Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (NAICS 51821): An Investor's Primer

1. Overview

NAICS 51821 is the federal statistical home of the cloud and data-center industry — the businesses that own the computers, run the data centers, and rent out storage, processing power, and hosting to everyone else. (NAICS is the North American Industry Classification System, the standard the U.S. government uses to sort businesses by activity.) It covers the physical and digital plumbing of the internet: renting raw servers and storage (Infrastructure as a Service, or IaaS), ready-made software environments (Platform as a Service, or PaaS), cloud storage, application and web hosting, colocation (housing a customer's servers in your building), and data processing [1].

This is one page up the taxonomy from the detailed industry primer. It exists mainly to hold this level's official statistics and to point you to the full write-up. The economics, companies, risks, and how-to-invest detail all live in the child primer — 518210.

2. What's inside — and why this level equals its one child

NAICS is a nested hierarchy: broad sectors at the top, progressively finer industries below. At the 5-digit "industry" level, code 51821 contains exactly one 6-digit national industry: 518210, which carries the same name. There is no second child to aggregate, no sibling activity split off into a separate code. The 5-digit rollup and the 6-digit leaf describe the identical set of businesses — the U.S. and the international (NAICS-wide) definitions simply line up one-to-one here.

Practically, that means every figure, company, and dynamic you would attribute to "NAICS 51821" is the same as for 518210. Rather than repeat the full industry primer, this page gives the rollup statistics for this level and hands you off. For the complete treatment — how the money works, the investable universe of public and private operators, demand drivers, regulation, consolidation, and risks — read the 518210 primer.

3. Size (this level's figures)

Because 51821 equals 518210, the ground-truth federal statistics are identical to the child's:

Average pay near $147,700 marks a high-wage, high-skill workforce, and the modest headcount (about 630,000) against enormous capital deployment reflects an industry whose output comes from machines and power, not people [2]. The Herfindahl-Hirschman Index (HHI, a finer concentration gauge) is suppressed in the federal data, so no HHI-based conclusion can be drawn.

Undercount caveat — read the concentration numbers carefully. A four-firm concentration ratio of only 37.2% looks strikingly un-concentrated for an industry most people picture as a three-company cloud oligopoly [3]. The reason: the giants are largely classified elsewhere. Federal firm-based tabulations file each company under its primary activity, so most of Amazon Web Services' revenue rolls up under Amazon (retail), Microsoft Azure under Microsoft (software publishing), and Google Cloud under Alphabet (search/internet publishing). AWS alone generated roughly $129 billion in 2025 [4] — by itself nearly 40% of this whole measured industry's $329.5 billion of 2022 receipts, yet almost none of it counts inside 51821. The federal figures therefore best describe the independent hosting, colocation, and data-processing firms — the long tail — while the hyperscalers sit in adjacent codes. Separately, County Business Patterns and the Economic Census cover only employer businesses, excluding self-employed operators and most government facilities, so small and government-run infrastructure are further understated. The supplied data does not quantify the missing share, so no adjustment is made here.

4. Investable universe — where value concentrates

Because this level has one child, the whole investable universe is that of 518210 (see the child primer for tickers, scale figures, and private owners). In brief, value concentrates in a few tiers:

• Hyperscale cloud, embedded in diversified parents (Amazon/AWS, Microsoft/Azure, Alphabet/Google Cloud, Oracle, IBM) — the largest revenue pools, but each is one segment of a bigger company classified outside this code.
• Data-center landlords structured as real estate investment trusts (REITs — companies that own income-producing property and pay out most profit as dividends), chiefly Equinix and Digital Realty.
• AI-focused "neoclouds" (e.g., CoreWeave) that rent graphics-processing-unit (GPU) capacity for AI workloads.
• Web hosting, managed hosting, and edge/data platforms — a long, fragmented tail.
• Privately held operators (QTS, CyrusOne, Vantage, Aligned, and others), reachable indirectly through listed alternative-asset managers.

Few public stocks are "pure" plays, because the biggest operators are buried inside larger parents. The child primer lays out each tier with names and approximate scale.

5. How the money works

The unifying model, detailed in the child primer, is converting capital and power into rentable computing capacity, then keeping it full at a healthy spread over its cost. REIT economics genuinely apply to the data-center landlords (Equinix and Digital Realty are REITs, so funds from operations, or FFO — cash earnings after adding back property depreciation — is their headline metric); regulated-utility rate-base and mining-cost frameworks do not apply. Cloud and neocloud operators sell on consumption or committed terms, so the metrics are gross margin, utilization, net revenue retention, and backlog / remaining performance obligations (RPO — contracted future revenue). Across every layer, capacity is priced in power, not floor space — energized megawatts (MW) are the binding constraint — and the business is capital-intensive and front-loaded: operators spend billions before revenue arrives.

6. Demand drivers

Same as the child: artificial intelligence (AI) is now the dominant driver of both revenue and capital spending; cloud migration continues as enterprises move workloads off their own servers; data growth, streaming, analytics, and cybersecurity add steady volume; and electric-power availability has become the effective ceiling on growth, with data-center electricity use rising fast enough to draw grid-planning and policy attention. See 518210 for the specifics and figures.

7. Regulation

There is no single regulator. The industry sits at the intersection of land/power/environmental permitting (now the most active front, with numerous projects delayed or blocked and several states weighing moratoriums), health-data rules (the Health Insurance Portability and Accountability Act, or HIPAA), government-cloud security authorization (FedRAMP, the Federal Risk and Authorization Management Program), competition scrutiny from the Federal Trade Commission, and a patchwork of state privacy and cybersecurity statutes plus export controls on advanced AI chips. The child primer covers each regime.

8. Consolidation

The industry is tiered and consolidating: a tight hyperscale oligopoly at the top, two well-capitalized public data-center landlords in the middle, and a wave of private-equity roll-ups — data-center M&A hit record levels in 2024, with tens of billions of disclosed private-equity spending, because contracted, often inflation-linked cash flows suit infrastructure funds. Neoclouds are the newest entrants. Full detail and deal figures are in 518210.

9. Risks

The principal risks — carried in full by the child primer — are AI overbuild / demand air-pockets (capacity built on debt against forecasts), power cost and availability, community and regulatory backlash, customer concentration (acute for neoclouds), hardware obsolescence (GPUs can lose value before facilities depreciate), financing and interest-rate risk, cybersecurity and outages, and disclosure/structure risk (public parents may not break out cloud economics; private owners disclose even less and are often highly leveraged).

10. How to invest & outlook

Public-market routes span cloud-platform parents (diversified, with limited segment transparency), data-center REITs (landlord-style, dividend-paying, rate-sensitive), AI-cloud developers (higher growth and leverage), managed-hosting and edge names, and listed alternative-asset managers for indirect exposure to private platforms. Private-market routes reach the large share of privately held capacity through infrastructure and private-equity funds, direct and co-investment stakes, and private credit financing construction — concentrated, income-generating, but illiquid and disclosure-light. Tickers, a diligence checklist, and the full forward-looking outlook are in the 518210 primer.

In short: the near-term story is dominated by the AI capital cycle, and the likely winners hold secured power, dense connectivity, strong counterparties, disciplined capital spending, and manageable leverage — not just the AI label. Because 51821 is 518210, that primer is the complete reference for this level.

---

Sources

Drawn from the child primer (518210), which carries the full source list.

1. U.S. Census Bureau, "2022 NAICS: 518210 Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services." https://www.census.gov/naics/?details=518210&input=518210&year=2022
2. U.S. Census Bureau, County Business Patterns 2023 (establishments 18,544; employment 629,527; annual payroll $92.96B; Q1 payroll $24.05B). https://www.census.gov/data/datasets/2023/econ/cbp/2023-cbp.html
3. U.S. Census Bureau, 2022 Economic Census — Concentration of Largest Firms, NAICS 518210 (receipts $329.47B; 12,054 firms; CR4 37.2%, CR8 41.9%, CR20 50.3%, CR50 61.2%; HHI suppressed). https://data.census.gov/table/ECNSIZE2022.EC2200SIZECONCEN?q=EC2200SIZECONCEN
4. Amazon.com, Form 10-K for FY ended December 31, 2025 (AWS segment ~$129B). https://www.sec.gov/Archives/edgar/data/1018724/000101872426000004/amzn-20251231.htm