IBM $10 billion quantum bet targets fault tolerant machine by 2029

IBM has pledged more than $10 billion over the next five years to build what it calls the world's first large scale, fault tolerant quantum computer. The company disclosed the plan on June 2, 2026, framing it not as an experiment but as an industrial program with a delivery date attached: 2029, the year IBM says its Quantum Starling machine will come online.

The IBM $10 billion quantum bet lands at a moment when quantum computing has stopped being a purely academic pursuit and started attracting the kind of capital, and the kind of federal attention, once reserved for semiconductors and artificial intelligence. Within three weeks of IBM's announcement, President Trump signed two executive orders on quantum, and the Commerce Department steered $2 billion in CHIPS and Science Act money toward nine quantum firms. IBM sat at the center of all of it.

What the $10 billion actually buys

IBM's five year outlay is not a single line item. The company says the money will cover research and development, capital expenditure, manufacturing scale up, ecosystem partnerships and mergers and acquisitions. In practice, that means IBM is trying to build the machine, the factory to produce its chips, and the supply chain around both at the same time.

The headline deliverable is IBM Quantum Starling, targeted for 2029. IBM describes it as the first large scale, fault tolerant quantum computer, capable of running circuits with 100 million quantum gates across 200 logical qubits. The company puts that at roughly 20,000 times more operations than today's systems can reliably execute. The distinction that matters is fault tolerance: current quantum machines are noisy, and errors accumulate faster than useful computation can proceed. A fault tolerant system uses many physical qubits to encode a smaller number of stable logical qubits, which is why 200 logical qubits is a far more demanding target than any raw qubit count suggests.

Starling is being built at IBM's historic Poughkeepsie, New York campus, the same site where the company built mainframes for generations. IBM has applied to construct a facility of roughly 511,000 square feet there, a project that involves demolishing two existing buildings. The scale of the real estate is itself a signal: this is a manufacturing program, not a lab bench.

IBM $10 billion quantum bet

The IBM $10 billion quantum bet did not emerge in a vacuum. It arrived just as the federal government moved to put public money and policy weight behind domestic quantum capacity. On May 21 and 22, 2026, the U.S. Commerce Department announced $2 billion in grants and equity stakes spread across nine quantum companies, drawing on CHIPS and Science Act funds. IBM received the largest slice, about $1 billion, and matched it with $1 billion of its own cash.

That structure, federal money matched dollar for dollar by a private company and paired with equity stakes, echoes the way Washington has approached advanced semiconductor manufacturing. It treats quantum computing as strategic infrastructure rather than speculative venture, and it ties government support directly to physical build out on American soil.

The other recipients filled out a map of the U.S. quantum industry. GlobalFoundries took $375 million. Roughly $100 million each went to D-Wave Quantum, Rigetti Computing, Infleqtion, Atom Computing, PsiQuantum and Quantinuum, with $38 million going to Diraq. The spread across superconducting, photonic, trapped ion and neutral atom approaches suggests the government is hedging across competing technical bets rather than crowning a single winner.

Anderon and the pure play quantum chip foundry

Part of IBM's federal award funds a new venture called Anderon, billed as the first pure play quantum chip foundry in the United States. Headquartered in Albany, New York, Anderon is set to run a 300mm quantum wafer fabrication facility, the same wafer size that anchors modern semiconductor manufacturing.

The word "foundry" is deliberate. In the classical chip world, a foundry manufactures chips designed by others, and the emergence of dedicated foundries was what turned semiconductors from a vertically integrated business into a global industry. A pure play quantum foundry implies the same ambition for qubits: a facility that can fabricate quantum processors at volume and to a repeatable standard, rather than hand building one device at a time.

Locating both Starling and Anderon in New York State concentrates a quantum manufacturing corridor in the Hudson Valley and Capital Region. Poughkeepsie handles the flagship system; Albany handles the wafers. Together they represent an attempt to keep the most sensitive parts of the quantum supply chain, the chips themselves, inside the country.

Two executive orders reshape the policy backdrop

On June 22, 2026, President Trump signed two executive orders on quantum at the White House. The first launched a national push to build a scientific research quantum computer at a Department of Energy facility by 2028, and to field Pentagon quantum sensor projects on the same 2028 timeline. The second ordered federal agencies to begin migrating to quantum resistant encryption by the early 2030s.

That second order speaks to a threat sometimes called "harvest now, decrypt later." Adversaries can collect encrypted data today and store it, betting that a future quantum computer will be able to break the encryption protecting it. Ordering agencies to move to post quantum cryptography is a defensive counterweight to the same technology IBM is racing to build.

The signing itself functioned as a show of industry alignment. IBM CEO Arvind Krishna attended alongside Alphabet President Ruth Porat, Nobel winning physicist John Martinis, Commerce Secretary Howard Lutnick and Energy Secretary Chris Wright. Krishna said "sound policy, sustained investment and public private partnership are vital" to U.S. quantum leadership, a line that reads as a summary of the whole strategy on display.

IBM's existing quantum fleet and customer base

IBM is not starting from a standing position. The company says it already operates more than 90 quantum computing systems worldwide, a fleet it claims is larger than all competitors combined. Those machines serve a network of more than 325 to 340 Fortune 500 companies, startups, universities and government agencies.

That installed base is arguably IBM's strongest argument. Building a fault tolerant machine is a hardware problem, but making quantum computing commercially useful is also a software, tooling and workforce problem. A company with dozens of systems already running, and hundreds of large customers already writing code against them, has a running start on the parts of the ecosystem that cannot be bought with capital expenditure alone.

It also gives IBM a credibility that matters to policymakers weighing where to place public funds. When the Commerce Department chose to route its single largest quantum award to IBM, the existing fleet and customer network were part of what it was buying: not just a promise of a 2029 machine, but a company already delivering quantum access at scale.

Quantum stocks rally on the funding news

Investors treated the news as a catalyst. IBM shares rose roughly 5% following the June 22 executive orders, and had already climbed about 1.7% to 4% in the days after the June 2 investment announcement. The moves were part of a broader quantum stock rally that also lifted GlobalFoundries, D-Wave, Rigetti and Infleqtion.

Rallies driven by policy and pledges carry their own risk, since a delivery date in 2029 gives markets years in which enthusiasm can outrun results. But the fact that the rally spread across the sector, rather than concentrating in IBM alone, suggests investors read the combined federal funding and executive orders as validation of quantum computing as a category, not just a bet on one company's roadmap.

For IBM specifically, the market reaction rewards a strategy of matching its own capital to public money and attaching its name to a hard, public deadline. That deadline is also the risk. Having told the world it will deliver a fault tolerant machine by 2029, IBM has given itself a benchmark against which every future quarter will be measured.

Closing the technical gap to 200 logical qubits

The gap between today's noisy machines and Starling is enormous, and it is worth being clear eyed about it. Running 100 million gates across 200 logical qubits requires error rates and error correction schemes that no one has yet demonstrated at that scale. IBM's own framing, roughly 20,000 times more operations than current systems, is a measure of how far the company still has to travel.

What separates this effort from earlier quantum optimism is the surrounding apparatus. A dedicated 300mm foundry in Albany addresses the manufacturing repeatability problem. A 511,000 square foot facility in Poughkeepsie provides the physical space for a large system and its cryogenic infrastructure. Federal funding and executive orders supply demand signals and a policy runway. None of that guarantees the physics will cooperate, but it removes several of the non technical excuses for failure.

The IBM $10 billion quantum bet is ultimately a wager that quantum computing has reached the stage semiconductors reached decades ago, where progress depends less on a single breakthrough and more on industrial execution: building factories, training people, securing supply chains and hitting deadlines. Whether Starling arrives on schedule in 2029 will be the clearest test yet of whether that thesis is right.