In 1997, IBM (IBM 0.93%) marked an AI milestone: Its Deep Blue program defeated reigning world chess champion Gary Kasparov, proving that brute force computing power could simulate human thinking. One grandmaster on the scene said that "the future of humanity is on the line," while Newsweek called it "the brain's last stand." The Wall Street Journal estimated that the match brought IBM $100 million worth of publicity.
It's hard to believe that this occurred before three of the "Magnificent Seven" companies were even founded, while a fourth, the AI juggernaut Nvidia (NVDA 0.04%), was years away from going public as a small-cap company. Today, 28 years after that AI triumph, some wonder if it's IBM that's making a "last stand" as the $15.7 trillion AI revolution unfolds.
Major AI savings and a quantum computing breakthrough
For decades, IBM could seemingly do no wrong. A 1985 New York Times article titled "The Daunting Power of IBM" reported that the company "of course, has been the No. 1 computer company virtually ever since the dawn of the computer age more than 30 years ago," and that IBM's hold on market share was so entrenched it raised antitrust concerns. The old office saying, "Nobody ever gets fired for buying IBM," reflected its computers' status as the gold standard. From its 1962 IPO to the dawn of the 21st century, IBM shares soared 3,800%.
In recent years, IBM has been an afterthought for investors, in the wake of its failure to capture a meaningful share in the $752 billion cloud market while Amazon and Microsoft carved out dominant positions. Even Warren Buffett took a capital loss on his shares of IBM when he sold in 2017, though dividends helped the investment squeak into positive return territory. The company was never considered among the Magnificent Seven tech giants soaring as they ushered in a new era.
NYSE: IBM
Key Data Points
Yet shares of IBM are up 45% year to date, edging Nvidia's return of 42%. IBM stock is outperforming shares of the largest, most famous AI stock amid promising signs that its enterprise solutions for clients in the AI space is booming.
In March, IBM's CEO Arvind Krishna sat down with TIME magazine to explain his company's bet on smaller AI models, tailor-made for specific clients where accuracy is paramount, and management's 2020 decision not to build a large AI model in the mold of OpenAI's ChatGPT or Alphabet's Bard (later renamed Gemini). Explaining IBM's decision to eschew large, generalist AI models, he pointed out that they could be orders of magnitude more expensive while only being marginally more powerful, and that smaller models could be more accurate for specific client needs, saying, "If I'd like to control a blast furnace, it needs to be correct 100% of the time."
Six months later, there are signs that this decision is paying off. In Q3 2025, the company's "AI book of business," or the total value of its AI-related deals in software, hardware, and consulting, grew 26% year over year, from $7.5 billion to $9.5 billion. Software revenue rose 10%, while infrastructure revenue climbed 17%. Given the strength of these numbers, management raised its 2025 revenue growth outlook to over 5%.
While not a huge number, it comes as IBM improves margins, with its gross profit margin climbing 1.2% to 57.3%. The improving profitability comes as IBM expects to achieve $4.5 billion in annual savings from improved productivity, thanks to AI solutions, well ahead of the $2 billion annual savings goal set in 2023.
With IBM's AI solutions consulting segment growing by $1.5 billion in Q3, it stands to reason that the company itself would be a major beneficiary of new AI-driven efficiencies. IBM's own success in using its AI solution systems to save billions of dollars a year for itself can only help it grow its clientele in the AI software market sector, which is forecast to hit $467 million in 2030, up from $122 billion last year.
That's an average annual growth rate of 25%, which is slow compared to the 41% average annual growth expected for quantum computing through 2030. According to the research firm Marketsandmarkets, the sector will hit $20.2 billion by then, which makes IBM's latest feat in the sector noteworthy.
In "pole position" for the global quantum computing race?
On Friday, Reuters reported that IBM had developed an algorithm to correct quantum computing errors that can run on widely available semiconductor chips. Quantum computing, which can be used to solve mathematical problems that would take conventional computers thousands of years to solve, has been prone to errors that severely restrict the technology's usefulness for now. But according to Jay Gambetta, IBM's director of research, that's changing, with IBM's correction algorithm implementation being "10 times faster than what is needed." IBM shares surged more than 7% on the news.
According to Gambetta, IBM completed the algorithm a year ahead of schedule. By 2029, it plans to build a quantum computer called Starling. It's projected to run 20,000 times more operations than today's quantum computers, and could revolutionize industries like drug development and materials discovery by simulating how molecules react with speed and efficiency that's millions of times greater than that of quantum computers.
Last month, IBM unveiled Europe's first IBM Quantum system, marking its second installation outside the United States and the company's commitment to global leadership in quantum computing. Its existing global fleet of quantum computers, and recently inked partnership with the chipmaker Advanced Micro Devices to develop "quantum-centric supercomputers," has led The Wall Street Journal to conclude that IBM is in "pole position in a new race between Big Blue, Google, Microsoft, and a horde of start-ups."
So, is IBM a buy?
From growing revenue, improving margins, and leadership in what could be the next decade-defining tech story after AI, the path for IBM to reclaim its former glory looks clear.
Despite shares hitting an all-time high this week, the stock looks reasonably valued, with a price-to-earnings ratio of 24 that is well below the S&P 500 average of 31.
The company also pays a dividend yielding 2.2%, which is larger than that of any of the Magnificent Seven stocks, and its earnings have surprised analysts on the upside for each of the last four quarters.
It's very possible that IBM continues to be an afterthought in the battle for AI market share. But its billions of dollars in annual cost savings already realized in 2025 are a reminder that the global $15.7 trillion AI revolution just might be big enough for everyone. For investors seeking income, value, and a solid chance to back a rising industry leader, IBM is a buy.