Information (U.S.) — NAICS 51

A Histometrics sector rollup primer for a general investing audience — relevant to both public-market and private investors. NAICS = North American Industry Classification System, the U.S. government's standard scheme for grouping businesses by activity. This is a two-digit sector, the broadest level in the taxonomy, sitting above six three-digit subsectors. This page synthesizes the six already-written child primers plus our ground-truth federal statistics for the 51 level; it does not re-research the sector from scratch. Its distinctive value is the contrast across the six children — how differently they are sized, owned, concentrated, and invested. For company-by-company detail, follow the child links.

---

1. Overview

NAICS 51 — Information — is the federal statistical home of the businesses that create, package, move, and monetize information rather than physical goods. It is where the U.S. government files software and streaming, telecom networks and cloud data centers, film studios and record labels, book and news publishers, broadcasters and social platforms, and the search engines and libraries that help people find it all. If there is a single "information economy" in the American statistical system, this is it — roughly $2.21 trillion of receipts and 3.8 million jobs in our ground-truth file.[1]

But "sector 51" is less a single industry than a federation of six very different ones, and that heterogeneity is the whole story at this level. Its children run from a near-monopoly (web search) to a fragmented cottage economy (independent publishers and producers); from the most heavily licensed businesses an investor can buy (spectrum-holding carriers and broadcasters) to the most lightly regulated (software as a product); from mature dividend-paying utilities (telecom) to the single fastest-growing capital cycle in the economy (AI data centers). A blended "Information sector" average tells you almost nothing. The only useful analysis is child by child, and this page is built to compare them.[2][3][4][5][6][7]

Two forces nonetheless run through all six and are worth holding in mind throughout:

• Build-once, serve-many economics. Everywhere in 51, the prize is a costly-to-create information asset — a piece of code, a music catalog or film library, a fiber network, a search index, a data center — that can then be reproduced or delivered to each additional customer at near-zero marginal cost. That gives the whole sector high operating leverage and rewards scale, recurring revenue, and ownership of durable rights or scarce infrastructure over undifferentiated capacity.
• Artificial intelligence (AI) as the universal swing factor. AI touches the economics, regulation, and risk of every child at once — a product to sell (software), the demand engine of a capital boom (cloud/data centers), a disruptor of the core mechanism (search), a cost lever and copyright question (film, music, publishing, broadcasting), and a driver of network-transport demand (telecom). It is the one force that reprices the entire sector simultaneously.[3][6][7]

---

2. What's inside — the six children and how they differ

NAICS 51 splits into six three-digit subsectors. They share the build-once logic but sit in different economic universes. Read this contrast before any sector-level number. All receipt shares are of the $2.21T sector total from our federal file.[1]

Definitions. CR4 = the combined revenue share of the four largest firms (CR8/CR20/CR50 extend that to the top 8/20/50). HHI = Herfindahl-Hirschman Index, a 0–10,000 concentration score; U.S. antitrust agencies treat below 1,000 as unconcentrated and above 1,800 as highly concentrated.[8] REIT = real-estate investment trust; ETF = exchange-traded fund; PE = private equity. Firm shares are of the 91,800-firm sector total and sum slightly above 100% because ~518 firms operate in more than one subsector (see §3).

Five contrasts worth holding onto.

1. Two children are two-thirds of the money. Telecom (30.2%) and publishing-including-software (26.7%) together are ~57% of sector receipts; add broadcasting/content and computing infrastructure and four children carry ~87%. Any headline about "the Information sector's revenue" is, to the first order, a statement about telecom networks and software.

2. Firms and dollars point in opposite directions. The two children with the most firms — publishing (~33%) and motion-picture/sound (~30%) — include vast cottage tails of tiny publishers, self-releasing authors and musicians, and micro-producers; motion-picture/sound is ~30% of firms but under 6% of receipts. Telecom, by contrast, is ~13% of firms but 30% of receipts: few, giant, capital-heavy carriers. Search is the extreme — 3% of firms, but 88 cents of every one of its revenue dollars in four companies.[5][7][2]

3. Concentration is real inside the children and invisible at the top. The sector's HHI of just 231.6 and CR4 of 24.4% make Information look almost perfectly competitive — well below telecom (1,248), broadcasting (544), or motion pictures (421), and below the CR4 of five of the six children. That is a pooling artifact, not a finding: the sector's leaders sit in different subsectors and do not compete with one another (a fiber carrier is not a rival to a record label or a search engine), so blending six industries into one $2.2T base dilutes every leader's share. Real market power — the search monopoly, the wireless oligopoly — lives at the child level and below. Never read competition off the 2-digit code.[1][8]

4. Investability inverts as you move down the list. Software (inside 513) is arguably the deepest, most liquid public-market exposure in the whole economy; telecom offers both a deep public menu and deep private access; but search's dominant asset is ownable only inside one diversified giant, its disruptors are private, and the biggest cloud, streaming, and satellite assets are booked inside parents or private companies (see §3–4). Growth and public access frequently sit at opposite ends of the same child.

5. This is a recently re-drawn sector. The 2022 NAICS revision reshaped 51 — gathering software publishing into 513 alongside content publishing, creating 516 to hold streaming and social alongside legacy broadcast, and splitting out 518 (cloud/data centers) and 519 (search/libraries). The sector's internal boundaries are new, which is one reason so much economic activity lands in an unexpected code (see the undercount caveat).[9]

Everything below covers the sector as a whole, flagging where the children pull apart.

---

3. Size (this level's rollup figures)

These are our ground-truth federal statistics for NAICS 51. Receipts, firm counts, and concentration are from the 2022 Economic Census (EC); establishments, employment, and payroll are from 2023 County Business Patterns (CBP) — different surveys, different years, counting different units, so do not mechanically reconcile them.[1]

Four things to read carefully.

Receipts roll up almost exactly; firms de-duplicate. The six children's receipts sum to $2,214.6 billion, matching the sector total to within about $4 million on $2.2 trillion — a clean rollup.[1][2][3][4][5][6][7] Firm counts sum to 92,318 against the reported 91,800 — 518 more, which is the Census avoiding double-counting: firms active in more than one subsector (for example a cable company that is both a telecom carrier and a content network, or a conglomerate that both publishes software and streams video) are counted once at the sector level but in each child below. Employment and payroll are only partially reconstructable from the children — where clean subsector CBP figures exist (telecom ~912,600; computing infrastructure ~629,500; search/other ~409,200; motion-picture/sound ~333,000) they are consistent with the level, with publishing and broadcasting making up the balance of the 3.81 million total.[2][5][6][7]

The sector looks unconcentrated, but the number is a floor on reality. An HHI of 231.6 and a CR4 of 24.4% would, on their face, describe a wide-open market. As §2 explains, that is arithmetic, not competition: pooling six industries with different leaders drags every share down, and even the top 50 firms collect only 60.8% of a $2.2T base — a measure of breadth, not health. The concentration that matters (a ~90% search monopoly, a three-carrier wireless field) is inside the children. Treat the sector HHI as a signal that Information is diverse, not that it is competitive.[1][7][5]

High revenue and pay per worker — a capital- and IP-heavy sector. On a (year-mixing, directional) basis, receipts run about $581,000 per employee and average pay about $149,000 — both well above the economy-wide average, reflecting that output here comes from code, networks, spectrum, catalogs, and machines rather than headcount, and that the workforce skews high-skill. Payroll is only ~26% of receipts.[1]

Undercount caveat — unusually large here, and it hides both scale and the growth edge. These are employer, primary-activity counts: they exclude nonemployers, the self-employed, freelancers, and government activity, and they file each firm under its single main business — which in this sector systematically misplaces the giants. The leakage, drawn from the children, is severe and one-directional against the most dynamic parts:

• The biggest cloud is largely elsewhere. Amazon Web Services rolls up under Amazon (retail, outside 51 entirely); Microsoft Azure sits in software (513) and Google Cloud in search (519), not in the computing-infrastructure child (518) that describes their activity — so 518's CR4 of 37.2% badly understates a business most people picture as a three-company oligopoly.[6]
• The biggest ad-video platform is mis-filed. YouTube's advertising (~$60 billion in 2025) is booked with Alphabet under web-search portals (519), not with broadcasting/content (516).[4][7]
• The fastest-growing telecom asset is outside the sector. SpaceX's Starlink (~$11.4 billion of connectivity revenue in 2025) is classified inside SpaceX as a space-vehicle manufacturer (336414), outside 51 altogether — larger by itself than the entire measured satellite industry.[5]
• The creative and solo tail is uncounted. Gig film crews and self-releasing musicians (independent artists, 711510), 2.6 million self-published book titles a year, solo developers and open-source work, podcasters and newsletter writers — none appear in these employer counts.[2][3]
• Public libraries are excluded. As a governmental activity, the ~$22B+ institutional library economy is almost entirely off these books.[7]

Read $2.21 trillion as the employer, primary-activity core of a materially larger information economy — one whose most disruptive dollars (hyperscale cloud, satellite broadband, ad-funded video, AI answer engines) are frequently booked outside the child, or the sector, that bears the activity's name. The federal file carries no sector-wide profit margin, cash-flow, capital-spending, or growth series; none is invented here.

---

4. Investable universe (where value concentrates across the children)

The defining fact at this level is that value, growth, and public access do not line up — they concentrate in different children, and often at opposite ends of the same one. Tickers below are for orientation only; reserve valuation for §10 and the child primers.

• Software (inside 513) is where public-market investability is deepest. An outsized share of the sector's most ownable value is here: Microsoft (MSFT), Oracle (ORCL), Salesforce (CRM), Adobe (ADBE), Intuit (INTU), ServiceNow (NOW), plus dozens of pure-play cloud, security, data, and gaming names and diversified software ETFs — with a vast venture and PE layer (Thoma Bravo, Vista, Silver Lake) beneath. Content publishing, the other half of 513, is the opposite: scarce pure-plays (NYT, News Corp, and scholarly houses RELX, Wiley, Informa) atop mostly private or foreign owners.[3]
• Telecom (517) is the most "ownable" child, but mature. Deep public menu — dividend carriers AT&T (T), Verizon (VZ), T-Mobile (TMUS); cable Comcast (CMCSA), Charter (CHTR); wholesale fiber Lumen (LUMN), Cogent (CCOI); tower REITs American Tower (AMT), Crown Castle (CCI), SBA (SBAC) — plus an unusually deep private layer (fiber-infrastructure funds, a large high-yield bond market, family-owned Cox).[5]
• Broadcasting/content (516) is a barbell. Nearly all the value and market cap sit in mega-cap streaming/social/networks — Meta, Netflix (NFLX), Disney (DIS), Comcast, Warner Bros. Discovery (WBD), Spotify (SPOT) — reachable via equity and Communication Services ETFs; the broadcast-station half is broad but shallow (Nexstar, Sinclair, Gray, Scripps; radio's leveraged iHeartMedia) with much held privately and no sector ETF.[4]
• Computing infrastructure (518) hides its giants and sells its landlords. Hyperscale cloud is embedded in diversified parents (Amazon, Microsoft, Alphabet, Oracle, IBM); the clean listed plays are the data-center REITs Equinix (EQIX) and Digital Realty (DLR), plus AI "neoclouds" and a large private/PE-owned tail reachable through listed alternative-asset managers.[6]
• Search (519) has no pure play. The franchise that is this child's economics is ownable only inside Alphabet (GOOGL/GOOG), with Microsoft a bundled secondary; the consequential disruptors — OpenAI, Anthropic, Perplexity — are private; the only listed pure-plays are trace-sized image licensors (Getty, Shutterstock).[7]
• Motion picture & sound (512) is inverted within itself. In film, the big money (production/distribution) is embedded in conglomerates and the only clean public pure-plays are theaters (Cinemark (CNK), AMC, IMAX); in music, the big money (labels and publishers) is listed through the majors (UMG, Warner Music (WMG), Sony (SONY)), while the studio tail is private.[2]

The practical read. For a public investor, Information is most accessible as software and telecom, plus mega-cap streaming and platforms, plus data-center REITs — but the sector's single largest and fastest-growing assets (AWS, Starlink, YouTube, the AI answer engines) are either buried inside diversified parents or entirely private, so "buying the sector" means buying a specific segment inside a bigger enterprise. For a private investor — including the private-credit, business-development-company (BDC), and closed-end-fund (CEF) audience — the opportunity set is far broader: fiber and tower infrastructure funds and telecom high-yield debt, data-center construction finance, software buyouts, music-royalty and film-slate funds and media asset-backed securities (ABS — bonds repaid by contracted royalty or subscription income), and pre-IPO stakes in the private platforms.

---

5. How the money works

Despite the size and ownership gaps, all six children run on the same underlying template: spend heavily to build or acquire an information asset that is expensive to create once, then earn off it repeatedly at near-zero marginal cost. High fixed cost, low incremental cost, and therefore high operating leverage — profit accrues to whoever owns the durable asset (rights, network, index, catalog, data center) and gets squeezed away from anyone who merely rents capacity or produces for hire. The sector's revenue models cluster into four archetypes, and most large firms blend several:

• Recurring subscription / access. Software-as-a-service (SaaS) renewals, streaming and music subscriptions, broadband and wireless plans, scholarly-journal and data subscriptions, cloud committed-use contracts. Sticky, cash-generative, judged on retention and net revenue retention. This is the direction every child is moving — away from one-time transactions toward revenue that repeats on its own.[3][4][5][6]
• Advertising. Search auctions (cost-per-click), social and video impressions, broadcast spots, retransmission-adjacent ad inventory. Roughly impressions × price, very high gross margin, cyclical, and increasingly AI-targeted.[7][4]
• Licensing of owned rights. Film libraries and music catalogs (masters and compositions), news and image syndication, and a fast-rising line of AI-training/citation licensing — recurring, annuity-like, and often traded as a financial asset priced on a multiple of royalty income.[2][3][7]
• Usage / capacity rental. Cloud compute and storage (priced in energized megawatts, not floor space), telecom data, colocation. Capital-intensive and front-loaded: operators spend billions before revenue arrives, then earn a spread over the cost of the plant.[6][5]

The metrics that fit this sector are enterprise-value/EBITDA (earnings before interest, taxes, depreciation and amortization), free-cash-flow (FCF) yield, recurring-revenue and retention measures, average revenue per user (ARPU) and churn, and catalog/library/backlog quality — not box-office, subscriber, or download headlines (which are not operator profit) and not regulated-utility rate-base or REIT funds-from-operations math, which apply only to the specific pockets that are literally REITs (data-center and tower landlords) and nowhere else in the sector.

---

6. Demand drivers

The children answer to different customers, but the drivers rhyme, and one dominates:

• Digitization and the migration to digital (the master trend). Attention, commerce, and advertising keep moving from analog and legacy channels to digital ones — from broadcast and print to streaming and platforms, from wired voice to broadband and wireless data, from on-premises servers to the cloud. That single dynamic is a tailwind for software, cloud, streaming, and search and a headwind for broadcast, print, and legacy voice.[4][3][6]
• AI — now the biggest single driver in parts of the sector. It is the demand engine of the computing-infrastructure capital cycle, a new product line for software, a usage driver and disruptor for search, and a cost-and-licensing question for all content.[6][7][3]
• Connectivity and data consumption as utilities. Households and businesses treat broadband, mobile, and cloud as essential; U.S. wireless data traffic and streaming viewing keep climbing, supporting ARPU even where subscriber counts are flat.[5][4]
• The advertising cycle. Digital ad revenue reached ~$294.6 billion in the U.S. in 2025 (up ~13.9%), and it moves search, social, broadcast, and ad-supported streaming together — cyclical, but from a structurally growing digital base.[10]
• Content and IP volume. More film, TV, music, games, and creator content lifts licensing, sync, production, and the software and infrastructure that produce and distribute it.[2][3]
• Institutional, R&D, and public budgets (the least-cyclical pocket) — enterprise software and data spend, library and university budgets, federal rural-broadband subsidy.[3][7][5]

The shared caveat: audience and enterprise demand for information looks durable, but durable demand does not guarantee operator profitability — a handful of platforms can gain users while squeezing the prices they pay content owners, carriers, and service houses.

---

7. Regulation

Information spans nearly the full regulatory range — from the most heavily licensed businesses an investor can buy to the most lightly touched — so there is no single rulebook. Three fronts cut across the sector, and two apply only to the licensed children:

• Copyright and AI — the shared central story. Copyright is what turns a film library, music catalog, backlist, database, or codebase into an asset, and whether AI developers must pay to train on it is unsettled and bears on the value of every content-owning business in 51. Landmark actions span the sector: The New York Times Co. v. OpenAI/Microsoft, Bartz v. Anthropic (2025, with a ~$1.5 billion settlement), and the U.S. Copyright Office's ongoing AI work.[15][3][2]
• Antitrust and merger review. The Department of Justice (DOJ) and Federal Trade Commission (FTC) review deals under the 2023 Merger Guidelines,[8] and enforcement is live across the sector — most consequentially United States v. Google (2024 liability finding; December 2025 remedies judgment) in search, alongside scrutiny of platform, software, telecom, and media consolidation.[12]
• Privacy, data, and platform law. State privacy statutes, the EU General Data Protection Regulation, FTC advertising rules, Section 230 platform-liability and child-safety regimes, and the foreign-ownership law that forced the TikTok divestiture touch the platform, software, and data-heavy children.[4][3]
• Spectrum and licensing (telecom and broadcast only). Terrestrial carriers and over-the-air stations operate on public spectrum under Federal Communications Commission (FCC) license, with ownership caps, the ~$9-billion-a-year Universal Service Fund, retransmission and political-broadcasting rules, and (post-2024) no automatic judicial deference on net-neutrality questions. These are among the most regulated businesses in the sector; software and publishing, at the other extreme, need no license to operate.[5][4]
• Infrastructure permitting (computing). Land, power, and environmental permitting has become the binding constraint on data-center growth, with projects delayed or blocked and some states weighing moratoriums.[6]

Net: regulatory intensity varies more within this sector than in almost any other, but two levers — copyright/AI outcomes and antitrust — can reprice whole children at once.

---

8. Consolidation

Every child is consolidating, at different stages and for different reasons, but with one common thread: scale is defense, and buyers are paying up for scarce, hard-to-replicate assets — spectrum, orbital slots, dense fiber, secured power, deep catalogs, and defensible workflows — while outsourcing or exiting the commoditized parts.

• Telecom (517): a fiber roll-up wave (Verizon–Frontier closed; Charter–Cox pending) and a wireless field collapsed to three nationals; buyers pay for spectrum and fiber routes.[5]
• Broadcasting/content (516) and motion-picture/sound (512) overlap on one landmark deal: Paramount Skydance agreed to acquire Warner Bros. Discovery (~$110 billion enterprise value; temporarily halted by a federal judge — not a completed deal), while Comcast plans to separate NBCUniversal's cable networks and Nexstar closed its purchase of Tegna in local TV; music runs a multi-year catalog-buying wave financed increasingly with ABS.[4][2][13]
• Publishing/software (513): strategic megadeals (Broadcom–VMware) and a re-accelerating take-private wave led by software-focused PE.[3]
• Computing infrastructure (518): record data-center M&A and PE roll-ups chasing contracted, often inflation-linked cash flows.[6]
• Search (519): a stable ~90% one-firm structure that barely consolidates (the threat is a new product, AI answer engines, not another search box).[7]

The cross-cutting judgment is the same everywhere: scale does not guarantee margins. Durable operators own defensible rights, networks, or capacity sold on differentiated terms — not undifferentiated capacity at volatile prices. (A NAICS sector is not automatically an antitrust market; read the concentration figures as directional, not as a legal finding.[8])

---

9. Risks

Shared across the sector, weighted differently by child:

• AI is double-edged, everywhere. It can open new products and licensing lines or commoditize the work, cannibalize per-unit pricing, raise compute costs, and devalue libraries and content faster than costs fall — with the legal status of training on copyrighted work unresolved.[6][7][3][15]
• Platform bargaining power and concentration. A few platforms and carriers are simultaneously the biggest customers and competitors of the content owners and service houses beneath them, able to strip upside and push risk downstream.[4][7]
• Interest-rate and capital-cycle sensitivity. The capital-heavy children (telecom, data centers) carry heavy debt and re-rate with rates; music/film catalogs are valued like bonds; high-growth software is priced on far-future cash flows. Add the risk of overbuild — AI data-center capacity and overlapping satellite constellations built on debt against forecasts.[5][6][2]
• Secular decline of the legacy pockets. Over-the-air broadcast, print publishing, and legacy wired voice are in structural, not cyclical, erosion — the migration of §6 draining one bank of the river.[4][3]
• Regulatory and legal risk. Antitrust remedies (search), spectrum/ownership and USF rules (telecom/broadcast), privacy and child-safety law (platforms), copyright/AI outcomes (all content), and data-center permitting can each reprice a child.[12][5][6]
• Measurement opacity. As §3 stresses, employer-only, primary-activity federal statistics understate the sector and misplace its biggest assets (AWS, Starlink, YouTube, AI answer engines); many of the most important operators are private or embedded inside diversified parents. A recognizable brand may have little economic connection to the child it appears to belong to — always look through to the specific segment and owner.[1][6]

The sector's central tension: information demand is not shrinking — it is moving and reordering. The question for each child is whether its financial durability outlasts the shift in where attention, data, and dollars accrue.

---

10. How to invest & outlook

Match the vehicle to the child — there is no single "Information" trade.

• Public routes. The deepest, most liquid exposure is software (Microsoft, Oracle, Adobe, Intuit; ServiceNow, CrowdStrike, Snowflake; or software/cloud ETFs) and telecom (dividend carriers T, VZ, TMUS; tower REITs AMT, CCI, SBAC; wholesale fiber LUMN, CCOI). Streaming and platforms come through mega-cap equity and Communication Services ETFs (Meta, Netflix, Disney, WBD, Spotify). Data-center landlords EQIX and DLR are the clean way to own the AI capital cycle; the hyperscalers themselves come bundled inside Amazon, Microsoft, and Alphabet. Search is ownable only through Alphabet (GOOGL). Music is ownable directly through the majors (UMG, WMG, SONY); film only through conglomerates or the small theater pure-plays (CNK, AMC, IMAX). Value every one of these with sector-appropriate tools — EV/EBITDA, FCF yield, recurring-revenue and retention quality, catalog/backlog strength — not box-office, subscriber, or download headlines.
• Private routes, where most non-conglomerate and BDC/CEF capital actually works: fiber and tower infrastructure funds and telecom high-yield debt; data-center construction and equipment finance; software buyouts and growth equity; music-royalty, film-slate, and content-library funds and media ABS; and pre-IPO stakes in the private platforms and AI challengers (OpenAI, Anthropic, Perplexity; a future Starlink IPO would be a landmark repricing event). The diligence questions repeat across children: who owns the durable asset, who controls distribution, how the royalty/recoupment/utilization waterfall works, and what happens if a title, slate, network, or facility underperforms.

Outlook (a judgment, not a fact). Across all six children, value should keep migrating from legacy to digital and accruing to whoever owns the scarce, durable asset — code and defensible workflows, deep catalogs and rights, spectrum and fiber, secured power and dense data centers, the search index. AI is the reordering force: it will widen the gap between owners of hard-to-replace assets and sellers of easily copied capacity, drive the near-term capital cycle in computing infrastructure, and unsettle the search economics that have been stable for two decades. The children will move apart, not together — software, cloud, and platforms compounding; telecom and broadcast managing mature or declining cores; music and streaming growing; print and legacy voice eroding. The sector-wide questions to watch: how AI-and-copyright law resolves, how far antitrust reshapes search and media, the path of interest rates for the capital-heavy and catalog-valued businesses, and whether the AI infrastructure build is met by durable demand or overbuilt. These are judgments, not guarantees — and Information is best underwritten child by child, never off the deceptively blended 2-digit average.

For the full treatment of any child — complete company tables, unit economics, detailed regulation, consolidation, and per-metric sourcing — read the six child primers: 512 Motion Picture & Sound Recording · 513 Publishing · 516 Broadcasting & Content Providers · 517 Telecommunications · 518 Computing Infrastructure, Data & Hosting · 519 Web Search, Libraries & Other Information Services.

---

Sources

Sector-level figures are from Histometrics' ground-truth federal file for NAICS 51 (U.S. Census Bureau, 2022 Economic Census and 2023 County Business Patterns). Remaining citations are drawn from the six child primers and the third-party sources they carry.

1. Histometrics ground-truth federal file, NAICS 51 — receipts, firms, CR4/CR8/CR20/CR50 and HHI (2022 Economic Census); establishments, employment, annual and Q1 payroll (2023 County Business Patterns). U.S. Census Bureau. https://www.census.gov/naics/
2. Histometrics child primer — NAICS 512 Motion Picture and Sound Recording Industries (receipts $130.9B; CR4 35.2%; HHI 420.6; film/music contrast, company tables, and full sources). primer-512-DRAFT.md
3. Histometrics child primer — NAICS 513 Publishing Industries (receipts $592.3B; CR4 20.2%; software-vs-content contrast, SaaS economics, and full sources). primer-513-DRAFT.md
4. Histometrics child primer — NAICS 516 Broadcasting and Content Providers (receipts $345.5B; CR4 38.7%; HHI 544.2; streaming-vs-broadcast barbell and full sources). primer-516-DRAFT.md
5. Histometrics child primer — NAICS 517 Telecommunications (receipts $669.3B; CR4 64.7%; HHI 1,248.4; terrestrial/satellite/other contrast and full sources). primer-517-DRAFT.md
6. Histometrics child primer — NAICS 518 Computing Infrastructure Providers, Data Processing, Web Hosting and Related Services (receipts $329.5B; CR4 37.2%; HHI suppressed; cloud/data-center economics and full sources). primer-518-DRAFT.md
7. Histometrics child primer — NAICS 519 Web Search Portals, Libraries, Archives, and Other Information Services (receipts $147.0B; CR4 87.9%; HHI suppressed; search-vs-libraries contrast and full sources). primer-519-DRAFT.md
8. U.S. Department of Justice / Federal Trade Commission. "2023 Merger Guidelines" (HHI thresholds; a NAICS group is not automatically an antitrust market). https://www.justice.gov/atr/2023-merger-guidelines
9. U.S. Census Bureau. "2022 NAICS Manual and Definitions — sector 51 (Information) and subsectors 512, 513, 516, 517, 518, 519," including the 2022 reclassification of software publishing and the creation of subsector 516. https://www.census.gov/naics/reference_files_tools/2022_NAICS_Manual.pdf
10. Interactive Advertising Bureau and PwC. "Internet Advertising Revenue Report: Full Year 2025" ($294.6B; +13.9% year over year). https://www.iab.com/
11. Nielsen. "The Gauge — Streaming reaches 47.5% of U.S. TV viewing (December 2025)." https://www.nielsen.com/
12. U.S. Department of Justice / U.S. District Court for the District of Columbia. "United States v. Google LLC" — liability opinion (Aug. 2024) and Final Judgment (Dec. 2025). https://www.justice.gov/atr/
13. Warner Bros. Discovery / Paramount Skydance merger disclosures (~$110B enterprise value; federal-judge pause); Comcast NBCUniversal cable separation; Nexstar–Tegna close. Associated Press; U.S. SEC. https://apnews.com/
14. Variety. "YouTube Revenue for Full-Year 2025 Topped $60 Billion"; SpaceX/Starlink ~$11.4B 2025 connectivity revenue (classified in SpaceX, NAICS 336414). https://variety.com/
15. NPR / The Authors Guild. "Bartz v. Anthropic" (2025 settlement); "The New York Times Co. v. OpenAI/Microsoft"; U.S. Copyright Office, "Copyright and Artificial Intelligence." https://www.copyright.gov/ai/