IBM is asking investors to treat quantum computing less like a science project and more like the next strategic chip industry. The payoff is still measured in years, but the spending is starting now.
Arvind Krishna is putting a large marker down on a technology that still makes plenty of investors nervous. IBM has pledged more than $10 billion over five years for quantum computing, while the U.S. Commerce Department has proposed $1 billion in support for a new IBM quantum chip foundry called Anderon, with IBM planning to match that amount.
That is not just another research budget. It is a statement about where IBM thinks the next computing race is moving after the first wave of artificial intelligence infrastructure. The company is telling customers, shareholders and Washington that quantum will require the same kind of manufacturing discipline that classical chips required, only with a different set of physics and a much smaller margin for error.
As Barron's recently reported, Krishna used New York Tech Week to argue that IBM still sees a quantum revolution coming, even as quantum stocks have cooled and investors have become more selective about long-term technology stories. That matters because IBM is not selling a small startup dream. It is trying to make quantum look like a credible extension of a profitable enterprise technology business.
IBM's current roadmap points to useful quantum advantage examples before the end of the decade and a fault-tolerant quantum computer by 2029. The company has already laid out plans for IBM Quantum Starling, a large-scale fault-tolerant system expected to be housed in Poughkeepsie, New York, and built around error correction techniques meant to keep fragile qubits useful for longer computations.
The promise is easy to understand. Quantum machines could eventually model chemistry, materials and biological systems in ways that classical computers struggle to handle. That is why companies in pharmaceuticals, finance, logistics and national security keep watching the field, even when commercial revenue remains small. If the technology works at scale, it does not just make existing software faster. It creates a different kind of problem-solving tool.
The caution is just as important. Quantum advantage is not the same thing as a mass-market product, and fault tolerance is not a switch that gets flipped overnight. Today's quantum processors remain sensitive to noise, and many useful workloads still depend on better algorithms, better error correction and more reliable hardware. IBM can be right about the direction and still face years of engineering work before customers see routine business value.
Anderon changes the argument
The most interesting part of IBM's plan may not be the headline investment number. It may be Anderon. The foundry is being positioned as America's first pure-play quantum chip foundry, with IBM contributing cash, intellectual property, assets and skilled workers alongside federal support. MarketWatch reported that IBM's Jay Gambetta described the shift as a move from building systems to speeding up the engineering phase, where reliable wafers and higher throughput become central.
That changes the supply chain discussion. For the last several years, U.S. semiconductor policy has been dominated by AI chips, advanced packaging and leading-edge logic. Quantum chips are different. They need specialized fabrication, control systems and cryogenic integration. If Anderon can serve IBM and other quantum builders, it becomes a shared industrial base rather than only an internal IBM project.
Washington clearly sees that strategic angle. Axios reported that the Commerce Department's broader quantum push totals about $2 billion across nine companies, including IBM and several venture-backed players such as Quantinuum, PsiQuantum, Atom Computing and Diraq. The government is spreading its bets because nobody knows which architecture will win. In quantum, superconducting circuits, trapped ions, photonics and silicon spin approaches are all still fighting for position.
For IBM, the foundry also helps answer a practical question. If quantum is going to move beyond laboratory systems, someone has to make chips repeatedly, reliably and at meaningful scale. That is not as exciting as a breakthrough demonstration, but it is often where technology markets are actually won.
The core business buys IBM time
IBM's advantage is that it is not depending on quantum revenue to survive. In its latest reported quarter, revenue rose to $15.92 billion from $14.54 billion a year earlier, adjusted earnings beat expectations at $1.91 a share, and free cash flow reached $2.2 billion, up about $300 million. Software revenue grew 11% to $7.05 billion, while infrastructure revenue jumped 15% to $3.33 billion.
That gives Krishna room to make a long bet without asking investors to ignore the present business. Red Hat, hybrid cloud, automation, data products and mainframe demand are doing the work today. Quantum is the option on what comes next.
The risk is that IBM's quantum story becomes too useful as a capital markets narrative. Every large technology company now wants to show it has a credible post-AI horizon. Quantum fits that need because it sounds enormous, strategic and still just far enough away to resist easy measurement. Investors should welcome the ambition, but they should keep asking for milestones that can be checked.
The next things to watch are simple. IBM needs to show that its quantum advantage work produces useful examples, that Anderon can attract customers beyond IBM, and that the 2029 fault-tolerant roadmap keeps moving from presentation slides into working systems. If those three things happen, IBM's quantum bet will look less like a side project and more like a serious attempt to build the next American computing supply chain.
Also read: Magnetar is putting AI bots in the analyst seat • Meta must reopen WhatsApp to rival AI assistants in Europe • Sandstone brings fresh AI money to in-house legal teams
