IBM was one of the earliest frontrunners in artificial intelligence. The company, known for designing some of the world’s first personal computers, built the first AI to defeat a world champion at chess, in 1997, and then the first AI to win the quiz show Jeopardy, in 2011.
In 2025, the AI frontier looks very different. The splashiest headlines often focus on AI models created by the likes of Google and OpenAI, which cost billions of dollars to train and can perform as generalists answering all kinds of questions. But IBM’s approach is different. The company has zeroed in on building smaller AI tools with a focus on reliability, and on helping its clients apply them to specific use cases.
[time-brightcove not-tgx=”true”]
IBM’s strategy speaks to one side of a debate currently raging in Silicon Valley and on Wall Street. Will the economic gains from AI mostly accrue to the few big companies investing billions in so-called “foundation models” like OpenAI and Google? Or will they instead flow to the thousands of companies that use AI to make themselves more productive?
Lately, the tide has appeared to be turning in IBM’s direction. At the end of January, the release of a highly efficient open-source AI model by the Chinese lab DeepSeek led many Wall Street analysts to conclude that large U.S. tech companies might struggle to recoup their huge investments, since similar technology to theirs would be more cheaply available elsewhere. The same week, IBM announced its most recent earnings, which showed a 10% increase in its bespoke AI software sales, beating analysts’ expectations. Its stock price jumped 12% on the news, reaching an all-time high and valuing the company at some $240 billion.
To Arvind Krishna, the former leader of IBM’s research division who has been CEO since 2020, the DeepSeek news felt like a validation of his strategy. “Smaller models, with much less compute applied to train them, can be successful,” he says in an interview with TIME. That, he suggests, might not be good news for the big tech companies at the forefront of the AI race. “I think it’s going to drive economic returns that are different—because if you spend a hundredth of the cost of training [an AI model] and you can deploy it on a much smaller infrastructure, everybody has to be competitive,” he says. “So I think it is going to put pressure on [their] economics.”
IBM, of course, isn’t just an AI company. It runs a cloud computing service, designs all kinds of software, and runs a consulting business to help clients knit them all together. It’s also a major investor in quantum computing research—the quest to build an entirely different kind of computer, based on quantum principles, which could carry out some calculations billions of times faster than existing machines. Krishna is bullish that this research will soon yield even bigger breakthroughs, saying that before 2030 he expects “we will see something remarkable happen.”
If it does, Krishna is quick to add, much of the value will accrue not only to IBM but also to its clients. But he also says that such a breakthrough could help IBM return to a dominant position in the tech industry, similar to the one that it held for much of the late 20th century as the world’s biggest PC manufacturer. “Assuming the timeline and the [quantum] breakthroughs I’m talking about happen, I think that gives us a tremendous position and the first mover advantage in that market, to a point where I think that we would become the de-facto answer for those technologies,” he says. “Much like we helped invent mainframes and the PC, maybe in quantum we’ll occupy that same position.”
This profile is published as a part of TIME’s TIME100 Impact Awards initiative, which recognizes leaders from across the world who are driving change in their communities. The next TIME100 Impact Awards ceremony will be held on Feb. 10 in Dubai.
