Web Search Portals and All Other Information Services (NAICS 519290)

A Histometrics industry primer for public- and private-market investors. NAICS = North American Industry Classification System, the U.S. government's standard for grouping businesses by activity.

1. Overview

This industry is built around one deceptively small-sounding activity — running the websites that let people search the internet — plus a grab-bag of "all other information services" that don't fit anywhere else (news-clipping and syndication shops, telephone-based recorded-information lines, and stock-photo agencies) [2].

The economic engine is simple to state and extraordinarily profitable to run: attract users, organize the world's information, capture the moment of commercial intent, and monetize it — mostly through advertising, and secondarily through subscriptions, licensing, and data services. In practice the search-portal piece is one of the most concentrated activities in the entire U.S. economy. When someone types a purchase-intent query ("cheap flights to Denver," "personal-injury lawyer near me"), a search engine auctions that click to advertisers in real time — a near-zero-marginal-cost, very-high-margin business.

What makes 2025–2026 the first genuine inflection in two decades is artificial intelligence (AI). AI "answer engines" — ChatGPT, Perplexity, Claude, and Google's own AI Overviews — increasingly answer questions directly, without the ten blue links [24][25][26]. At the same time, a U.S. federal court has declared Google an illegal monopolist in general search and ruled on remedies [9]. For the first time, the core mechanism of the industry is under real pressure.

Exposure runs two ways. In public markets there is no clean pure-play search company: the activity sits inside diversified giants (Google's parent and Microsoft), with the only true pure-plays being small visual-content licensors. In private markets, the most consequential new entrants — the AI answer engines and the privacy-focused search challengers — are almost all still private. Sections 4 and 10 cover the specific ways in.

2. What it is and how it's structured

In scope (NAICS 519290): establishments that operate websites using a search engine to build and maintain searchable databases of internet addresses and content ("web search portals"), plus other information services not classified elsewhere — the official illustrative examples are news clipping/syndication, telephone-based recorded-information services, and stock-photo agencies [2].

The business models inside the code span a wide range:

• Large consumer portals: Google, Bing, Yahoo, DuckDuckGo, Brave, Ecosia.
• Niche information providers: clipping services, specialized databases, stock-photo marketplaces, non-real-estate title-search shops.
• AI search / answer engines: products that synthesize web information rather than returning ranked links.

This is a 2022-vintage code. Under the older NAICS 2017 system, search portals lived inside 519130 ("Internet Publishing and Broadcasting and Web Search Portals"). The 2022 revision dismantled that grouping, re-sorting the Information sector by what the content is and how the business makes money rather than whether it's delivered online [3]. Search portals were carved out into 519290; publishing and broadcasting moved elsewhere.

What it explicitly EXCLUDES — this matters, because most of "Big Tech" is not in this code [2]:

• Social networks and media streaming (Meta/Facebook, YouTube, Netflix) → 516210 (Media Streaming, Social Networks, and Other Media Networks and Content Providers).
• Software publishers (apps, operating systems) → 513210; other internet/owned-content publishing → NAICS 513.
• Web hosting, data processing, and cloud infrastructure (Amazon Web Services, Google Cloud, Microsoft Azure) → 518210.
• Internet access and telecommunications carriers → 517.
• Libraries and archives → 519210.

So a single company like Alphabet is split across codes: Google Search is a 519290 activity, YouTube is 516210, Google Cloud is 518210, and Android/Chrome as software is 513210. Only the search-portal slice is counted here — a fact that drives the measurement quirks in Section 3.

Ownership mix: bar-belled. On one end, a handful of trillion-dollar public technology companies operate the largest consumer gateways. On the other, a long tail of small, mostly private "other information services" firms and private-equity-owned portals. There is very little in the middle. Federal concentration data describe revenue concentration, not the public/private split.

3. How big it is

Core figures below are our ingested federal statistics — authoritative for Census and Small Business Administration (SBA) data [1]. Dollar values are converted from the thousands in which the underlying files report them.

The Herfindahl-Hirschman Index (HHI) — the standard single-number concentration score — is suppressed in the federal file for this code, so no value is stated here [1].

Two figures tell the story. CR4 of 89.6% means the four largest firms collect roughly nine of every ten dollars of revenue — one of the most concentrated employer industries the Census measures; a CR4 near 90% signals extreme dominance even without the HHI [1]. And revenue of ~$144 billion against only ~9,253 employees works out to about $15.6 million of revenue per employee — an impossible figure for any real operating workforce.

The undercount / misclassification caveat (important). That per-employee figure is a statistical artifact, and there are two reasons the labor totals understate reality:

1. Establishment-level assignment. The Economic Census assigns each establishment to one industry by its primary activity. The establishments that book search-advertising revenue land in 519290 — but the vast workforce that actually builds and runs search (engineers, salespeople, operations) is classified under the parent companies' other establishments (software publishing, data processing, research). The revenue is real, but it is effectively Google's revenue wearing an industry label, while the headcount is a rounding error.
2. CBP coverage. County Business Patterns counts only employer establishments with paid staff. It excludes the self-employed, non-employer businesses, firms without an employer identification number (EIN), and most government operations [4] — a real limitation for a category full of tiny information shops and independent contributors.

Read the receipts as economically meaningful and the headcount as an undercount.

For scale outside the Census frame: U.S. search-advertising revenue reached about $114 billion in 2025, inside a total U.S. digital-ad market of roughly $294.6 billion (up ~13.9% year over year), per the Interactive Advertising Bureau (IAB) and PricewaterhouseCoopers (PwC) [6]. Alphabet's global "Google Search & other" revenue line alone was $224.5 billion in 2025 [7] — larger than the entire U.S. search-ad market, because it is worldwide and spans more than the domestic 519290 slice. The federal data provide no industry-wide time series for profit margins, query volume, input costs, capital spending, or private-company valuations, so those are not estimated here.

4. The investable universe

NAICS is establishment-based, while public companies report consolidated segments — so the names below are economic exposures, not one-to-one matches to the code. There is no pure-play public search company; the only clean public plays in this NAICS bucket are small visual-content licensors.

Thomson Reuters (TRI) is a further adjacent exposure — news/wire-service and professional-information licensing inside a broader legal-and-tax data company.

Major private operators and owners (where the real disruption and much of the challenger activity sits):

• OpenAI, Perplexity, Anthropic — AI "answer engines" (ChatGPT Search, Perplexity, Claude web search) competing directly for information-seeking activity [24][25][26]. Accessible mainly through late-stage venture and pre-IPO secondary markets.
• Yahoo — Apollo Funds completed its acquisition in 2021; Verizon retained a ~10% stake at closing. No later public cap table is available, so treat the precise split as historical [20].
• DuckDuckGo — independent, majority-owned by founder Gabriel Weinberg and team [21].
• Brave Software — privately held operator of Brave Search and the Brave browser [22].
• Ecosia — steward-owned; its structure bars outside owners from extracting profits or selling the company [23].

Takeaway. Buying Alphabet or Microsoft gives you search inside a bundle of cloud, ads, hardware, and software — you cannot isolate the search economics or value it on a standalone multiple. The only pure public plays in this NAICS code are the visual-content licensors (Getty, Shutterstock), a tiny fraction of the code's revenue. The pure search disruption is happening in private hands.

5. How the money works

Two very different economic engines share this code.

Engine 1 — the search-advertising auction (the money). A search engine gives the product away free and monetizes user intent:

1. The operator crawls and indexes the web.
2. Users submit queries.
3. Advertisers bid in a continuous, real-time auction for placement against specific keywords. Payment is usually cost-per-click (CPC) — the advertiser pays only when a user clicks [13].
4. Placement isn't just highest bid. Each ad gets an Ad Rank ≈ bid × Quality Score (a relevance/click-through measure), and the winner typically pays just above the next competitor's effective bid — a second-price-style outcome, not their own maximum [13].

The unit economics: near-zero marginal cost per query against enormous fixed costs (crawling, indexing, data centers). Once built, incremental ad revenue is almost pure gross margin, which is why search operating margins run far above most industries. The levers owners watch: query volume, coverage / ad load (how many queries show ads), CPC, click-through rate (CTR), and revenue per search. A large offsetting cost is traffic acquisition cost (TAC) — payments to browsers, device makers, and partners to be the default or to send traffic.

Recent reported figures put numbers on the machine. Alphabet reported paid clicks up 6% and aggregate CPC up 7% in 2025, with TAC of $59.9 billion [7]. Microsoft reported search-and-news-advertising revenue up 13% (up ~20% excluding TAC), driven by higher search volume and revenue per search [8]. A separate cross-advertiser benchmark put the average Google Ads CPC near $5.26 in 2025, up ~13% year over year — a different measure (a sample of advertiser accounts, not Alphabet's blended network CPC), but pointing the same direction: price per click is rising [12].

A subtle 2025 dynamic ties this to AI: when an AI-generated answer satisfies a query without a click ("zero-click"), the pool of monetizable clicks shrinks, so the same advertisers compete for fewer clicks — pushing CPC up even as volume softens [12][14]. Higher price partly offsets fewer clicks, for now.

Engine 2 — content licensing and subscriptions (the "other information services"). Stock-photo agencies (Getty, Shutterstock) and news/wire syndicators make money by licensing content — per-image, per-article, subscription, or archive deals — and professional databases sell recurring access. The newest, fast-growing revenue line is AI licensing: paying-per-use deals in which AI companies license text and images to train and cite their models. Shutterstock books a growing "Data, Distribution & Services" line largely tied to AI-data licensing [16], and wire services such as the Associated Press (AP) and Reuters have signed content-licensing deals with Google, Microsoft, Meta, and OpenAI [27].

This is not a manufacturing-style capacity-utilization business. The binding constraints are the cost and quality of crawling, indexing, distribution, data centers, bandwidth, content rights, and — increasingly — AI inference (the compute cost of generating each answer). High-intent commercial search tends to be more resilient in a downturn than broad brand advertising.

6. What drives demand

Reported facts:

• Digital-ad budgets. U.S. digital advertising reached ~$294.6 billion in 2025, up ~13.9%, per IAB/PwC [6]; search specifically grew but decelerated versus the prior year. Search revenue tracks the broader ad economy, which is cyclical.
• Purchase intent and e-commerce. Search monetizes best on commercial queries (travel, insurance, legal, local services); more online shopping means more valuable clicks.
• Query volume and usage share. StatCounter's June 2026 U.S. web-usage estimates were Google 86.7%, Bing 8.7%, Yahoo 2.6%, DuckDuckGo 1.5% — usage estimates, not revenue or query-share [10]. Alphabet attributed 2025 search growth partly to more queries from user adoption and mobile usage [7].
• The AI-answer shift (the swing factor). Pew Research Center found 65% of U.S. adults at least sometimes encountered AI summaries in search results, including 45% who saw them often [11]. Third-party trackers estimate AI assistants now capture a meaningful and rising share of informational query volume, with non-branded informational traffic to content sites down materially — the first structural demand shift away from link-based search in the industry's history (these AI-referral estimates are directional, not audited) [14].
• Content scarcity and AI licensing. For the "other information services" side, a genuinely new demand source is AI companies needing licensed, rights-clean text and images — a market that barely existed before 2023 [16][27].

Forward-looking judgment: demand should keep following four themes — commercial search, mobile/local discovery, professional information, and AI-assisted research. The open question is whether AI expands total information consumption and monetization or answers queries without sending users to monetized links.

7. Regulation

• Antitrust — the defining event. A federal court ruled in August 2024 that Google unlawfully maintained monopolies in general search and general search-text advertising under Section 2 of the Sherman Act. In the remedies phase, the court declined to order a structural breakup (it did not force a sale of Chrome or Android) and instead imposed behavioral remedies. The December 2025 final judgment bars certain exclusive default-placement contracts and requires Google to make search results/data available to qualified competitors through a syndication application programming interface (API) for a five-year term, subject to privacy safeguards [9]. Reported fact: the breakup was avoided. Forward-looking: the durability and competitive bite of the data-sharing and anti-exclusivity remedies remain the central open question, and appeals may continue. (Google separately lost a distinct ad-tech monopoly case.)
• Privacy — a state-led patchwork. There is no comprehensive federal consumer-privacy statute; instead a growing set of state laws (California, Virginia, Colorado, Texas, and more) create consumer rights and compliance duties, several requiring businesses to honor browser opt-out signals [31]. The California Consumer Privacy Act (CCPA) is the anchor regime [29]. Because search advertising depends on user data and targeting, tighter rules raise compliance cost and can dampen ad performance.
• Children's privacy. The Federal Trade Commission (FTC) enforces the Children's Online Privacy Protection Act (COPPA), which generally requires notice and verifiable parental consent before covered services collect personal information from children under 13 [28].
• Copyright and AI content. Copyright law remains unsettled around web crawling, training data, AI-generated outputs, and compensation for publishers — directly shaping the "other information services" licensing market and AI-answer product design [30][27].

Regulation cuts both ways: it can reduce incumbent advantages and shift traffic to challengers, or it can raise compliance costs and restrict the data and advertising practices that fund the industry.

8. Competitive dynamics and consolidation

• A near-monopoly, challenged from a new direction. By usage, Google holds roughly 87% of U.S. search, with Bing, Yahoo, and DuckDuckGo splitting most of the rest [10]. That structure has been stable for 15+ years and survived every prior "Google killer." The credible threat now is not another search box but a different product — conversational AI answer engines that bypass the results page entirely [11][14].
• Scale effects reinforce incumbency. More users generate more behavioral feedback; more data improves ranking, personalization, and ad targeting; more advertiser demand deepens auction liquidity; and browser, device, operating-system, cloud, and ad relationships compound distribution. The federal concentration figures — CR4 of 89.6% and CR50 of 98.8% — quantify how top-heavy the employer-business category is, though they do not name firms or equal search-query share [1].
• Distribution is the moat. Google's dominance has rested on being the default on browsers and phones — precisely the practice the antitrust remedy targets [9]. Whoever controls the default entry point to information controls the economics.
• AI-answer fragmentation. The AI-assistant market is itself competitive and shifting fast, with ChatGPT, Google's Gemini, Anthropic's Claude, and Perplexity all contesting share [14][24][25][26]. Google's counter is AI Overviews / AI Mode, folding AI answers into its own dominant surface to defend the franchise with distribution.
• Consolidation, in both directions. Private equity acquires and cost-optimizes mature portals (Apollo/Yahoo) [20]; acquisition platforms combine legacy internet properties (Bending Spoons/AOL) [18]. In stock imagery, the proposed Getty–Shutterstock "merger of equals" (announced January 2025) drew competition review and was ultimately abandoned in July 2026 — leaving the two largest agencies independent as they face both AI-generated-image competition and AI-licensing opportunity [17].

Forward-looking judgment: entry is easiest in a narrow niche (privacy, enterprise/professional search, a trusted vertical) but very hard in general search, where distribution, data, brand, and infrastructure advantages compound.

9. Principal risks

• AI disintermediation. The core structural risk: if users get answers without clicking, the click-auction economics erode. This is a real, measurable shift already underway, not a hypothetical [11][14].
• AI cost inflation. Richer AI answers require substantially more compute, data-center capacity, and licensed content — raising cost per answer even as they may cannibalize ad clicks.
• Antitrust overhang. Remedies could weaken Google's default-distribution advantage; appeals, the separate ad-tech case, and international regulators (notably the EU) add uncertainty [9].
• Concentration risk for the code itself. Because ~90% of the code's revenue is one company, any "search industry" exposure is really a bet on Google — and on a business Google is voluntarily cannibalizing with its own AI [1][7].
• Advertising cyclicality. Search revenue is tied to ad budgets and consumer spending; a downturn, or a shift toward retail-media and social platforms, compresses CPC and budgets quickly.
• Copyright and content-rights cost. Search and AI products face licensing costs, litigation, and possible restrictions on crawling and summarization [30][27].
• Privacy and security. Tighter tracking rules, breaches, or child-safety failures can lower ad effectiveness and create fines and reputational damage [28][29][31].
• Dependency and opacity. Smaller providers often depend on external indexes, APIs, browser distribution, or cloud infrastructure; private operators disclose little about revenue, margins, and ownership terms.

10. How to invest, and the outlook

Public routes.

• Diversified giants: Alphabet (GOOGL/GOOG) is the only way to own the dominant search franchise; Microsoft (MSFT) offers Bing exposure bundled with far larger cloud and software businesses. In both, search is a segment, so valuation multiples (price-to-earnings and the like) reflect the whole company, not search alone — do not treat either as a pure search bet.
• Pure-play "other information services": Getty Images (GETY) and Shutterstock (SSTK) are the clean, small-cap public plays in this NAICS code — each roughly $1 billion in revenue, and now independent after their merger fell through. The thesis is visual-content plus AI-data licensing, against AI-generated-image substitution.
• Adjacent information services: RELX (LexisNexis) and Thomson Reuters (TRI) offer subscription-heavy professional-information and AI-licensing exposure inside broader data companies; Bending Spoons (BSP) offers legacy-portal exposure via AOL.

Private routes.

• The frontier is private: OpenAI, Anthropic, and Perplexity are the AI-answer challengers most likely to reshape search economics, reachable mainly through late-stage venture and pre-IPO secondaries [24][25][26]. DuckDuckGo, Brave, Ecosia, and Apollo-owned Yahoo are the traditional-search private holdings [20][21][22][23]. Private investors are, in effect, buying the disruptors public investors can't yet touch directly.
• Diligence focus for private deals: monetized searches, revenue per search, advertiser retention, TAC, API/index dependence, content-rights ownership, subscription churn, AI cost per answer, gross margin, and regulatory exposure — underwrite unit economics and data rights, not user growth alone.

Near-term drivers to watch (forward-looking).

1. Does AI cannibalize or extend the auction? Whether Google can monetize AI Overviews/AI Mode at rates comparable to classic search ads is the single biggest swing factor for the whole profit pool [11][14].
2. Antitrust remedy enforcement. How aggressively the anti-default and syndication/data-sharing remedies are enforced determines whether a real second search competitor can emerge [9].
3. AI-licensing revenue ramp. For the pure-plays, the pace at which AI-content licensing scales — and whether courts require AI firms to license rather than train freely — sets the ceiling [16][27][30].
4. Ad-cycle and CPC trajectory. With volume softening but CPC rising, watch whether price gains keep offsetting click erosion [6][12].

Bottom line. On paper a tiny industry (992 establishments, ~9,000 employees), 519290 is really a ~$144-billion, near-monopoly information-and-advertising machine hiding behind a bland statistical label [1]. For the first time in twenty years its core mechanism is under genuine threat — not from a rival search box, but from AI that answers before you click. Reported facts point to extreme concentration and a real AI-driven traffic shift; the base case is that the industry evolves rather than disappears, with value shifting from ranked links toward answers, agents, commerce, subscriptions, and licensed data. The public ways in are diversified proxies; the pure disruption is still private; and whether Google's franchise proves durable or the profit pool migrates to AI answer engines is the open question that defines this industry's next five years.

---

Sources

1. Histometrics ingested federal statistics — NAICS 519290: U.S. Census Bureau, County Business Patterns (2023) and Economic Census / Concentration of Largest Firms (2022); SBA size standard (2023). Authoritative for Census/SBA figures in this primer. https://www.census.gov/naics/?input=519290&year=2022&details=519290
2. U.S. Census Bureau. "2022 NAICS Definition: 519290 — Web Search Portals and All Other Information Services." 2022. https://www.census.gov/naics/?details=519290&input=519290&year=2022
3. U.S. Bureau of Labor Statistics. "Reconstruction of CES time series: implementing the NAICS 2022 redefinitions," Monthly Labor Review. 2024. https://www.bls.gov/opub/mlr/2024/article/reconstruction-of-ces-time-series-implementing-the-naics-2022-redefinitions.htm
4. U.S. Census Bureau. "County Business Patterns Methodology" (coverage limits: excludes non-employers, self-employed, most government). 2026. https://www.census.gov/programs-surveys/cbp/technical-documentation/methodology.html
5. U.S. Small Business Administration. "Table of Size Standards." 2023. https://www.sba.gov/document/support-table-size-standards
6. Interactive Advertising Bureau (IAB) / PricewaterhouseCoopers (PwC). "Internet Advertising Revenue Report: Full Year 2025" (U.S. digital ad revenue ~$294.6B, +13.9%; search ~$114B). 2026. https://www.iab.com/insights/internet-advertising-revenue-report-full-year-2025/
7. Alphabet Inc. Form 10-K for fiscal year ended Dec 31, 2025 ("Google Search & other" revenue $224.5B; paid clicks +6%, CPC +7%; TAC $59.9B). 2026. https://www.sec.gov/Archives/edgar/data/1652044/000165204426000018/goog-20251231.htm
8. Microsoft Corporation. "2025 Annual Report" (search & news advertising revenue $13.9B; +13%, ~+20% ex-TAC). 2025. https://www.microsoft.com/investor/reports/ar25/
9. 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). 2024–2025. https://www.justice.gov/atr/media/1421546/dl?inline=
10. StatCounter Global Stats. "Search Engine Market Share, United States of America" (June 2026 web-usage estimates). 2026. https://gs.statcounter.com/search-engine-market-share/all/United-States-of-America
11. Pew Research Center. "Americans Have Mixed Feelings About AI Summaries in Search Results" (65% at least sometimes see AI summaries). 2025. https://www.pewresearch.org/short-reads/2025/10/01/americans-have-mixed-feelings-about-ai-summaries-in-search-results/
12. WordStream / Search Engine Land. "2025 Google Ads Benchmarks" (average CPC ~$5.26, +~13% YoY) and coverage of PPC in zero-click search. 2025. https://www.wordstream.com/blog/2025-google-ads-benchmarks
13. Google Ads Help / "Pay-per-click." Auction mechanics: Ad Rank ≈ bid × Quality Score; second-price-style pricing; CPC. 2025. https://support.google.com/google-ads/answer/1704431
14. Digital Applied. "AI Search Engine Statistics 2026: Market Share Data" (directional estimates of AI referral and informational-query share). 2026. https://www.digitalapplied.com/blog/ai-search-engine-statistics-2026-market-share
15. U.S. Securities and Exchange Commission. "Getty Images Holdings, Inc. Form 10-K, FY2025" (revenue ~$981M). 2026. https://www.sec.gov/Archives/edgar/data/1898496/000162828026018160/gety-20251231.htm
16. U.S. Securities and Exchange Commission. "Shutterstock, Inc. Form 10-K, FY2025" (revenue ~$990M; Data, Distribution & Services / AI-data licensing). 2026. https://www.sec.gov/Archives/edgar/data/1549346/000154934626000018/sstk-20251231.htm
17. United Kingdom Competition and Markets Authority. "Getty Images / Shutterstock merger inquiry" (deal announced Jan 2025; abandoned July 2026). 2026. https://www.gov.uk/cma-cases/getty-images-slash-shutterstock-merger-inquiry
18. Bending Spoons. "Final Prospectus as Filed" (2026 IPO; AOL acquisition completed Jan 2, 2026). 2026. https://bendingspoons.com/documents/financials/2026/Bending%20Spoons%20Final%20Prospectus%20As%20Filed.pdf
19. RELX. "2025 Annual Report" (LexisNexis; professional/legal information). 2026. https://www.relx.com/~/media/Files/R/RELX-Group/documents/reports/annual-reports/relx-2025-annual-report.pdf
20. Apollo Global Management. "Apollo Funds Complete Acquisition of Yahoo" (Verizon retained ~10%). 2021. https://www.apollo.com/insights-news/pressreleases/2021/09/apollo-funds-complete-acquisition-of-yahoo-161530593
21. DuckDuckGo. "Who Owns DuckDuckGo?" 2026. https://duckduckgo.com/duckduckgo-help-pages/company/who-owns-duckduck-go
22. Brave Software. Company site. 2026. https://brave.com/
23. Ecosia. "About Ecosia" (steward-ownership structure). 2026. https://support.ecosia.org/article/406-about-ecosia
24. OpenAI. "Introducing ChatGPT Search." 2024. https://openai.com/index/introducing-chatgpt-search/
25. Perplexity. "What Is Perplexity?" 2026. https://www.perplexity.ai/help-center/en/articles/10352155-what-is-perplexity
26. Anthropic. "Claude web search now available globally on all plans." 2025. https://claude.com/blog/web-search
27. Digiday. "A timeline of the major deals between publishers and AI tech companies" (AP, Reuters licensing to Google/Microsoft/Meta/OpenAI). 2024–2025. https://digiday.com/media/a-timeline-of-the-major-deals-between-publishers-and-ai-tech-companies-in-2025/
28. Federal Trade Commission. "Children's Online Privacy Protection Rule (COPPA)." 2026. https://www.ftc.gov/legal-library/browse/rules/childrens-online-privacy-protection-rule-coppa
29. California Privacy Protection Agency. "California Consumer Privacy Act of 2018 (CCPA)." 2026. https://cppa.ca.gov/regulations/
30. U.S. Copyright Office. "Copyright and Artificial Intelligence." 2025. https://www.copyright.gov/ai/
31. Choozle / IAPP. "U.S. State Privacy Legislation: Key Updates and Impacts on Digital Advertising" (state privacy-law patchwork; browser opt-out signals). 2025. https://choozle.com/blog/u-s-state-privacy-legislation-key-updates-and-impacts-on-digital-advertising/