The fastest-growing companies in the Fortune 500

Fortune ranks the 500 largest U.S. companies by revenue every year, and the top of that list barely moves. The more useful question is which of those companies are expanding fastest — adding revenue, profit and market value at a rate their peers cannot match. Growth at this scale is unusual. A business already selling tens of billions of dollars a year needs an entirely new source of demand before its numbers move much at all.

The companies that pull it off tend to share three traits. They identified a structural shift in the economy before it was obvious. They had the balance sheet to build capacity once demand arrived. And they converted that demand into earnings rather than sales volume alone. Over the past several years, a handful of forces have driven the fastest-growing Fortune 500 companies. Artificial intelligence infrastructure and the electricity load created by data centers account for much of the movement. The rest traces to the recovery in ride-hailing and delivery, a U.S. housing market held back by years of underbuilding, and a new class of drugs for diabetes and obesity.

This list covers companies with sustained, documented expansion inside the Fortune 500, along with the products, contracts and strategic decisions behind it. It does not reproduce a single year's ranking, which shifts annually and depends on Fortune's own revenue, earnings and shareholder-return calculations. The focus is growth that can be traced to an identifiable cause: Nvidia $NVDA's position in AI chips, Constellation Energy's nuclear fleet, Eli Lilly $LLY's manufacturing buildout for GLP-1 drugs.

The piece also covers how the ranking is built, since the fastest-growing companies and the largest companies by revenue answer different questions and rarely share names. Fortune's score blends revenue growth, earnings-per-share growth and three-year annualized total shareholder return into a single figure, with minimum thresholds for size and listing history. Knowing which forces produced the last wave of fast movers is the most practical way to recognize the next one. That wave may come from the power grid, the pharmaceutical pipeline or the industrial suppliers feeding data center construction.

Illustration: Quartz

Fortune's Fastest-Growing Companies ranking is built on a formula, not on reputation or momentum. The magazine scores candidates on three measures: revenue growth, earnings-per-share growth and three-year annualized total return to shareholders. Those scores are combined into a single ranking, so the top spot does not simply go to whichever company posted the biggest sales jump.

The blended approach exists because any one measure can mislead on its own. A retailer opening locations at speed can report strong revenue gains while margins stay flat and profit barely moves. A software firm can lift earnings per share through cost cuts and buybacks without selling much more than it did a year earlier. A company must show strength across sales, earnings and market return to score well, which screens out businesses expanding in only one direction.

Eligibility rules shape the outcome as much as the scoring does. Fortune applies minimum thresholds for revenue, market value and share price. A company also needs enough trading history for a three-year return to be calculated at all. Recently listed companies are therefore absent even when their underlying business is expanding quickly. The result is a ranking weighted toward established public companies rather than new arrivals.

The arithmetic of scale matters too. A company with hundreds of billions of dollars in annual revenue must add tens of billions more to match the percentage gain a mid-size firm gets from winning a few large contracts. That is why the Fortune 500 revenue ranking and the growth ranking rarely share many names. Several of the largest U.S. companies by size never surface on the growth list at all.

Treat the two as answers to separate questions. One identifies which U.S. companies are largest at a single point in time. The other identifies which are changing shape fastest. The second is the better guide to where demand in the economy has moved — toward AI hardware, electricity generation or pharmaceuticals, depending on the year.

Illustration: Quartz

Nvidia $NVDA's expansion traces to a single product category: graphics processing units redeployed as the computing engine for training and running artificial intelligence models. The company did not enter a new business. It found that the chip architecture it had refined for video game rendering suited the parallel math that large language models require.

That overlap moved Nvidia from a respected component supplier to one of the most valuable public companies in the world. Its data center division, selling accelerators such as the H100 and the Blackwell generation that followed, now dwarfs the gaming segment that defined the company for most of its history. Microsoft $MSFT, Amazon $AMZN, Google $GOOGL and Meta $META are among the largest buyers, filling server halls with racks of Nvidia silicon.

Hardware is only part of the position. Nvidia spent years building CUDA, the programming platform developers use to write software that runs on its chips. Machine learning frameworks were built against CUDA first. Moving an existing codebase to a rival architecture means rewriting and revalidating work that already functions. That software layer helps explain why competing accelerators have struggled to take share even when their specifications look comparable on paper.

Manufacturing sets the ceiling on how fast the company can grow. Nvidia is fabless — it designs chips and contracts production to Taiwan Semiconductor Manufacturing Company. Output depends on TSMC $TSM's leading-edge capacity and, for the most complex parts, on advanced packaging that stacks high-bandwidth memory next to the processor. Both have acted as bottlenecks, with orders running ahead of supply.

Customer concentration is the clearest risk attached to the growth. A small group of cloud providers and AI developers accounts for a large share of data center revenue. Nvidia's trajectory is tied to capital spending decisions made inside a handful of boardrooms. U.S. export controls on advanced chip sales to China add a second variable. They cut off a market the company had served at scale, leaving results more dependent on North American and allied demand.

Illustration: Quartz

Vistra Corp, a Texas-based independent power producer, has grown on the back of an old business suddenly facing new demand: generating and selling electricity. Data center construction, factory electrification and the move toward electric vehicles have pushed U.S. power consumption upward after a long stretch of flat demand. Vistra sells into that market.

The company's asset mix explains much of the advantage. Vistra owns natural gas plants, nuclear capacity, coal units and grid-scale battery storage across several states, including a large position in Texas through the ERCOT grid. Dispatchable generation — power that can be called on at any hour — matters more as grids absorb heavy industrial loads. Those loads run constantly. They do not pause overnight or on still, cloudy days.

Vistra enlarged its nuclear fleet through the 2024 purchase of Energy Harbor, which added the Beaver Valley plant in Pennsylvania and the Davis-Besse and Perry plants in Ohio. Nuclear output has become valuable to technology companies because it is carbon-free and available around the clock, a pairing wind and solar cannot match without substantial storage attached.

Vistra also runs one of the largest competitive retail electricity businesses in the U.S., selling directly to households and commercial customers under brands including TXU Energy. The retail arm pairs with generation as a hedge against price swings, since the volatility that raises costs for a retailer lifts revenue for a generator.

Investors reassessed the whole category as a result. Independent power producers were long treated as slow, predictable holdings valued mainly for steady dividends. Load growth reframed companies like Vistra as businesses with real volume expansion ahead of them, and comparable repricing reached peers with similar nuclear and gas portfolios.

The same forces define the risk. Vistra's trajectory depends on the data center buildout continuing at its present pace and on wholesale power prices holding firm. It also depends on long-term supply talks converting into signed contracts rather than stalling at the letter-of-intent stage.

Illustration: Quartz

Uber $UBER Technologies moved from heavy losses to consistent operating income and free cash flow. That change, rather than raw revenue expansion, is what put the company back in growth-stock territory. The underlying businesses were already large. What shifted was how much of each dollar of gross bookings reached the bottom line.

Ride-hailing recovered fully from the pandemic collapse in urban travel. Trip volumes and monthly active platform consumers climbed past their pre-2020 levels. Airport pickups, commuting and nightlife traffic all came back, and the company rebuilt driver supply in its core U.S. and European markets.

Delivery became a second pillar rather than a side experiment. Uber Eats pushed beyond restaurant orders into grocery, convenience and retail partnerships, which lifted average order values and gave couriers work outside peak dining hours. Running both networks through one app also held down customer acquisition costs. A rider already in the system can be converted into a delivery customer without fresh marketing spend.

Advertising added a third revenue line with different economics. Sponsored restaurant listings, in-app placements and promoted retail products carry far higher margins than moving a person or a meal across a city. The logistics network behind those ads is already paid for, so incremental ad revenue arrives with little added cost.

Capital discipline completed the shift. Uber sold its self-driving division, Uber ATG, to Aurora Innovation in 2020 and later brought an outside investor into its freight brokerage arm, Uber Freight. Geographic expansion narrowed to markets where the company held defensible positions. Spending concentrated on mobility and delivery instead of a long list of adjacent bets.

Uber reported its first full year of operating profit in 2023, a milestone the company had missed through a decade of expansion. The distinction matters for growth rankings that weigh earnings growth alongside revenue. A business crossing from losses into profit can post percentage gains that a steadily profitable peer cannot match. Bookings alone had grown for years without producing that result.

Illustration: Quartz

Builders FirstSource sells lumber, roof trusses, windows, millwork and prefabricated components to homebuilders across the U.S. Its expansion has followed the durability of new home construction rather than the volume of existing home sales. Those two markets separated sharply once mortgage rates rose. Resale listings thinned as owners stayed put to keep cheap loans, while builders kept starting houses to meet demand that had nowhere else to go.

Scale came largely through acquisition. The company has absorbed regional lumber yards, component manufacturers and distribution operations, assembling a national footprint that can serve a production builder working in several states at once. Its 2021 merger with BMC Stock Holdings was the largest of those deals, joining two of the sector's biggest suppliers. Size brings purchasing leverage with mills and manufacturers. It also lets the company commit to delivery schedules that a single-market supplier cannot match.

The more consequential shift has been in what it sells. Raw framing lumber is a commodity, priced by the market and subject to swings that can erase margin in a quarter. Value-added products — wall panels assembled in a plant, engineered floor systems, pre-hung doors, window packages ready to install — carry steadier pricing and better margins. They also solve a problem builders cannot fix on their own. Skilled framing crews are scarce. A house built from factory-made panels needs fewer carpenter hours on site and goes up faster, which shortens the period a builder pays interest on construction financing.

That positioning explains why the business did not repeat the collapse it suffered after 2008. The U.S. underbuilt housing for more than a decade following the financial crisis, leaving a shortfall relative to household formation that higher borrowing costs did not erase. The company remains exposed to lumber prices and to any genuine downturn in housing starts. Its growth, though, traces to a supply deficit measured in years rather than to a short cyclical upswing.

Illustration: Quartz

Constellation Energy runs the largest nuclear fleet in the U.S., and its expansion since the 2022 spinoff from Exelon $EXC rests on one reversal. Reactors once treated as candidates for early retirement became the power assets technology companies most wanted. Data centers need carbon-free electricity around the clock, and Constellation already owned the plants that produce it.

The turnaround came fast. For most of the 2010s, cheap natural gas from shale drilling pushed wholesale power prices low enough that several merchant nuclear plants could not cover operating costs. Some owners closed reactors outright. Others won state subsidies to keep them running. Nuclear generation was treated as a legacy liability rather than a source of earnings growth.

Two developments changed the math. The Inflation Reduction Act created a production tax credit for existing nuclear plants, putting a floor under revenue for reactors exposed to volatile wholesale markets. Then hyperscale data center operators, bound by their own emissions targets, sought firm clean power that solar and wind cannot supply overnight or during calm stretches.

Constellation's agreement with Microsoft $MSFT to restart a reactor at Three Mile Island in Pennsylvania became the clearest example. The unit was shut down in 2019 for economic reasons, not safety ones. It is being brought back as the Crane Clean Energy Center under a long-term power purchase agreement. Restarting a retired commercial reactor had no real precedent in the U.S. before this wave of demand. Constellation later reached a long-term agreement with Meta $META covering output from its Clinton plant in Illinois.

The company has also pursued growth through acquisition, agreeing to buy Calpine, a large natural gas and geothermal generator, to widen its generation mix beyond nuclear. That mix matters when selling to industrial customers who want clean attributes and reliable delivery at the same time. Constellation's path from here depends on how long data center load growth lasts and whether regulators sign off on the grid arrangements these contracts require.

Illustration: Quartz

Two footwear labels account for nearly all of Deckers Brands' expansion: Hoka and Ugg. The Goleta, California-based company, listed as Deckers Outdoor Corporation, built its recent revenue gains by moving both brands out of narrow niches and into mainstream retail, with no technology story or acquisition spree behind the growth.

Hoka started in France in 2009 as a specialist brand for ultramarathon runners. Its oversized, heavily cushioned midsole was designed to soften long descents on mountain trails, and Deckers later bought the brand outright. The thick-soled profile that once marked Hoka as an outlier turned into an everyday shoe. Models such as the Clifton and Bondi now sell to walkers, nurses, hospitality staff and shoppers who never enter a race. Hoka also pushed into trail, hiking and walking categories well beyond road running. Hoka's annual revenue passed $1 billion in Deckers' fiscal 2023, a scale few footwear brands reach.

Ugg followed a different path. Deckers acquired the brand in 1995, and for years it was identified almost entirely with one item: the sheepskin Classic boot, a product exposed to shifting taste. The company widened the line into slippers, clogs, sandals and reworked boot shapes. The Tasman slipper drew renewed attention from younger buyers through social platforms. A seasonal, single-product name became a year-round business.

Distribution strategy amplified both labels. Deckers pushed sales through its own stores and websites rather than leaning on department stores and third-party retailers, capturing retail margin instead of wholesale margin. It has also managed Hoka's supply deliberately, limiting discounting and controlling which retailers carry the line, which protects full-price selling.

Markets outside the U.S. remain the clearer runway. Hoka's recognition across Europe and parts of Asia still trails its U.S. profile, leaving room to grow where it is not yet an established name. Deckers split its stock six-for-one in September 2024, following a long climb in its share price.

Illustration: Quartz

Eli Lilly $LLY's expansion over the past several years rests on one drug molecule, tirzepatide, sold as Mounjaro for type 2 diabetes and as Zepbound for chronic weight management. It acts on both the GIP and GLP-1 receptors, and demand outran Lilly's manufacturing capacity for an extended stretch after launch, putting the drug on the U.S. Food and Drug Administration's shortage list.

The size of the patient population explains why the revenue line moved so sharply. Obesity affects a large share of U.S. adults, far more people than the diabetes population these medicines were first developed to treat. Weight reduction recorded in tirzepatide's clinical trial program exceeded what earlier prescription weight-loss drugs delivered. Prescribing spread from endocrinologists to primary care physicians.

Lilly answered the shortage by spending billions of dollars on production. The company expanded its manufacturing footprint in Indiana, added capacity in North Carolina and invested in sites in Ireland. Building active ingredient and fill-finish capacity for injectable medicines takes years, so that commitment came well before any certainty about long-term reimbursement.

Coverage has been the swing factor. Insurers and employers have moved unevenly on paying for the weight-loss indication, and Medicare's statutory restriction on covering drugs prescribed solely for weight loss limits how much of the market Lilly can reach at list price. Lilly also sells Zepbound in single-dose vials to self-paying patients, at prices below the drug's list price.

Novo Nordisk, the Danish maker of semaglutide, sold as Wegovy and Ozempic, is the direct rival. The two companies have competed on supply, pricing and new approvals at once. Lilly has pursued additional indications for tirzepatide, including obstructive sleep apnea in adults with obesity, which broadens the clinical argument for insurance coverage.

The growth mechanism differs from the chipmakers and power producers elsewhere on this list. It runs on prescription volume for a chronic medical condition rather than on a construction cycle.

Illustration: Quartz

The buildout of data centers for AI workloads has pulled far more companies into its orbit than the chipmakers and cloud providers at the center of the story. Transformer manufacturers, switchgear suppliers, cooling system makers, electrical contractors and water utilities near large server clusters all face demand conditions that did not exist a few years ago.

Grid equipment is the clearest example. A new data center campus cannot connect to the transmission system without large power transformers, substations and switchgear. Manufacturing capacity for that equipment is limited and slow to expand. Utilities, developers and industrial customers compete for the same production slots, and lead times on some high-voltage gear now stretch into years. Suppliers that spent decades in a low-growth replacement market hold order books they cannot clear quickly.

Cooling is a second pressure point. Servers running AI training jobs generate far more heat per rack than conventional computing equipment, and traditional air cooling struggles at those densities. Operators have moved toward direct liquid cooling, which circulates coolant to cold plates mounted on the chips themselves. The shift creates demand for pumps, piping, heat exchangers and coolant distribution units from firms that previously served industrial and commercial HVAC markets.

Water has turned into a siting issue. Evaporative cooling systems consume substantial volumes of water. That has drawn scrutiny from local governments and residents in regions already dealing with supply stress. Some operators have responded with closed-loop or air-cooled designs in drier areas, trading higher electricity use for lower water consumption.

Commercial real estate has reorganized around the same trend. Data center development has become one of the most active construction categories in the U.S. In established markets such as Northern Virginia, the limiting factor is no longer tenant demand. Available land with an achievable grid interconnection sets the pace of new supply, and interconnection queue positions now shape development timelines as much as construction schedules do.