Intel Buys the Quantum Startup Its CEO Already Backed
Intel has bought a company its own chief executive already owned a piece of — and then, pointedly, stepped aside from negotiating over. That is the strangest and most telling detail of the chipmaker's confirmed move into quantum computing, announced on 26 October 2026: a definitive agreement to acquire Tel Aviv-based quantum software firm Classiq, a deal Israeli press reports at about $2.3 billion in cash and stock. Intel itself has disclosed no price at all, and that gap between the announcement and the reported number is where this story actually lives.
Strip away the press-release language and the logic is simple. Intel has spent more than a decade trying to build quantum hardware out of silicon. Classiq builds the software that tells quantum machines what to do without requiring the user to be a quantum physicist. Put the two together and Intel can finally sell something resembling a complete quantum computer — hardware, software and the factory that makes the chips — rather than a promising component in search of a customer. Whether Intel, of all companies, can hold that combination together is a much harder question, and its own history gives sceptics plenty of ammunition.
The deal, in plain terms
Intel announced on 26 October 2026 that it had entered into a definitive agreement to acquire Classiq, described in the announcement as "a leading provider of quantum computing software". The combination, Intel said, would pair Classiq's synthesis platform with its spin-qubit manufacturing and existing software to "create a user-friendly full-stack quantum computing platform, enabling the promising technology to mature into a mainstream computing business".
The transaction is subject to regulatory approvals and other customary closing conditions, and is expected to close in the first half of 2027. Until then, the two companies will continue to operate as separate, independent companies. Intel also promised that after closing it would keep supporting Classiq's existing multi-platform integrations, keeping the software "modality- and hardware-agnostic" — in other words, still usable on rivals' machines, not just Intel's own.
That last promise matters more than it sounds. Classiq's whole pitch is that a bank, a carmaker or a hospital should not have to bet on one kind of quantum hardware before the hardware race is settled. If Intel quietly turned the software into a funnel for its own chips, it would destroy the neutrality that made Classiq valuable. Watch whether that pledge survives contact with Intel's sales targets. It is the easiest promise in this announcement to make and the hardest to keep.
We checked Intel's announcement itself, line by line: it contains no purchase price, no revenue figure for Classiq and no timetable for a combined commercial product. Everything in the next section about money comes from reporting, not from Intel — and readers should hold the two categories apart.
The $2.3 billion number Intel never said
Calcalist's Ctech and Haaretz both report the deal at about $2.3 billion in cash and Intel stock, and Bloomberg, as carried by Reuters on 27 October, reported the same figure citing people familiar with the matter. Reuters noted that it could not independently verify the report, that Intel declined to comment on the reported value, and that Classiq did not immediately respond to its request for comment. Globes, meanwhile, reported $2.2 billion. The honest summary is this: roughly $2.3 billion, reported, not confirmed — and a $100 million spread between Israeli outlets that should stop anyone quoting the figure as gospel.
If the reported figure is right, it would be the largest acquisition in quantum computing to date, by a wide margin. The previous record was IonQ's $1.075 billion acquisition of Oxford Ionics, announced in June 2025 and completed in September 2025. Intel would be paying more than double that, and — again on reported figures — more than three times Classiq's roughly $600–700 million valuation at its 2025 funding round. That is either a confident bet on a platform, or a very expensive way of admitting you cannot build the software layer yourself. Possibly both.
The reported spoils, all via Ctech and Ynet, are striking for a company of about 150 people: employees collectively hold roughly 10% of Classiq, worth around $230 million at the reported price, Intel has reportedly set aside more than $100 million in retention grants, and each of the three founders is expected to receive between $80 million and $100 million. Those are reported figures, not Intel disclosures, but they explain why a retention package of that size exists at all: in a 150-person company, losing a dozen key engineers is not attrition, it is an amputation.
What Classiq actually does
Classiq was founded in Tel Aviv in 2020 by Nir Minerbi, its chief executive; Amir Naveh, its chief product officer; and Yehuda Naveh, its chief technology officer and Amir's father. Minerbi and Amir Naveh are graduates of the Israel Defense Forces' Talpiot programme; Yehuda Naveh previously led IBM's quantum computing programme in Israel. There is something faintly wonderful, and very Israeli, about a father-and-son founding team selling a company to Intel — and about Minerbi, 35, describing the founding problem with unusual candour.
"We founded Classiq to solve an almost embarrassing problem: no one knew how to communicate with a quantum computer," Minerbi told Ctech. "It was a gamble, because quantum computers barely existed at the time."
The product is, in essence, an abstraction layer. A developer describes what a quantum programme should do in high-level terms; Classiq's platform synthesises the optimised circuit that a quantum computer actually runs, and it can target different hardware. Intel's announcement reaches for the grandest possible analogy, saying Classiq plays a similar role in quantum computing to the one electronic design automation (EDA) and computer-aided design (CAD) software played for classical chip design — "providing a common software platform that could abstract the ability to design chips from the complexity of the underlying hardware".
That analogy is not accidental, and it leads directly to the deal's most awkward subplot. Lip-Bu Tan, Intel's chief executive, used to run Cadence Design Systems, the EDA giant. In a September interview on the QuantumJane podcast, he described Classiq as an attempt to build the "next Cadence" for quantum computing. He had also put his own money into the company long before he ran Intel, investing in its $33 million Series B round in February 2022 alongside chip-design veteran Harvey Jones, and remaining among its backers through the $110 million Series C in May 2025.
The CEO who backed it, then recused himself
"I am delighted to have Nir, Yehuda, Amir and the entire Classiq team join Intel," Tan said in Intel's announcement. "For more than a decade, Intel has made remarkable progress in quantum computing by applying its semiconductor innovation, advanced manufacturing prowess and systems engineering expertise to the development of scalable silicon spin-qubit technology. Joining forces with Classiq transforms Intel's Quantum Computing Group into a full-stack quantum computing business."
Minerbi struck a matching note: "Since our founding, Classiq has focused on making quantum computing practical and accessible by enabling developers, researchers and enterprises to create scalable quantum applications without requiring deep expertise in quantum circuit design. Together, Intel and Classiq can help unlock the benefits of quantum computing and accelerate its adoption sooner."
What Intel's announcement does not mention is Tan's personal stake. Haaretz describes it as relatively small, and notes that he does not sit on Classiq's board. According to sources close to Classiq cited by Ctech and Ynet, Tan recused himself from the acquisition discussions because of his financial interest, and did not participate in the negotiations. That is the correct behaviour, and it is also the minimum the situation requires. A chief executive buying a company he personally backed is the kind of fact that should be volunteered in the announcement, not left for Israeli business pages to surface. Intel's silence on it is a small thing, and it is not a good look.
We verified each of the internal links in this article against the live blog before writing — they all return, rather than merely being assumed to exist — because a quantum software story that ships with broken links would be an irony too far.
Why Intel wants the software layer now
Intel says it has been investing in quantum computing since 2015, and claims it is "the only company capable of producing qubit chips and complex cryogenic control electronics at-scale, using equipment and processes common to conventional silicon-based semiconductor manufacturing". Its bet is on silicon spin qubits — information stored in the spin of single electrons held in silicon — which, if it works, could be manufactured in something resembling an ordinary chip fab rather than a bespoke physics laboratory.
The reported hardware track record is real but modest: Intel's Tunnel Falls, a 12-qubit silicon research chip introduced in 2023, is being used by researchers at Argonne National Laboratory to study qubit interactions, according to TechStrong. Twelve qubits is not a computer that changes the world; it is a laboratory instrument. That is precisely why the software matters. Nobody buys a 12-qubit chip. People might buy a platform that lets them write one programme and run it on whatever hardware eventually wins.
Classiq's customer list, as reported by Ctech and in the company's own 2025 funding announcement, includes BMW, Rolls-Royce, HSBC, Deloitte, Toshiba and Citi, with Mizuho also named by the company. The reported use cases are concrete rather than futuristic: BMW simulating car aerodynamics, HSBC building a fraud-detection algorithm, and Italy's Intesa Sanpaolo pricing options over shorter timeframes, according to SiliconANGLE. Ctech also names Ichilov Hospital and the Bank of Israel as users. Revenue, by contrast, is small and the reported figures disagree with each other — Ctech puts annual revenue in the tens of millions of dollars, Ynet says it remains below $100 million, and Haaretz cites an annualised run rate of several million dollars last year that has since grown more than fivefold. Treat any single revenue number in this story with suspicion; the range itself is the fact.
Classiq has raised about $200 million in total, according to Ctech and Reuters. The $110 million Series C in May 2025, led by Entrée Capital, brought the total to $173 million at the time; a subsequent extension, including a strategic round reported in November 2025 that added AMD Ventures, Qualcomm Ventures and IonQ, took it past $200 million. Until recently, Ynet reports, Classiq had been considering a Wall Street listing through a SPAC merger. The Intel deal, if it closes, is a cleaner exit — and a much richer one than a listing at the 2025 valuation would likely have delivered.
The ghost at the table: Habana and Granulate
Here is where the celebration should stop. Intel's record with Israeli acquisitions is, to put it bluntly, shoddy. It paid about $2 billion for Habana Labs in 2019, an AI chipmaker that should have given it a fighting chance against Nvidia; Haaretz and Ctech both report that Habana's operations were effectively dismantled, and Ynet's account has nearly all of its original founders, managers and engineers gone by 2024, most of them leaving as soon as their four-year retention period ended. It paid about $650 million for Granulate in March 2022; Ctech reported in October 2024 that Intel was shutting Granulate down entirely and laying off its workforce of more than 100 people. Its $15.3 billion purchase of Mobileye in 2017 — still Intel's largest Israeli deal, and the yardstick that makes Classiq its second-largest — ended with Mobileye taken public again in October 2022 at a $16.7 billion valuation: above what Intel had paid, but far below the $50 billion it had reportedly hoped for, with Intel retaining voting control.
None of that proves Classiq will fail. It proves that buying clever Israeli engineering is a skill Intel has repeatedly failed to demonstrate, and that retention grants alone do not keep founders who have just banked $80 million-plus. The reported plan for Classiq to become Intel's global development centre for quantum computing — reported by SiliconANGLE, not stated in Intel's own announcement — would at least give the team a centre of gravity rather than absorption into a business unit. But "centre" is a word Intel has used before, and Habana's people know how that story ended.
There is also the regulatory shadow. The deal needs approvals, and Intel's planned $5.4 billion purchase of Tower Semiconductor collapsed after regulators, including authorities in China, declined to approve it. A quantum deal involving an Israeli company, an American chipmaker and technology with obvious dual-use sensitivity is not going to glide through every jurisdiction unexamined. The first half of 2027 is a target, not a promise, and Intel's own announcement says the transaction "may not be completed in a timely manner or at all".
Our take: the right bet by a company that has to prove it can keep one
This is, on the merits, the most interesting acquisition Intel has made in years — and also the one its history least entitles it to be trusted with. The strategic logic is genuinely sound: quantum computing's bottleneck has never been only physics, it has been usability, and whoever owns the layer that lets ordinary developers build quantum programmes owns the toll road whatever hardware wins. Intel buying that layer, rather than trying to grow it inside a hardware culture, is a recognition of its own limits that deserves credit. It is also, at a reported $2.3 billion for a company with small and disputed revenue, a price that only makes sense if Intel believes the software layer becomes the platform — the Cadence, the Windows, the thing everything else runs through.
But belief is cheap and integration is expensive. Intel missed mobile, then watched the AI boom from the sidelines while its own AI chip bet withered; buying its way into the next platform is not the same as earning it, and the graveyard of Habana and Granulate says the hard part starts the day after closing. If Tan — the Cadence veteran who spotted this company four years before he ran Intel — can keep a 150-person Tel Aviv team intact, independent-minded and hardware-neutral inside a 100,000-person chipmaker, this could be the deal that finally breaks Intel's acquisition curse. If he cannot, Classiq will join a familiar list, and the $2.3 billion will read, in hindsight, as the price of another lesson Intel paid for but did not learn. The next eighteen months, through regulatory review and into the first year after closing, will decide which story this was.
The wider race, meanwhile, does not pause for Intel. IonQ is building its own full stack from the opposite direction, IBM's Qiskit remains the incumbent software platform Classiq must displace in developers' minds, and Microsoft and Google are chasing practical machines of their own. Intel's move raises the stakes for all of them — which is why, even with the price unconfirmed and the close more than a year away in the worst case, this is a deal worth covering now rather than waiting for the paperwork. Quantum computing has waited decades for its mainstream moment; Intel has just paid a reported fortune to be standing in the right place when, or if, it arrives.
Related coverage on TechPulse USA: Oracle's AI-cloud expansion shows how the cloud giants are hedging their hardware bets; Microsoft's Maia 300 bet on custom AI silicon; and how Claude redesigned quantum circuits with far fewer gates.
Intel's Classiq deal, answered
Has Intel confirmed the $2.3 billion price? No. Intel's 26 October announcement discloses no price. About $2.3 billion in cash and stock is reported by Ctech, Haaretz and Bloomberg (via Reuters); Globes reported $2.2 billion. Reuters could not independently verify the figure.
When does the deal close? Expected in the first half of 2027, subject to regulatory approvals and customary closing conditions. Until then the companies operate independently.
What does Classiq's software actually do? It lets a developer describe what a quantum programme should do in high-level terms and synthesises the optimised circuit a quantum computer runs, across different hardware types — Intel's EDA/CAD analogy for quantum.
Will Classiq's software still work on rivals' machines? Intel promised to keep supporting Classiq's existing multi-platform integrations after closing, keeping it modality- and hardware-agnostic. Whether that pledge survives Intel's sales targets is the open question.
Why should anyone trust Intel with this acquisition? On the merits the logic is sound — usability is quantum's bottleneck — but Intel's record (Habana effectively dismantled, Granulate shut in 2024, Tower's $5.4B deal collapsed on regulatory approval) is the honest counterweight. The next eighteen months decide it.
Sources: Intel's 26 October 2026 announcement (Intel investor relations / Business Wire); Dow Jones Newswires via Morningstar, 26 October 2026; Reuters, 27 October 2026 (Bloomberg report on the $2.3 billion figure); Haaretz, 26 October 2026; Calcalist's Ctech, 26–27 October 2026; Ynet, 26 October 2026; Globes, May 2025 (Series C); Dealroom, October 2026; SiliconANGLE, 26 October 2026; TechStrong, 27 October 2026. The $2.3 billion price, founder and employee proceeds, retention grants and the 2025 valuation are reported figures, not Intel disclosures.
By Daniel Rao. Daniel covers semiconductors, AI infrastructure and the business of computing for TechPulse USA.
Comments
Post a Comment