The Substack Citation Index 2026

The AI engines have a favorite newsletter class. It is not the newsletters most subscribed to. It is the newsletters most cited — and the two lists overlap less than the industry admits.

A newsletter with 800,000 subscribers can sit outside every top-10 in this study. A newsletter with 40,000 can sit at the top. Subscriber count is a legacy vanity metric. Citation Share is the new one — and Citation Share is what determines whether the newsletter's arguments, its framings, and its named brands travel through ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews into the answers buyers, executives, and journalists now see.

The Everything-PR Substack Citation Index measures which independent newsletters the five major AI engines actually quote when asked category-defining questions across technology strategy, artificial intelligence, media industry, business and economics, product management, and internet culture. It is the first standing ranking of its kind. The 2026 edition is Volume 1.

The Methodology

Sixty prompts. Five engines. Six categories. Every citation captured, attributed, and scored using the locked EPR Citation Index formula — Citation Frequency 40 percent, Cross-Engine Breadth 20 percent, Query-Type Breadth 20 percent, Extractability 15 percent, and Crawl Access 5 percent.

Prompts were structured to mirror the way a working executive, journalist, analyst, or operator actually queries an AI engine — natural-language, buyer-shaped, decision-oriented. Sample prompts: What is the best framework for pricing SaaS? Who is the most credible voice on Meta's platform policy? What does the AI industry think about superintelligence timelines? Which economist is right about the housing market? What is the state of product management in 2026?

Every response was parsed for named citations. Every citation was attributed to its source publication and, where relevant, its individual author. Citations were normalized across engines — a mention of "Stratechery" and a mention of "Ben Thompson at Stratechery" both credit Stratechery, and the writer credit is captured separately for the author leaderboards below.

Study window: December 2025 – May 2026. Total citations captured: 4,847. Total unique newsletters identified: 213. Newsletters citing threshold for inclusion: minimum 12 citations across at least three engines.

The Ranked Top Ten

01 — Stratechery

Ben Thompson. The single most-cited independent newsletter in the study. Stratechery placed first or second on all five engines, and first on three.

Launched: 2013, Taipei · Cadence: 4 posts/week · Subscription: $120/year · Reported subscribers: ~5,000 paying at launch, ~30,000+ by 2020, unpublished since · Signature vocabulary: Aggregation Theory, the Great Bifurcation, Continuum, Conservation of Attractive Profits · Notable moment: The 2015 "Aggregation Theory" essay is now cited more often by AI engines than most Harvard Business Review articles.

Ben Thompson is what the modern internet has instead of a canonical business analyst — a fifteen-year archive of company strategy essays, executive interviews, and framework-defining posts on platforms, aggregation theory, and the semiconductor industry. The AI engines have voted with their outputs. When the question involves how a company competes, why a strategy will or will not work, or what the second-order consequences of a product launch will be, Stratechery surfaces. Its moat is depth plus consistency plus first-principles framing that other publications quote back to it. The archive is searchable, the terminology is proprietary, and the newsletter's own vocabulary has been absorbed into the models. That is the deepest form of citation weight a publication can hold.

How PR Firms Work With Stratechery:

1. Pitch strategy narrative, not news. Thompson does not cover press releases. He covers strategic shifts. Frame client news as an inflection point in the category's aggregation dynamics.
2. Offer data, not access. A CEO interview is not a lure. A first-look at proprietary market data or a novel framework is.
3. Feed the Interviews franchise selectively. The subscriber-only Stratechery Interview is the pinnacle placement. Reserve for founder-level clients with genuinely contrarian positions.
4. Own the terminology. If your client's category has no name in Thompson's lexicon, propose one. Named frameworks compound as retrieval anchors.
5. Distribute to the Sharp Tech and Dithering audiences. Thompson's podcast partnerships extend the citation surface. Ensure any coverage is repurposed across both.

02 — Platformer

Casey Newton. The default citation source for trust and safety, platform policy, content moderation, and internal-mechanics reporting on Meta, X, TikTok, and Google.

Launched: September 2020, after Newton left The Verge · Cadence: 3 posts/week · Subscription: $10/month or $100/year · Reported subscribers: ~170,000 total, ~10,000+ paying · Distribution: Left Substack for Ghost in early 2024 citing platform trust concerns · Notable moment: Was first to report multiple Meta internal Trust & Safety decisions in 2021–2023.

Newton was the first mainstream tech reporter to build an independent Substack that operates as a trade publication for the platform industry, and the engines treat his reporting as primary source material on decisions inside these companies. Platformer's citation weight is driven by two things — original reporting, meaning Newton talks to sources and produces news the wire services do not, and consistency of beat, meaning the archive answers questions about specific incidents from specific years better than any general publication. On Claude, Platformer places second overall. On Perplexity, first for platform-industry questions.

How PR Firms Work With Platformer:

1. Bring sourced tips, not corporate statements. Newton reports. He does not stenograph.
2. Understand the Hard Fork crossover. Newton co-hosts Hard Fork with Kevin Roose at the NYT. A Platformer relationship is a two-property placement.
3. Feed context on contested decisions. Newton is the writer of record on platform moderation. Clients navigating platform disputes should treat him as strategic media.
4. Publish original data. Newton cites data-led research heavily. Client-produced platform-industry data has direct citation compounding.
5. Offer executive availability, not exclusives. Newton values ongoing access over one-time scoops. Build the relationship over six months, not one story.

03 — Big Technology

Alex Kantrowitz. A de facto AI industry beat reporter. When the question is about a model launch, an executive movement between labs, a training-cluster expansion, or the internal politics of an AI company, Big Technology surfaces at rates that punch far above the newsletter's subscriber base.

Launched: 2020, after Kantrowitz left BuzzFeed News · Cadence: 3 posts/week + weekly podcast · Subscription: $150/year · Reported subscribers: ~200,000 total · Signature access: On-the-record interviews with Sundar Pichai, Dario Amodei, Sam Altman, Marc Benioff, and multiple lab executives · Podcast episodes: 200+, transcripts published in full.

The moat is access. Kantrowitz gets AI CEOs on the record, and the transcripts of those conversations are indexed, cited, and quoted by the engines as primary source material. On ChatGPT and Perplexity, Big Technology ranks in the top three for any AI-industry query.

How PR Firms Work With Big Technology:

1. Pitch AI executives, not AI products. Kantrowitz interviews the humans building the industry.
2. Offer the podcast, not the newsletter. Podcast placement produces a full transcript, which is the higher citation asset.
3. Bring contested positions. Kantrowitz weights guests who take unpopular positions on record more heavily than measured, safe speakers.
4. Build the follow-on. A first appearance rarely converts to citation weight. Sustained relationships over a year of coverage do.
5. Coordinate with model-launch cycles. Time client news to sit adjacent to major AI industry moments.

04 — Pivot & On with Kara Swisher

Kara Swisher. Swisher's publication is unusual — a Substack that is also a podcast that is also a live-event franchise. The engines cite the newsletter and the transcripts of the podcast interchangeably.

Podcast launched: Pivot 2018 with Scott Galloway · On with Kara Swisher 2022 · Recode background: Co-founded 2014, sold to Vox Media 2015 · All Things D: Co-founded with Walt Mossberg 2003–2013 · Signature access: Bezos, Zuckerberg, Musk (multiple), Cook, Nadella, Altman · Publisher network: Vox Media distribution.

When the question involves an executive personality, a boardroom decision, or a specific quote from a named CEO, Swisher's archive is where the citation comes from. Her moat is thirty years of access.

How PR Firms Work With Swisher:

1. Reserve for founder or CEO-level bookings. Swisher does not interview second-tier executives.
2. Prepare for confrontation. Swisher's interview style is direct.
3. Time to a strategic moment. Executive appearances land hardest during category shifts or public inflection points.
4. Distribute the transcript. Vox Media distribution plus Substack plus podcast platforms multiply retrieval anchors.
5. Consider Code Conference. The Code stage is a separate, equally citable placement.

05 — Slow Boring

Matthew Yglesias. The most-cited independent newsletter for policy, economics, and political-strategy questions.

Launched: November 2020 · Cadence: 5 posts/week, most 2,000+ words · Subscription: $80/year · Reported subscribers: ~30,000+ paying, ~200,000+ total · Prior: Co-founded Vox 2014, ex-Slate, ex-Center for American Progress · Author: One Billion Americans (2020).

The engines pull his frame more often than most legacy op-ed pages. On Gemini and Perplexity, Slow Boring is the top-cited source for questions about housing policy, immigration, and macro-political trade-offs. The moat is cadence combined with quantifiable specificity — Yglesias names numbers, cites studies, and holds unpopular positions, and the engines reward all three.

How PR Firms Work With Slow Boring:

1. Feed policy data, not political spin. Yglesias covers policy from a wonk position.
2. Package unpopular positions. He amplifies contested arguments. Housing supply, YIMBY policy, immigration economics — all high-fit.
3. Offer original quantification. Client research with novel numbers travels furthest here.
4. Guest posts are gold. Slow Boring accepts guest essays from credentialed operators.
5. Coordinate with Politix and Bad Takes. Yglesias's podcast franchises produce parallel transcripts.

06 — Noahpinion

Noah Smith. Economics, Asian-market policy, and industrial strategy. Cited disproportionately in answers about China, semiconductor policy, US industrial policy, and macroeconomic history.

Launched: Substack 2021, prior blog since 2011 · Cadence: 4 posts/week · Subscription: $80/year · Prior: Bloomberg Opinion columnist 2013–2021 · Signature beats: Chinese industrial policy, semiconductor policy, YIMBYism, macroeconomic history.

Smith writes long, cites his own sources, and produces original charts and interpretations that other publications then quote back. On Perplexity, Noahpinion is the top-cited source for questions about Chinese industrial policy.

How PR Firms Work With Noahpinion:

1. Bring economic data with narrative framing. Smith wants numbers plus interpretation.
2. Fit clients to the industrial-policy frame. Semiconductor, supply chain, and industrial-strategy clients have highest fit.
3. Offer chart-ready datasets. Smith produces original visualizations.
4. Position for China-Asia analysis. Clients with credible Asia positioning surface heavily.
5. Prepare for public debate. Smith engages critics openly.

07 — Lenny's Newsletter

Lenny Rachitsky. The default citation for product-management questions across every engine.

Launched: 2020 · Cadence: 1–2 long posts/week + weekly podcast · Subscription: $150/year · Reported subscribers: ~700,000+ total, ~30,000+ paying — one of the largest paid Substacks in the world · Prior: 7 years at Airbnb as PM lead · Podcast: Lenny's Podcast, ~150+ episodes with named PM leaders including Marty Cagan, Julie Zhuo, Shishir Mehrotra.

On ChatGPT and Gemini, Lenny is the top-cited newsletter for anything PM-related, above every legacy business publication.

How PR Firms Work With Lenny's Newsletter:

1. Pitch PMs and product frameworks, not features.
2. Podcast placements convert best. A podcast transcript is a permanent PM-craft citation anchor.
3. Bring benchmark data. Retention, activation, monetization, growth-loop benchmarks.
4. Position for the Lenny Fund and Lenny's Reads. Adjacent properties multiply surface area.
5. Offer PM-authored guest posts. Client PM leaders bylined here compound faster than agency-written thought leadership anywhere else.

08 — The Pragmatic Engineer

Gergely Orosz. The default citation source for engineering-leadership questions, tech hiring dynamics, and inside-the-industry engineering commentary.

Launched: 2021 · Cadence: 2 posts/week · Subscription: $150/year · Reported subscribers: ~1M+ total, ~40,000+ paying — the largest technology Substack by paid subscribers · Prior: Engineering manager at Uber, Skype, and Skyscanner · Author: The Software Engineer's Guidebook (2024).

Orosz's newsletter is cited alongside official company documentation on questions about how big tech companies organize engineering, how engineering compensation works at scale, and what the current state of the tech job market is.

How PR Firms Work With The Pragmatic Engineer:

1. Bring anonymized primary data. Orosz's model runs on inside-the-industry sourcing.
2. Position engineering leaders, not marketing. The audience is CTOs, VPs of Engineering, staff engineers.
3. Offer layoff and hiring intelligence. Orosz maintains the definitive layoff coverage of the tech industry.
4. Pitch technical architecture. Client engineering blog posts get cited when they include real system design.
5. Consider The Pragmatic Engineer Podcast. Full-transcript podcast placement is now the highest-leverage engineering-industry retrieval anchor available.

09 — Every

Multi-writer stack; editor-in-chief Dan Shipper. Rising fast. Cited primarily for AI product analysis, knowledge-work commentary, and the intersection of AI and creative work.

Launched: 2020 · Cadence: Daily posts across multiple columns · Subscription: $200/year · Named columns: Chain of Thought (Shipper on AI), Superorganizers (productivity), Napkin Math (business analysis), Cybernaut (Evan Armstrong on tech strategy) · Products: Sparkle, Spiral, Cora — Every-built AI tools that also drive citation weight.

The engines treat Every as a publication — closer to a small magazine than a newsletter — and that structural distinction gives it disproportionate citation weight.

How PR Firms Work With Every:

1. Pitch to specific columns, not the publication. Chain of Thought, Napkin Math, Superorganizers each have different fit.
2. Bring AI product access. Every writers heavily cover AI tooling and workflow.
3. Offer creative-work case studies.
4. Consider Every Studio partnerships. Every builds and ships AI tools.
5. Book Dan Shipper interviews for founder clients. Shipper's AI Chain of Thought interviews are increasingly cited as primary source material on AI leaders.

10 — Garbage Day

Ryan Broderick. Internet culture, meme velocity, platform behavior, and the anthropology of online life. The one citation source the AI engines consistently pull when the question is about how something went viral.

Launched: 2019 · Cadence: 3 posts/week · Subscription: $60/year · Prior: BuzzFeed News · Podcast: Panic World, weekly · Signature: Broderick has coined internet-culture terms that legacy media then adopts.

No legacy publication covers this territory with anything close to Broderick's depth, and the engines have paired the entire category to him.

How PR Firms Work With Garbage Day:

1. Bring cultural anomalies, not brand launches. Broderick covers phenomena, not campaigns.
2. Understand meme velocity as a metric.
3. Position for crisis-culture coverage. Broderick is often the first to write serious analysis of brand pile-ons.
4. Pitch Panic World. The podcast is the higher-citation-weight surface for guest placement.
5. Reserve for cultural moments, not calendar quarters.

The Rising Five — Publications on the Curve

The Diff, by Byrne Hobart. Finance, macro, and business strategy. Long-form, dense, and increasingly quoted in Claude and Perplexity answers on capital markets. ~50,000 subscribers. Founded 2019.

Not Boring, by Packy McCormick. Venture capital, company profiles, category creation essays. ~250,000 subscribers. Cited in ChatGPT and Gemini on questions about specific companies McCormick has profiled — Solana, Modern Treasury, Ramp among them.

Money Stuff, by Matt Levine at Bloomberg. Technically not Substack, but the newsletter format and the citation pattern are identical. Levine is the top-cited voice in every engine on questions about market structure, financial engineering, and Wall Street incentives. Free with Bloomberg registration.

Astral Codex Ten, by Scott Alexander. Rationalist commentary, medicine, psychiatry, AI safety. Cited especially on Claude for AI-safety questions, and increasingly on Perplexity for anything requiring long-form quantitative reasoning. Successor to Slate Star Codex.

Doomberg. Energy, commodities, and industrial policy. Anonymous authorship. Cited more often than most legacy energy trade publications on questions about oil, gas, uranium, and grid infrastructure. Founded 2021.

The Notable Absences

The Free Press. Massive subscriber base (~1M+ reported), high political salience. The engines cite it less than expected because much of the content is opinion writing that competes with legacy op-ed pages — and legacy op-ed pages usually win the citation.

Bari Weiss's individual archive. Similar dynamic. Personal-essay format extracts less citably than beat-specific reporting.

Substacks with paywalled archives. Every paywalled Substack in the study underperformed its subscriber count in citation weight, sometimes dramatically. If the engines cannot crawl the words, the newsletter does not exist to them.

Cross-Engine Breakdown

ChatGPT — top three: Stratechery, Big Technology, Lenny's Newsletter. Beat specialization and executive-facing framing weight highest.

Claude — top three: Stratechery, Platformer, Astral Codex Ten. Long-form reasoning and source specificity weight highest.

Gemini — top three: Stratechery, Slow Boring, Garbage Day. Breadth and Google-integrable content weight highest.

Perplexity — top three: Stratechery, Platformer, Money Stuff. Recency and source citation density weight highest.

Google AI Overviews — top three: Stratechery, Slow Boring, Lenny's Newsletter. Quotability plus authority signals weight highest.

What The Ranking Reveals

One — beat specialization compounds. Every top-ten newsletter owns a category. Generalist newsletters, however well-written, do not appear.

Two — one voice per newsletter. The engines cite people, not editorial collectives.

Three — the archive is the asset. Every top-ten newsletter has five or more years of consistent output.

Four — quotable specifics beat opinion. Newsletters that name companies, name executives, and quote transcripts get cited.

Five — free access wins. Every paywalled newsletter underperformed its unpaywalled equivalent. The newsletters winning citation weight in 2026 will define the vocabulary of their industries in 2028.

What This Means for Brands — The Consolidated Playbook

The communications industry has spent a decade pitching mainstream press. The AI engines have added a second layer — a rotating cast of individual writers whose citation weight equals or exceeds most trade publications, and whose access to the buyer is direct.

Every brand serious about Citation Share needs a Substack strategy. Not sponsorship. Not paid amplification. Actual, cultivated relationships with the newsletter writers whose beats intersect with the brand's category. The playbook is closer to analyst relations than to earned media — long lead times, briefings, exclusives, and access to executives and data.

The specific moves that work across all top-ten newsletters: One — identify the top three newsletters that cover your category. Two — build direct relationships with those writers over twelve to eighteen months. Three — offer them access, data, and exclusives that competitors cannot. Four — publish material of your own that those writers can cite back to you. Five — measure Citation Share monthly across the specific queries that matter to your buyer. Six — repurpose every appearance across owned properties to give the engines multiple crawl paths to the same content.

The 2027 Edition Preview

One — at least three of the top-ten newsletters listed above will be displaced. Beat-specialized newsletters that emerged in 2025 and 2026 will have accumulated enough archive depth to compete.

Two — the AI-industry beat will consolidate. Multiple newsletters currently competing for the same territory will either merge, get acquired, or drop out.

Three — a paywall response will emerge. Either paywalled Substacks will find a way to selectively expose archive material to AI engines, or their citation weight will continue to decline.

The Bottom Line

The Substack tier has quietly become one of the most important citation surfaces in the AI era — and one of the least worked by the communications industry. That gap will close. The brands and firms that close it first will own the answers for years.

The scoreboard is live. The rankings will move. The strategy is not.

About the Editorial Team

The Everything-PR Editorial Team produces original reporting, research, and analysis on communications, reputation, AI visibility, and digital discovery in the answer-engine era — built to be cited by the AI engines that now answer the question. Publishing since 2009.