Billionaire Jim Goodnight Built An Analytics Profit Machine. AI Is Forcing Its Reinvention.
ByPhoebe Liu,Reporter.
May 15, 2026, 06:30am EDT
Updated May 15, 2026, 10:36am EDT
Unlike most of today’s biggest AI companies, SAS—once America’s largest privately held software company—has always operated slowly, steadily and profitably. Competition from all sides and an upcoming leadership transition will test the company’s longstanding strategy.
Clad in a plain white shirt, Jim Goodnight, billionaire cofounder and CEO of analytics firm SAS, eases into a leather rolling chair in a Cary, North Carolina, meeting room that looks less like a corner office than a geology exhibit. Behind him are glistening gemstones. A clump of pyrite. Purple amethysts. A fossilized dinosaur egg—a 69-million-year-old Hadrosaurus found in the Gobi Desert. A meteorite. “It’s not something you want to get hit in the head by,” he deadpans.
SAS is 50. Its CEO is 83. And like the rocks on display, both are artifacts from an earlier time long before fast-growing, deeply-unprofitable AI shook the world. SAS analyzes large troves of data from its customers in real-time to help them make better business decisions.
“People like to dismiss us by saying, ‘well, that’s legacy software,’” says Goodnight, a statistics pioneer who helped define what analytics would be long before AI became an umbrella term for everything. “But it’s not. We’ve been improving it for 50 years.”
Now SAS has to prove that endurance isn’t the same thing as stagnation.
The company generates just over $3 billion in annual revenue from most of the Fortune 100—including 90% of the financial services companies and all of the health and life sciences firms, plus most every government department. It has stayed private, profitable and debt free.
The AI bo-m is stress-testing that posture. OpenAI, Anthropic and a swarm of newer data-and-analytics rivals are selling the future as a clean break from “legacy” incumbents. Hyperscalers like Microsoft and Amazon are bundling data and AI into cloud contracts. Public-sector competition is heating up. And inside SAS, the next chapter is no longer theoretical: Goodnight has been hinting for years at a leadership transition, including an IPO as a possible succession plan. “When we go public, we need a different CEO,” he says. “You don’t want an old fa-t like me going around trying to sell stock.”
For a company designed to outlast market volatility, an uncomfortable question is suddenly immediate: can SAS modernize fast enough to matter in the AI era—without abandoning the slow, profitable discipline that made it an outlier in the first place? And can it do it without Goodnight?
Goodnight is confident it can; he’s seen this cycle before: the dot-com bo-m, when he considered outside money and passed; the dot-com bust, which rewarded that restraint; failed investments, including an airline; and the 2022 market correction which may have forced SAS to delay its IPO plans. He’s unmoved by the idea that generative AI has rewritten the laws of business.
AI is “just picking the next word in a sentence based on probability,” Goodnight says, correctly, of large language models. “How’s that going to solve anything?” He thinks SAS’ decades of customer trust and “domain expertise,” particularly in finance, healthcare and government services, will help it retain its edge.
Yet Goodnight will likely leave SAS’ future in the age of AI to younger hands.
In recent years, he has ceded more of the daily operating work to a new generation of executives, especially chief technology officer Bryan Harris and chief operating officer Gavin Day. Goodnight says he’s training Harris and Day to take over, though he hasn’t yet decided which of the two he would like as CEO.
The plan they are inheriting is simple to describe and hard to execute: persuade customers that SAS is not the same company it was 50 years ago, sell them on AI that helps them make smarter business decisions instead of merely sounding like it might, and mold the products to meet every customer where it’s needed.
“Incumbency is our biggest headwind,” says Harris.
That incumbency can be seen in SAS’ sprawling North Carolina headquarters. Its 300-acre tree-lined property boasts a day care and doctor’s office, fields dotted with employees playing intramural soccer at lunchtime, one of the state’s few five-star hotels and dozens of docile sheep grazing underneath the company’s solar panels. Turn left from Analytics Drive onto Research Drive and walk down Binary Way, and you’ll be blinded by a shining silver sculpture of the mathematical constant pi. The company’s campus, as they call it, reflects Goodnight’s vision and SAS’ academic origins.
SAS, short for Statistical Analysis System, was born out of North Carolina State University where Goodnight—then a young faculty member fresh out of a statistics PhD—teamed up with Tony Barr in the late 1960s to create software that sifted through and analyzed N.C. State’s agriculture department data. After the tool attracted more than 100 outside customers, Goodnight, Barr, John Sall and Jane Helwig incorporated SAS Institute in 1976. Barr sold his 40% stake for $340,000 in 1979. Helwig, who died in 2021, left and sold her stake a couple of years later. Goodnight now owns two-thirds of SAS, making him worth $13.3 billion and the richest man in North Carolina; Sall owns the remaining third, a $6.5 billion stake.
From the beginning, the company was bootstrapped. Back when SAS software was sold as physical books, all staff—including the founders—would form an assembly line every time a new shipment of books arrived to unload the books into an employee’s basement, a tradition cofounder Sall recalls as “book brigades.”
When the phones from prospective customers stopped ringing, Sall says Goodnight—in keeping with his upbringing as a hardware shopkeeper’s son—forced the cofounders to split up SAS’ potential customers into four (grouped alphabetically) and do the marketing themselves.
The approach worked. SAS was cash-flow positive from day one and generated $600 million (revenue) on an estimated $300 million in operating income by 1996, Forbes previously reported. SAS grew steadily, always prioritizing profitability over the fastest possible growth, Sall says.
Along the way, as evidenced by its campus, SAS built up a reputation as a company that takes care of its employees. Extensive benefits—beginning with free M&Ms (11,000 pounds per week, company-wide) then expanding to on-site doctors and a pharmacy, subsidized on-site childcare and a hair salon—weren’t common in the ’80s and ’90s. It was Goodnight's retention strategy: keep employees happy, keep turnover low and avoid the expensive churn of bonuses and dilutive stock options.
He used to joke that 95% of SAS’ assets, its people, drove out the front gate every night. After the pandemic and a remote-work policy, the line no longer works quite the same way. “I can’t even get ’em to come in,” he says.
Three years ago, Harris brought Goodnight an idea he loved. SAS could use computer vision to analyze video feeds from farms and determine how illnesses spread among chickens. The tool would help farmers keep their flocks healthy. Goodnight ki-led it with a single question: “How much do the cameras cost? The farmers would never pay for that.”
From the perspectives of both customers like those farmers and SAS itself, Goodnight has been laser-focused on cost and profitability for decades. He criticized AI innovation for being 90% wasted dollars, and repeatedly emphasized SAS’ need to get further into the green.
The CEO credits SAS’ durability to that desire to stay profitable, even at the expense of rapid growth. While Anthropic has reportedly grown revenue at roughly 10 times year over year for three years, SAS’ revenue rose 9% last quarter, roughly in line with Morningstar's prediction that software companies will grow at around 10% per year through 2029.
Goodnight thinks the AI companies’ pace “needs to slow down.” But that doesn’t mean SAS has ignored the market. In 2023, the company announced a three-year, $1 billion investment to develop AI-powered products. “It looked like we were going to spend that much anyway, so we announced it,” Goodnight says flatly.
The problem is that SAS is hardly alone here. It is up against rivals that bet on AI first, and more heavily. On the mega-cap side, there’s Microsoft, Amazon and Oracle. Slightly newer entrants: Snowflake, Databricks, Alteryx and others. On the public sector side, Palantir has been siphoning U.S. government contracts from SAS and others. (Palantir’s U.S. government revenue grew by around double SAS’ total government revenue last year.)
SAS’ modus operandi is to meet customers wherever they are most anxious. The company works with nearly every major bank and the Big Four accounting firms, helping them use AI in ways that are secure, traceable and useful for fraud detection and financial risk. Healthcare, government, finance and other regulated industries are natural terrain for a company that has spent decades selling caution as a feature. Even there, the pressure is rising. Anthropic has been hiring industry experts and in May announced a suite of financial-services products that compete directly for the same customers.
“Everyone is in ‘coopetition,’” Harris says. Customers have asked SAS to integrate with its rivals, and the company has happily obliged.
That has made SAS uniquely malleable among its peers. If customers want their data analysis done in the cloud (Microsoft, Amazon, you name it), SAS can do it. If they want it done on premises, SAS will do that too—and in the programming language of your choice. That matters in hospitals and government agencies, especially when sensitive data and regulation collide here and abroad. In the executive building where customers are flown in for meetings, one screen recently read, “Welcome, U.A.E. Government Delegation.”
Harris thinks new revenue streams can come from digital twins—AI-rendered versions of complex physical facilities like manufacturing plants that are used to figure out a facility’s most efficient layout, predict safety incidents without putting workers at risk, and perform virtual testing—via a partnership with Epic Games. Paper products manufacturer Georgia Pacific, for example, uses them to test and train robots in its Savannah River Mill facility, keeping costs down and employees safe. Digital twins currently generate single-digit millions in revenue, but Harris believes the business can grow to $500 million within three or four years.
SAS is also experimenting with quantum computing for ultra-complex transactions, like in fraud detection for banks, that traditional computers can’t handle. Also in SAS’ plans: using data and AI to help sports teams. In December, SAS announced a partnership with Liverpool to use its products to market to the soccer team’s fans better. At SAS’ 50th anniversary conference, the company announced a smattering of new tools that incorporate AI agents.
“SAS has never met a problem they didn’t want to go after,” says IDC research director Kathy Lange, who previously worked at SAS and suggested that the company could benefit from more focus. “It’s a double-edged sword.”
Believing it’s the best way to sell some of his stake without needing to sell SAS for parts, Goodnight still wants an IPO. But five years after SAS first said it was preparing to go public, the window has narrowed, shifted and occasionally looked like a regret chute.“We don't want to go when all the money has been already used for SpaceX,”
The numbers also need work. Before hitting the roadshow, Goodnight wants to meet the Rule of 40, a common software company benchmark in which revenue growth rate and profit margin sum to 40. That might help the company defend its share price in public markets, especially when pitted against fast-growing competition. But with both components sitting at around 10%, Goodnight says SAS isn’t even halfway there.
For CFO Matt Parson, it’s optionality that’s the key here. SAS has to be ready for the public markets, but they can’t be the only path to helping Goodnight and Sall sell some of their stake. Why sell? The founders’ children aren’t planning to take over, but Goodnight and Sall might still like to leave them with some cash. They’ve yet to take much out of SAS: the company pays out a small dividend, but has invested most profits—“many billions of dollars”—back into the business over its lifetime.
In case an IPO isn’t possible, Parson thus wants to prepare the firm for other solutions: an acquisition or outside investment. The company routinely gets acquisition offers, but Goodnight hasn’t entertained any of them. (The last publicly reported bid was Broadcom’s $15-20 billion offer in 2021; it was progressing until Goodnight changed his mind.) A minority investment could be in the cards, according to Parson, if the right partner came along. If SAS can remain profitable, it can also stay as-is for the foreseeable future: private and founder-owned.
Sipping a cup of black coffee, this time in front of a piece of the Berlin Wall he helped smash, Goodnight is risk-adverse as ever. He is ready to stop being the face of the story he created.
“I wish people knew nothing about me,” he says, with a wink