analysis · 2026-05-17

Raw-Feed Licensing:
The Contract That
Doesn’t Exist Yet

Training-data licensing is contracted. Display licensing is contracted. The third category — the post-wire one — has no contract.
Spotify pays songwriters ~$0.004 per stream. Apple Music pays ~$0.008. The Copyright Royalty Board under Phonorecords IV sets the all-in mechanical streaming royalty at 15.1% (2023) → 15.35% (2027) of platform revenue. Per-rewrite LLM inference cost lands in the same band: $0.003–$0.02, local open-weight to higher-tier cloud. The numbers collide, and the contract category that should price them against each other — raw-feed licensing for downstream per-audience rewrite — has not been written. This piece walks through what the contract should specify, why it isn’t there, and who structurally doesn’t want it written.
$0.004
Avg Spotify per-stream
royalty (2025)
$0.003
Per-rewrite inference cost
local Mac fleet, open-weight
15.35%
Phonorecords IV mechanical
streaming rate by 2027
$3B+
MLC payouts since 2021
(scaffolding scale)
SPOTIFY $0.004/STREAM· APPLE MUSIC $0.008/STREAM· TIDAL $0.01284/STREAM· YOUTUBE MUSIC ~$0.005-0.007· PHONORECORDS IV 15.1%→15.35%· MECHANICAL RATE 12.7¢ (2025)· 1909 COPYRIGHT ACT· 1976 REVISION· DPRA 1995· MMA 2018· MLC $3B PAYOUTS· TOLLBIT 7000 SITES· TOLLBIT $24M SERIES A· 730% BOT-PAYWALL GROWTH· ARC XP 2000+ PROPERTIES· CHATGPT 87.8% AI-BOT TRAFFIC· RAW-FEED CONTRACT MISSING· SPOTIFY $0.004/STREAM· APPLE MUSIC $0.008/STREAM· TIDAL $0.01284/STREAM· YOUTUBE MUSIC ~$0.005-0.007· PHONORECORDS IV 15.1%→15.35%· MECHANICAL RATE 12.7¢ (2025)· 1909 COPYRIGHT ACT· 1976 REVISION· DPRA 1995· MMA 2018· MLC $3B PAYOUTS· TOLLBIT 7000 SITES· TOLLBIT $24M SERIES A· 730% BOT-PAYWALL GROWTH· ARC XP 2000+ PROPERTIES· CHATGPT 87.8% AI-BOT TRAFFIC· RAW-FEED CONTRACT MISSING·
FIG. 01 — THE THREE LICENSE CATEGORIES
Two contracts written, one missing
The AI-publisher licensing market sorts into three structural categories — and only two are contracted today
CATEGORY A
Training-data
Archive-shaped · One-shot · Fixed term
AP–OpenAI 2023 (archive 1985→)
Reddit–OpenAI 2024
Stack Overflow–OpenAI 2024
Shutterstock multi-deal
CATEGORY B
Display
Chat-shaped · Attribution-bound · Brand-tier priced
News Corp–OpenAI $250M/5yr
News Corp–Meta $150M/3yr
Axel Springer ~$13M/yr
FT $5–10M/yr · AP–Google
CATEGORY C
Raw-feed-rewrite
Post-wire-shaped · Per-audience derivative-work production
Mistral–AFP (2,300/day, structurally close but priced as display+RAG)

No standard contract.
No Standard
Contract
Training-data and display licensing assume the AI is a destination. Raw-feed-for-rewrite assumes the AI is an intermediate layer producing N derivative works for N downstream publication endpoints. That use case has no industry-standard pricing unit, no industry-standard attribution requirement, no industry-standard audit infrastructure. It just happens, unlicensed, in the gap.
FIG. 02 — THE COST COLLISION
Per-stream music royalty vs. per-rewrite inference cost
Both are units of derivative-work production at scale — and they sit in the same numerical neighbourhood
A · Music streaming royalty per stream · 2025
Spotify (avg)
$0.004
Apple Music (avg)
$0.008
Amazon Music
$0.006
YouTube Music Premium
$0.006
Tidal (highest)
$0.01284
Band: $0.003 — $0.013 per unit
B · Per-rewrite LLM inference · 600-word source
Local open-weight (Mac fleet)
$0.003
Cloud commodity (Haiku/4o-mini)
$0.007
Cloud mid-tier
$0.012
Cloud higher-tier
$0.020
50-site fan-out total
< $1
Band: $0.003 — $0.020 per unit
The collision is structural, not coincidental. Both rates are derivative-work production units operating at the same scale-economics — variable cost per piece of content, distributed across a pooled audience. If raw-feed licensing settled at a per-rewrite royalty in the same band ($0.005–$0.02), the wire cooperatives would have a defensible economic floor and the AI side would have a defensible variable-cost line item. Neither party has proposed this publicly.
FIG. 03 — THE 1909 PRECEDENT
The legal scaffolding music has and news doesn’t
117 years of statutory rate-setting, compulsory licensing, and collective collection infrastructure
1908
White-Smith Music Publishing v. Apollo — Supreme Court rules piano rolls aren’t “copies” of sheet music because humans can’t read them. Songwriters lose; mechanical reproduction unregulated.
1909
Copyright Act of 1909 — Congress overrides the Court; creates first compulsory mechanical license at 2¢ per unit. The original statutory rate-setting precedent.
1976
Copyright Act revision — Rate raised from 2¢ to 2.75¢ after 67 years frozen. Section 115 framework retained. Compulsory licensing extended to new media.
1995
Digital Performance Right in Sound Recordings Act — Extends mechanical licensing to digital downloads. Acknowledges new technology forms.
2018
Music Modernization Act — Establishes the Mechanical Licensing Collective. Blanket licensing for digital streaming services. Centralised collection infrastructure.
2023–27
Phonorecords IV (CRB) — Sets all-in mechanical streaming royalty rate at 15.1%→15.35% of platform revenue. Current statutory mechanical rate 12.7¢ per track.
2026
News raw-feed licensing — No statutory rate. No compulsory licensing regime. No central collective. No CRB-equivalent. The contract category exists structurally but has no scaffolding underneath it.
The pattern across 117 years: technology outruns licensing, lawsuit fails to protect rights-holders, Congress intervenes statutorily, rate-setting body resolves per-unit pricing, collective handles administration. News raw-feed licensing is currently at the “technology outruns licensing” step. The intervening steps will, on historical pattern, eventually follow — but they take decades. The Bartz $1.5B settlement and the NYT v. Perplexity complaint are the early lawsuit-failure-to-protect signals.
FIG. 04 — THE TOLLBIT GAP
The closest existing infrastructure stops short of raw-feed
TollBit operates ~7,000 publisher sites with two license types — neither addresses the post-wire category
LICENSE TYPE
USE CASE COVERED
STATUS
Summarization
AI cites or grounds an answer once with a single use of the page. Pricing per 1,000 pages accessed. RPM benchmark.
Contracted
via TollBit
Full Display
AI displays the complete text of an article once within its product. Per-1,000-pages pricing benchmarked against syndication rates.
Contracted
via TollBit
Model Training
Use of the content to train or fine-tune an AI model. TollBit explicitly does not permit either license type to extend to training.
Excluded
by both licenses
Raw-feed-rewrite
AI ingests the source feed and produces N differentiated rewrites for N downstream publication endpoints. The post-wire use case.
Not offered
as a license type
TollBit (founded 2023, ~7,000 publisher sites including TIME, Fast Company, Washington Post Arc XP, $24M Lightspeed Series A on top of seed) is the most-built piece of the raw-feed licensing infrastructure: detection, metering, rate-setting per 1,000 pages, payment routing, MCP-server integration. What the platform doesn’t have yet is the license category. Bot-paywall adoption grew 730% Q4 2024 → Q1 2025; ~20% of publishers earn revenue, in the hundreds-to-tens-of-thousands per month range. Necessary infrastructure, insufficient contract category.
FIG. 05 — FIVE CONTRACT SHAPES
What the missing contract could look like
Five plausible structures, scored on near-term feasibility · none currently leading
SH.
CONTRACT SHAPE
PRICING UNIT
NEAR-TERM
A
Per-rewrite royaltyMusic-streaming-mapped, pro-rata pool possible
$0.005–0.02 / rewrite
Medium
B
Per-source-story flat feeModified wire-subscription, simpler administration
Tiered $/story
High
C
Per-endpoint subscriptionExtension of existing AP/Reuters subscription model
$/endpoint/yr
Medium
D
Revenue-share on AI trafficAligns dollars with realised value · audit-heavy
% of attributed rev
Low
E
Statutory compulsory licenseCRB-equivalent for news · 1909-act-shaped
Statutory rate
Low (slow)
Near-term feasibility is not the same as long-term likelihood. The historical pattern (mechanical, broadcast, cable) suggests Shape E — statutory compulsory licensing — is where these gaps eventually settle, but on a 5–15 year timeline. The near-term outcomes (Shape A or B) will set the precedent the statutory regime eventually formalises. Whoever drafts the first major Shape A or B contract has disproportionate influence on what Shape E ends up codifying a decade later.
Per-stream music royalty and per-rewrite inference cost are in the same numerical neighbourhood because both are units of derivative-work production at scale. The contract that should price them against each other does not exist yet.
Thorsten Meyer · Raw-Feed Licensing · Post-Wire 02

Implications of the Missing Raw-Feed Contract

The lack of a standardized raw-feed licensing contract impacts multiple stakeholders: AI developers, publishers, and licensing bodies. Without clear legal and financial frameworks, parties risk disputes over rights, revenue sharing, and attribution, which could slow innovation and market growth. The situation also creates a mispricing of derivative works, similar to the historical misalignments in music royalties, potentially leading to legal challenges and regulatory intervention.

Amazon

AI raw feed licensing contracts

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Historical and Industry Context of Licensing Gaps

While training-data and display licensing are contractually well-defined, the third category—raw-feed licensing for downstream rewriting—remains unregulated. The economic collision between AI rewrite costs (~$0.003 to $0.02 per rewrite) and music streaming royalties (around $0.004 to $0.008 per stream) exemplifies the structural issues. The legal scaffolding for music, rooted in the 1909 Copyright Act and subsequent reforms, provides a precedent for how licensing gaps are eventually addressed. Currently, industry stakeholders are divided, with some preferring to maintain the status quo that favors their interests, delaying the creation of a comprehensive contract.

Thorsten Meyer notes that this situation mirrors the early 1900s, when the legal framework for music licensing was still evolving, and industry players were resistant to regulation that would limit their control or profits.

“The missing contract category for raw-feed licensing is a structural gap that could slow down AI innovation if not addressed.”

— Thorsten Meyer

Unresolved Legal and Industry Standoff

It remains unclear when or if a standardized raw-feed licensing contract will be established, as stakeholders such as AI labs, publishers, and search engines are divided. The exact terms, pricing models, and regulatory approaches are still under discussion, with no consensus yet reached.

Next Steps Toward Contract Formation

Industry stakeholders are expected to engage in negotiations over the coming months, possibly influenced by regulatory pressure or legal precedents. Policymakers and licensing bodies may also step in to propose frameworks, drawing lessons from historical licensing regimes like music royalties. The development of a formal contract will be crucial for clarifying rights, pricing, and attribution standards in AI content rewriting.

Key Questions

Why is there no standard contract for raw-feed licensing yet?

Stakeholders have conflicting interests and prefer to maintain control over the existing, unregulated gap, delaying the creation of a standardized framework.

How does the lack of a contract affect AI content rewriting?

Without clear licensing terms, there is legal uncertainty, potential disputes over rights and revenue, and a risk of slowing innovation in AI content generation.

What historical precedent exists for resolving such licensing gaps?

The music industry faced similar issues in the early 20th century, leading to the development of statutory licensing frameworks that eventually stabilized the market.

Who are the main parties involved in this licensing gap?

AI labs, brand-strong publishers, wire cooperatives, and search engines are the key stakeholders, each with differing interests and influence over the contractual process.

Source: ThorstenMeyerAI.com

You May Also Like

Canada: The Proof It Didn’t Keep

Canada demonstrated it can implement near-universal basic income through CERB in 2020, but political and fiscal limits have halted broader adoption.

Your Coding Agent Is an Attack Surface: The Claude Code Security Reckoning

Recent security flaws in Claude Code reveal attack surfaces in local configs and MCP integrations, risking token theft and code execution.

The SSD Squeeze: Why Storage Joined the Party

Record-breaking NAND shortages driven by AI demand and factory competition are causing sharp price increases in SSDs and enterprise storage.

The Machine Economy — Capital-Heavy, Human-Light, Trading With Itself

Analysis of the emerging machine economy where AI-driven firms operate with minimal human involvement, reshaping markets and economic structures.