U.S. Boiler, Tank, and Shipping Container Manufacturing (NAICS 3324): An Investor Primer

This is a rollup page for a four-digit North American Industry Classification System (NAICS) industry group. NAICS 3324 gathers three quite different metal-forming businesses under one heading. The distinctive value here is the contrast across the three children; each has its own leaf primer for the company-level detail.

1. Overview

NAICS 3324 covers factories that form metal into vessels and containers, from safety-critical pressure equipment down to the can in a vending machine.[1] Three child industries sit inside it: power boilers and heat exchangers (33241), heavy-gauge metal tanks (33242), and light-gauge metal cans, drums and boxes (33243). What they share is metalworking economics: expensive steel or aluminum content, welding and forming labor, high fixed costs, and short economic freight radii because empty vessels are bulky relative to their value.

What they do not share is almost everything else. One end of the group builds engineered, code-stamped pressure equipment for power plants and refineries; the other end mass-produces consumer packaging at extreme line speeds. They sell to different customers, answer to different regulators, run at very different capital intensities, and — most important for an investor — offer very different ways in. Cans are where the public stocks are; tanks are where the plant count and private ownership are; boilers are the smallest but highest-barrier slice.

Two things have sharpened since this page was last written. First, the three children are visibly not moving together: boiler-side drivers are constructive, the tank child's own employment and plant counts fell year over year, beverage and pet-food cans are growing while human-food cans shrink outright, and the drum-and-pail end is in an acknowledged industrial contraction.[4][5][6] Second, the quality of federal coverage now differs by child — the tank child has withdrawn its revenue, firm-count and concentration figures as unconfirmable, so the group's own Economic Census totals are no longer fully corroborated from below. Both points are treated in Sections 2, 3 and 8. This primer leads with the comparison, then covers the group as a whole.

2. What's inside — the three child industries and how they differ

The table below is the heart of the page. "Share of level" is shown three ways because the three measures disagree in an instructive manner: revenue, jobs, and plant count each rank the children differently.

Sources for the shares: revenue from the 2022 Economic Census; jobs and plants from 2023 County Business Patterns (CBP). Company and contract detail is carried by the child primers.[2][3][4][5][6]

Three notes on the shares. First, the ~60% revenue share of the can/light-gauge child overstates its economic weight relative to the other two. Beverage-can revenue carries a very high metal (aluminum) content that passes straight through to the customer, so revenue is inflated versus value actually added — Ardagh puts variable costs, which include metal, at roughly 75% of its cost of sales.[6] By jobs the three children are far more even (~27% / ~37% / ~37%), and by payroll the gap narrows to ~28% / ~33% / ~38%.[4][5][6]

Second, the tank child no longer has a revenue or firm-count figure. The 33242 page has withdrawn the 2022 Economic Census receipts total ($10.338 billion) and firm count (611) it used to carry, because its rebuilt research could not confirm them against a published source; employment and payroll are now its only reliable size anchors.[5] The group's own $43.739 billion total still stands on its own Economic Census source, so the residual after boilers and cans is roughly a quarter of the group — but that is an inference from the parent total, not a published figure, and it should not be quoted as one.

Third, the firm counts no longer reconcile at all. Boilers report 185 firms and cans 276 against the group's 1,066; the tank count is gone.[2][4][6] Even if it returned, firms could never be residualized cleanly, because a company active in more than one child is counted once at the group level but in each child it operates in. Establishment, employment and payroll counts, by contrast, still sum exactly (see Section 3).

The one-line takeaway: revenue and payroll point to cans, plants point to tanks, and barriers-to-entry point to boilers — and the ownership follows, with listed exposure concentrated almost entirely in the can child.

3. Size — the group as a whole

Federal figures for NAICS 3324 come from two U.S. Census Bureau programs: the 2022 Economic Census (revenue, firms, market concentration) and the 2023 County Business Patterns (establishments, employees, payroll). Because they cover different years they are complementary, not a single snapshot. Every figure below is a reported federal value; nothing is estimated or filled in for a suppressed cell.[2][3]

The 2023 CBP figures still reconcile perfectly: the children's establishments (237 + 678 + 422 = 1,337), employees (23,734 + 32,394 + 32,602 = 88,730), annual payroll (~$1.94B + ~$2.28B + ~$2.60B ≈ $6.82B) and first-quarter payroll (~$478M + ~$555M + ~$661M ≈ $1.69B) all sum exactly to the group, so the Section 2 jobs and plant shares are arithmetic, not estimates.[4][5][6]

The 2022 Economic Census figures do not. Coverage below the group line is now uneven: the can child carries a full Economic Census record (revenue, firms and concentration); the boiler child carries one at the industry level but its own leaf primer does not, having failed to reach the table and declined to estimate; and the tank child has withdrawn its Economic Census figures entirely.[4][5][6] The revenue, firm and concentration rows above therefore rest on the group's own source and are only partly corroborated from below. Treat them as this page's figures.

Pay per worker averages ~$76,900 at the group level but spans a wide range — from roughly $70,500 in heavy-gauge tanks to ~$81,700 in boilers, and, within the can/light-gauge child, from ~$88,100 in canmaking to ~$61,240 in drums and pails.[4][5][6] That spread is a fair proxy for how differently capital-intensive these businesses are: canmaking averages roughly 132 employees per plant against about 40 in drums and pails.[6]

Only one child carries a year-over-year comparison, and it points down: heavy-gauge tanks lost 4.7% of employment and 4.2% of establishments between 2022 and 2023, with nominal annual payroll off 1.3% even as first-quarter payroll rose — fewer workers paid more per head.[5] No equivalent two-year series is published on this page for the group or the other two children, so this is a signal about tanks, not about NAICS 3324.

Undercount caveat. These counts cover employer establishments and firms with payroll. They omit owner-only (nonemployer) shops, most government production, and fabrication buried inside a company classified under a different primary activity (for example captive naval, defense or food-processing plants). They also exclude economically linked activity in separate codes — notably drum reconditioning (NAICS 811310) and pure tank installation (construction). The undercount is smallest in the capital-intensive corners (boilers, high-speed canmaking) and largest where small, private, regional operators cluster — heavy-gauge tanks and the drum-and-pail end of the can child. In those corners, the true operator count runs modestly above the employer counts above. No suppressed value is stated anywhere on this page.

4. Investable universe — where value (and access) concentrates

The single most useful fact for a public-market investor is that listed exposure in NAICS 3324 is heavily lopsided toward one child, and it is not the child with the most plants.

• Cans (part of 33243) hold nearly all the clean listed exposure. This is the only corner of the group with several near-direct public pure plays: Ball, Crown Holdings, Silgan Holdings, Ardagh Metal Packaging and Sonoco Products, plus indirect captive exposure through Anheuser-Busch InBev, which owns Metal Container Corporation.[6] Two cautions travel with that menu: listed is not the same as independent — Ardagh Metal Packaging was approximately 76% controlled by Ardagh Holdings as of late 2025 — and segment revenue is not NAICS revenue, since Ball's $6.29 billion North and Central American beverage segment and Silgan's $3.14 billion metal-container segment both reach beyond the United States.[6] Ownership beyond the listed names spreads across pension funds (Ontario Teachers' in Trivium Packaging), family groups and captive plants.[6]
• Boilers (33241) are reached through diversified equipment makers. Babcock & Wilcox is the most direct listed name and also the smallest, at $587.7 million of 2025 revenue — under a tenth of the boiler child's reported sales. Graham Corporation ($245.3 million of fiscal-2026 sales, 85% of backlog defense-related) and BWX Technologies (a $853.1 million Commercial Operations segment within $3.198 billion of 2025 revenue) supply the higher-barrier nuclear and defense end, and Baker Hughes and Modine (fiscal-2026 Climate Solutions sales of $2.062 billion) the broad thermal end.[4] None discloses U.S. output in this code. Large private and strategically owned operators — Cleaver-Brooks, Kelvion, Babcock Power, Fulton, Metalforms Heat Transfer, API Heat Transfer — round out the field.[4]
• Heavy-gauge tanks (33242) are mostly a private-market game. Public access is either diversified (Baker Hughes, Worthington Enterprises), small and illiquid (Paul Mueller, TerraVest, which owns Highland Tank), or adjacent project exposure through engineering-and-construction firms such as Matrix Service that erect field-built storage rather than manufacture it.[5] The substantive operators are private or employee-owned: CB&I, bought from McDermott for roughly $450 million in December 2024 by a Mason Capital-led consortium; CST Industries; the employee-owned Tank Connection; TRIARC Tank, sold by Arcosa for $275 million in 2022; and the Tarsco businesses.[5]
• Other containers (part of 33243) are almost entirely private. Screening the narrow drum-and-pail definition yields one diluted public proxy — Greif, whose Durable Metal Solutions segment reported $1.37 billion of worldwide sales of which only $251.8 million was in the United States for the eleven months ended September 2025.[6] The rest is PE platforms (Mauser, Skolnik), family producers (North Coast Container, Schütz) and an employee-owned specialist (Cleveland Steel Container).[6]

One transaction now spans two children and is worth flagging at this level: Baker Hughes completed its acquisition of Chart Industries on 16 July 2026 at $210 per share in cash, bringing Chart's Heat Transfer Systems business (boiler-side exchangers; $1.468 billion of orders on $1.035 billion of 2024 sales, closing with $2.097 billion of backlog) and its Cryo Tank Solutions business ($624.2 million of 2025 sales) under one roof.[4][5] A single diversified buyer is now a meaningful counterparty in both 33241 and 33242.

There is no clean industry exchange-traded fund (ETF) and no single listed proxy for NAICS 3324. Tickers, listings, revenue, backlog and transaction benchmarks live in the three leaf primers. The net map: a public investor's menu is dominated by the can child; two of the three children are reached mainly through private-equity, family-business and employee-ownership channels.

5. How the money works

Three broad economic models coexist in the group, and which one dominates depends on the child.

• Project / engineered-to-order (most of boilers, much of heavy-gauge tanks). Revenue tracks customer capital spending, contract awards and backlog conversion rather than throughput. Large jobs use percentage-of-completion or milestone billing, so a revised cost estimate can recognize a loss immediately, and fixed-price terms convert metal or labor inflation into overruns.[4][5]
• High-volume manufacturing (cans, and the repeat end of drums/pails). Returns turn on line utilization, uptime, speed, spoilage and freight discipline; Silgan's metal-container plants generally serve customers within a 300-mile radius, and Ardagh's roughly 75/25 variable-to-fixed cost split explains why volume swings hit reported margins faster than the headline metal exposure suggests.[6]
• Aftermarket / service (replacement tubes and parts for boilers; inspection and repair for tanks; collection and reconditioning for drums). This layer is steadier and converts to cash better than new-build — Babcock & Wilcox's parts business alone was roughly 41% of 2025 revenue — and it deepens customer ties, though drum reconditioning also cannibalizes new-unit demand.[4][5][6]

Demand visibility is the sharpest divide in the group, and it runs almost the full available range. At one end, the can business is contracted years out: Silgan has roughly 90% of projected 2026 metal-container sales under multi-year arrangements and Ardagh Metal Packaging more than 80% of 2025 revenue under two- to seven-year agreements. In the middle sit the project businesses, where backlog gives visibility but its quality and margin are the open question. At the other end, the drum business is effectively spot — Greif reports that many industrial-packaging customers order weekly for same-week delivery, leaving almost no forward book.[4][5][6] Same freight economics, opposite ability to see the next twelve months.

Margins disperse just as widely, and the label on the box does not predict them. Within tanks alone, Matrix Service's storage-and-terminal segment earned a 4.0% gross margin on $365.9 million of fiscal-2025 revenue while Chart's Cryo Tank Solutions produced a 23.0% gross margin on $624.2 million of 2025 sales; Greif's Durable Metal Solutions ran a 20.6% gross margin, Graham a 23.5% gross margin, Silgan's metal-container segment an 8.3% adjusted margin on earnings before interest and taxes, and Ball's North and Central American beverage segment a 12% comparable operating margin.[4][5][6] Field-erected commodity fabrication and code-certified specialty equipment are different businesses wearing the same NAICS code.

The common thread across all three is metal cost and its pass-through. Carbon and stainless steel, aluminum can sheet, tinplate, nickel alloys, tubes and forgings are the dominant inputs, and most contracts carry metal-cost adjustment clauses. Prices have moved hard in both directions of the ledger: the boiler child's output producer price index rose 15.1% year over year to June 2026 while steel-mill products rose 16.9% — inputs outrunning output pricing — and the tank child's price index climbed from 163.1 in December 2020 to 305.8 in June 2026.[4][5] That means revenue can rise on metal inflation without profit dollars rising in step, and timing lags, tariffs, scrap and uncovered conversion costs still bite. The most useful cross-child operating measures are orders and book-to-bill, backlog and its margin (for the project businesses), unit shipments and price/mix excluding metal pass-through, line utilization and scrap (for the volume businesses), and aftermarket share.

6. Demand drivers

Because the three children sell into different end markets, they move on different cycles — a diversification the group offers as a whole, and one the current evidence makes concrete.

• Boilers ride electricity and data-center load growth (the U.S. Department of Energy put data centers at 4.4% of national electricity in 2023, projected at 6.7%–12% by 2028), plus power construction, liquefied natural gas (LNG) export, nuclear life-extension and new build, and routine industrial replacement during scheduled outages.[4][7] Those drivers do not convert into orders evenly, which is the point worth carrying up: developers planned 86 gigawatts of U.S. generating-capacity additions for 2026, but only 6.3 gigawatts is natural gas and 3.3 gigawatts combined-cycle — solar and batteries dominate the total and use little steam equipment.[4] LNG is the more mechanically direct pull, with North American export capacity forecast to rise from 11.4 billion cubic feet per day in 2023 to 24.4 billion by 2028.[8]
• Heavy-gauge tanks lean on the steadier water-infrastructure market — the Environmental Protection Agency's sixth drinking-water needs assessment identified $472.6 billion of required investment over 20 years, of which $47.6 billion is for constructing, rehabilitating or covering water-storage reservoirs — plus more cyclical energy, chemical, terminal and industrial-gas demand, with replacement and inspection cushioning the cycle.[5][9]
• Cans and other containers are split, and now split in direction as well as end market. Beverage cans and pet food are growing (Crown reported global beverage-can shipments up 5% in the first quarter of 2026, up 1% in North America; Silgan reported metal-container volume up 2% on double-digit pet-food growth), while U.S. human-food can shipments fell from 26.3 billion units in 2022 to 24.4 billion in 2023 and 23.7 billion in 2024.[6][10] Drums and pails are industrial-linked and contracting outright: Greif has warned that the multi-year industrial contraction was likely to continue into fiscal 2026.[6]

Two forces cut across the group. Installed-base replacement is generally steadier than new-plant or new-line construction everywhere. And substitution matters at the container end, where metal competes with plastic, glass, fibre and reusable intermediate bulk containers — winning on barrier protection, fill speed, hazardous-material certification and recyclability, losing on weight and cost. Circularity cuts differently by child: for cans it means collection (the U.S. aluminum beverage-can recycling rate fell to 43% in 2023 against a historical average of about 52%), while for drums it means literal reuse, which directly displaces new-unit demand.[6]

7. Regulation

Compliance is both a cost and a barrier to entry across the group, but the specific regime shifts with the product.

• Pressure equipment (boilers and tanks) centers on the American Society of Mechanical Engineers (ASME) Boiler and Pressure Vessel Code (BPVC) — Section I (power boilers), Section VIII (pressure vessels), Section III (nuclear). Certification, approved welding procedures, material traceability and nondestructive examination are gating requirements; the National Board Inspection Code, adopted in most U.S. and Canadian jurisdictions, governs installation, inspection and repair after the unit ships; and nuclear suppliers add Nuclear Regulatory Commission (NRC) quality assurance.[4][5][11]
• Tanks additionally follow American Petroleum Institute (API) Standard 650 (welded oil-storage tanks), American Water Works Association standards, Environmental Protection Agency (EPA) spill-prevention rules, and Build America, Buy America domestic-content requirements on federally assisted water projects.[5]
• Containers face food-safety and coating rules for cans (Food and Drug Administration food-contact requirements; EPA emissions standards for can coating, which can also reach cleaning, lining and painting at some drum plants) and hazardous-materials rules for drums (Pipeline and Hazardous Materials Safety Administration and United Nations marking under Title 49). The EPA's treatment of drum reconditioning under the "empty container" rule remains an open overhang — the agency continues to examine the risks after its 2023 advance notice, with no final replacement rule described as of July 2026.[6]

Two threads run through all three. The Occupational Safety and Health Administration (OSHA) governs welding, machine guarding and hot work everywhere — enforcement is real but not extreme, with 59 citations across 17 inspected heavy-gauge tank establishments and $281,000 of penalties in the year to September 2025.[5] And Section 232 metals tariffs are a live cost variable for every child: rates on steel and aluminum rose from 25% to 50% effective 4 June 2025, the regime now generally applies 50% to covered metal products and 25% to many derivative products, and a July 2026 proclamation authorized half-rate primary-aluminum imports for companies with approved U.S. smelting-investment plans.[5][6][12] Tariffs are no longer background; they move faster than most contracts reset.

8. Consolidation

At the group level NAICS 3324 is unconcentrated: an HHI of 328 (the sum of each firm's squared market share, where under 1,500 is the federal "unconcentrated" zone) and a four-firm share of 30.9%.[2] That aggregate blends three markets at different stages, and it hides more than it reveals — every child measured on a comparable 2022 basis is at least as concentrated as the group, which is what aggregation does when distinct markets are pooled into one denominator.

• The can/light-gauge child is the most concentrated measured slice: CR4 of 51.4% and an HHI of 844, with CR50 at 95.9%. Even that blends a tightly held can business with a more open drum-and-pail one — Ball estimates five companies make substantially all North American aluminum beverage cans and puts its own share at roughly 36% of regional shipments, while Silgan claims more than half of 2025 U.S. metal-food-container volume. Regulators forced eight plant divestitures in the last major can merger.[6]
• Boilers are fragmented on the current federal reading — CR4 of 33.9% and an HHI of 482 — though specific niches, notably nuclear components and naval propulsion, are far more concentrated.[4] The two vintages of evidence disagree in level: the 33241 page reports 33.9% from the 2022 Economic Census, while its leaf primer, unable to reach a current table, falls back on a Small Business Administration analysis of the 2012 Economic Census putting CR4 at 27.3%.[4] They agree in direction — no single company dominates nationally — but the gap is real and unresolved.
• Heavy-gauge tanks have no current concentration measure at all. The child has withdrawn the 2022 statistics this page previously carried (including the HHI of ~196) as unconfirmable. The last published federal ratios date from the 2002 Economic Census, when the four largest companies held 18.6% of industry value added, the top eight 27.7%, the top 20 41.0% and the top 50 57.5%.[5][13] Those numbers still read as the most fragmented of the three, and two decades of roll-up activity have probably raised them, but the evidence is old and on a value-added rather than shipments basis — it is a direction, not a level.

So the group's 328 understates the concentration of the can oligopoly, and for tanks it can no longer be compared against anything current. Deal activity is brisk across all three children regardless. In boilers: Miura bought Cleaver-Brooks in May 2024, Apollo-managed funds completed their acquisition of a majority interest in Kelvion in January 2026, and Metalforms Heat Transfer acquired Koch Heat Transfer's North American business in September 2024.[4] In tanks: the CB&I and TRIARC transactions and continuing roll-up of regional fabricators, with disclosed markers pointing to modest pricing for ordinary fabrication — TerraVest paid $78 million for Highland Tank in 2023, roughly four times trailing earnings before interest, taxes, depreciation and amortization (EBITDA).[5] In drums and pails: Mauser's lifecycle acquisitions, the Stavig family's North Coast Container platform and Pelican's purchase of Skolnik.[6] And Baker Hughes' Chart acquisition, noted in Section 4, consolidates across two children at once. Consolidation broadens aftermarket reach and product range but does not remove local fabrication capacity or customer-qualification requirements.

9. Risks

Shared across the group:

• Metal-cost and tariff mismatch: pass-through clauses may not fully or promptly recover steel, aluminum, energy or Section 232 tariff costs, and in boilers input prices have recently outrun output prices.[4]
• Fixed-price execution: in the project businesses (boilers, engineered tanks), estimating errors, welding rework or commissioning problems can erase a job's margin — one crude-terminal project's labor productivity alone cost Matrix Service $5.1 million.[5]
• Utilization and overcapacity: in the volume businesses (cans, drums), a few poorly timed new lines can depress regional pricing and fixed-cost absorption, and there is no published federal utilization series for the metal-container detail to check it against.[6]
• Skilled-labor scarcity: the Bureau of Labor Statistics projects just 2% growth in welder employment from 2024 to 2034 but about 45,600 openings a year, nearly all replacements — and code-qualified pressure-vessel welders and nuclear-quality personnel are a far narrower pool than the headline occupation.[5][14]
• Cyclicality and customer concentration: power, LNG, chemical, industrial and consumer capital spending arrive in waves, and large buyers hold leverage — Graham reported two customers each above 10% of revenue with 85% of backlog defense-related, and qualified metal suppliers on the can side are few.[4][6]
• Leverage: debt-funded PE roll-ups can turn stable operating earnings into volatile equity returns, and can understate assumed environmental and warranty obligations.

Child-specific: energy-transition risk to legacy boiler demand (heat exchangers are materially less exposed); product-liability, leak and environmental claims for tanks and drums; substitution and demand-mix swings for cans, where a shrinking human-food base can offset beverage and pet-food growth; unresolved EPA rulemaking on drum reconditioning; and reuse cannibalizing new-drum sales.[4][5][6]

10. How to invest & outlook

Public investors. Genuine listed exposure sits almost entirely in the can child — compare Ball, Crown, Silgan, Ardagh and Sonoco on normalized conversion earnings (not revenue inflated by metal pass-through) and leverage, remembering Ardagh Metal Packaging's small float against ~76% control. Boilers are reached through diversified equipment names (Babcock & Wilcox most directly; BWX Technologies and Graham for nuclear/defense; Baker Hughes and Modine for broad thermal exposure), and tanks only through diversified, illiquid or adjacent contractor names. Because no listed name discloses output in these codes, segment disclosure — not the corporate label — is the number that matters. Useful cross-name measures: organic orders, book-to-bill, backlog margin, contracted volume and renewal dates, free cash flow, net debt, and enterprise value to EBITDA. There is no single ETF or clean 3324 proxy.

Private investors. The private field is far richer than the public one and covers two of the three children outright — regional tank and pressure-vessel fabricators, boiler service companies, and drum-and-pail platforms owned by PE, families and employees. Diligence should normalize EBITDA for project adjustments, test metal-escalation clauses and their lags, examine backlog margin job by job, verify code certifications and United Nations markings, inspect environmental history, and separate recurring aftermarket earnings from volatile new-build revenue. Sanitary stainless, cryogenic and aftermarket-service niches tend to earn better valuations than commodity fabrication without service revenue — the roughly four-times-EBITDA mark on ordinary tank fabrication is a floor for plain assets, not a benchmark for qualified ones.[5]

Outlook — editorial judgment. The group as a whole looks stable to modestly constructive, not a growth sector, and the three children are pulling apart rather than moving in step. Electricity and data-center load, LNG, nuclear refurbishment, water-system replacement, domestic-content rules and new beverage formats give multi-year support; against that, only 6.3 of 2026's planned 86 gigawatts of generating capacity is gas-fired, fabricated-metal capacity utilization stood at 76.9% in June 2026, the tank child's employment and plant counts fell year over year, and the drum-and-pail end is in an acknowledged contraction.[4][5][6] Metal inflation, tariffs, scarce skilled labor and fixed-price execution can absorb much of the upside. Across all three children, the spread between winners and losers is set at the plant and contract level — qualified niche leaders with recurring service revenue and disciplined terms — rather than by the industry's overall volume. For the full analysis of any child, read its leaf primer.

Sources

1. U.S. Census Bureau, 2022 NAICS Manual / Definitions (NAICS 3324 and children), 2022, https://www.census.gov/naics/reference_files_tools/2022_NAICS_Manual.pdf
2. U.S. Census Bureau, 2022 Economic Census — Concentration by Largest Firms, NAICS 3324 (receipts, firms, CR4/CR8/CR20/CR50, HHI), 2025.
3. U.S. Census Bureau, 2023 County Business Patterns, NAICS 3324 (establishments, employees, payroll), 2025, https://data.census.gov/table/CBP2023.CB2300CBP
4. Power Boiler and Heat Exchanger Manufacturing Primer — NAICS 33241 (companion child primer; carries the underlying company, regulatory and demand citations).
5. Metal Tank (Heavy Gauge) Manufacturing Primer — NAICS 33242 (companion child primer).
6. Metal Can, Box, and Other Metal Container (Light Gauge) Manufacturing Primer — NAICS 33243 (companion child primer).
7. U.S. Department of Energy, Report Evaluating Increase in Electricity Demand from Data Centers, 2024, https://www.energy.gov/articles/doe-releases-new-report-evaluating-increase-electricity-demand-data-centers
8. U.S. Energy Information Administration, North American LNG Export Capacity Is on Track to More Than Double by 2028, 2025, https://www.eia.gov/todayinenergy/detail.php?id=64128
9. U.S. Environmental Protection Agency, EPA's 6th Drinking Water Infrastructure Needs Survey and Assessment, 2023, https://www.epa.gov/dwsrf/epas-6th-drinking-water-infrastructure-needs-survey-and-assessment
10. Can Manufacturers Institute, U.S. Food Can Shipment History, 2025, https://www.cancentral.com/wp-content/uploads/2025/04/CMI-Food-History-for-CC-1970-2024.pdf
11. American Society of Mechanical Engineers, Boiler and Pressure Vessel Code — 2025 Edition, 2025, https://www.asme.org/codes-standards/bpvc-standards/bpvc-2025
12. The White House, Further Adjusting the Tariff Regimes for Imports of Aluminum, Steel, and Copper, 2026, https://www.whitehouse.gov/presidential-actions/2026/06/further-adjusting-the-tariff-regimes-for-imports-of-aluminum-steel-and-copper-into-the-united-states/
13. U.S. Census Bureau, 2002 Economic Census: Concentration Ratios in Manufacturing, 2006, https://www2.census.gov/library/publications/economic-census/2002/manufacturing-reports/subject-series/ec0231sr1.pdf
14. U.S. Bureau of Labor Statistics, Occupational Outlook Handbook: Welders, Cutters, Solderers, and Brazers, 2025, https://www.bls.gov/ooh/production/welders-cutters-solderers-and-brazers.htm