In late September 2026, on a stage in Toronto at the festival that had been called Elevate until its opening night, Christian Weedbrook offered the room some career advice. “The take home is don’t go to film school,” said the founder and chief executive of Xanadu, who failed the course twice before he sold heat packs at a flea market. The audience laughed, as he meant it to. Earlier that week he had sold almost a tenth of his shares in the company. Christine Li, the BNN Bloomberg anchor interviewing him, wanted to talk about that as well.

Weedbrook’s story is one of stubbornness rewarded. He describes a grade average of “pretty much zero”, a side door into mathematics through an arts degree, and a PhD in which he was “never the best”. In 2016 he founded a company to build quantum computers out of light, at a time when, he says, that was “not a thing”. Each time Li pressed him on the share price, his own sale or the new shares the company keeps issuing, he came back to the same idea. “Take a 10, 20-year approach,” he said of the stock.

It is sound advice for a founder, and it has served this one well. The difficulty is that Xanadu chose the public markets, and those markets keep a much shorter clock than any founder. Since its shares began trading on 27 March, that clock has been set by investors who took their money back before the listing, by a lock-up that ended and by a share facility drawn on while the price was high. The patience Weedbrook asks of his shareholders is patience that the structure he picked does not always reward.

He has a strong case for picking it. Xanadu’s private rounds grew from a $9 million seed in 2018 to a $100 million Series C in 2022 that made it a unicorn, a startup valued at $1 billion or more. “It never got any easier,” Weedbrook said, and for the Series B he recalls “438 nos” before one lead investor said yes. The public route was another world. He says the company raised its $275 million private placement, shares sold to chosen investors alongside the listing, “within four weeks”.

He was frank about why Xanadu merged with a SPAC, a listed shell company that raises cash first and finds a business to merge with later, in place of a traditional flotation. “The SPAC allows you to go public without a real product, without real revenue,” he told the room. “We have single digit millions.” The filings agree, with revenue of $2.8 million in the first quarter of 2026 and $1.5 million in the second. An ordinary listing would have asked harder questions of those numbers. For a company that burns capital the way quantum hardware does, the SPAC was a door built for exactly its shape.

The price of that door is that SPAC investors may take their money back before the merger closes. Xanadu’s deal was planned to raise about $500 million, with $225 million from Crane Harbor’s trust and $275 million from the private placement. It closed with about $302 million. Weedbrook put the redemptions at 88 per cent, saying “we only got to keep 12% of the money”, and blamed markets that turned fearful in the fortnight before the listing. What remained was a small pool of freely traded shares, and small pools move sharply. “It’s been a wild ride since going public, particularly with the share price,” he conceded.

The lock-up, the period after a listing in which insiders may not sell, ran for six months. On 22 September, the day the festival opened, Weedbrook sold 4.6 million shares on the Toronto Stock Exchange and other markets for about C$32.8 million, “for personal financial planning purposes”, according to his early warning filing. On stage Li put the sale at 33 million shares, which appears to confuse the proceeds with the share count. Weedbrook’s own description was closer. “I sold 10% of my shares,” he said. “I still have 90%, so I’m obviously all in still.”

Nobody reasonable would grudge him that. He said that until the sale, with the company now in its 11th year, “100% of my wealth was in Xanadu, and it’s illiquid”, and the filing shows he still controls about a quarter of the votes. Founders who never take money off the table are rarer than founders who say they never will. The point lies in the timing, which belongs to the structure. The lock-up ended that same day for the early shareholders who held more than 95 per cent of the stock, and The Globe and Mail reported the sale under the headline that he had sold as the lock-up lifted. He told the paper he does not plan to sell more.

Then there is the facility. In May Xanadu agreed with Yorkville Advisors to sell up to $300 million of new shares over three years, and in the second quarter it raised $67.2 million by selling 5.5 million shares at an average net price of $12.28. When Weedbrook sold in September, his shares went for between US$5.01 and US$7.27 in the United States. The company, in other words, issued stock at roughly twice the price its founder received a few months later. That is good treasury work, and the finance chief has said Xanadu will draw on it only when it believes “conditions are favorable to the Company, including its shareholders”. It is the investors who end up holding those new shares who carry the swing.

Weedbrook’s answer to all this is the long view again. “If quantum computing plays out the way we and others think,” he said, the dilution “will be not meaningful compared to the true value of this industry and hopefully Xanadu.” He says he loves reading about Warren Buffett and Charlie Munger, and the argument borrows their patience for a company with little revenue. The long view turns every short-term cost into a rounding error. It does so only if the long term arrives on time.

That brings the conversation to the clock Xanadu has set for itself, and to a question with a much wider audience than its shareholders. By the end of August the company had published a roadmap with 200 logical qubits in 2029, 500 in 2030 and more than 1,000 in 2031. A logical qubit is a reliable unit of quantum information built from many fragile physical ones, so that errors can be caught and fixed. Li quoted the target as 500 by 2029, a year ahead of the published plan. Weedbrook let it pass, calling the goal “a stretch goal”, and adding, “It is possible. We’re still on time.”

Li then observed that about 1,200 logical qubits would be enough to break Bitcoin. That figure matches a whitepaper posted in March 2026 by researchers at Google Quantum AI, written with co-authors at Stanford University and the Ethereum Foundation. It found that Shor’s algorithm, the method for cracking the codes behind Bitcoin’s signatures, could run on fewer than 1,200 logical qubits and fewer than 90 million Toffoli gates, a kind of basic logic operation. The same paper sorted future machines by clock speed. It placed photonic machines, Xanadu’s kind, among the fast designs that could attack a payment in the minutes before it is confirmed.

Weedbrook’s reply was dry and sensible. “You can’t really use that as a business model,” he said, since no investor funds a plan to break the law. New codes that a quantum computer cannot crack are coming, he added, but “you can’t be complacent”, and “banks and other institutions need to move as fast as possible.” Of the wider internet he said simply, “I still think it needs to move much faster.”

Here his long view meets everyone else’s short one. A machine at Xanadu’s 2031 target would sit in the range the whitepaper describes, although logical qubits are not interchangeable and error rates and gate counts decide what a machine can run. Nobody is suggesting Xanadu would attack anything. The point is that its roadmap is also a deadline for other people. Every year it slips buys time for banks and Bitcoin developers, and every year it holds takes that time away.

So the company’s two audiences want opposite things from the same schedule. Shareholders need the roadmap to hold, because the long view only pays if the machine arrives. The banks, governments and Bitcoin developers replacing the world’s encryption should plan as though it will hold, whatever any founder says from a stage about how far away the threat may be. Weedbrook, who told the room there was no worry “at this stage”, also told banks not to wait.

Even the proof he offered of quantum speed needed a footnote. Weedbrook told the room that Borealis, Xanadu’s 2022 machine, did in two minutes what the world’s fastest supercomputer would need about seven million years to do. The Nature paper that reported the result gives a different comparison. It says the best classical methods would need more than 9,000 years to produce a single sample of the kind Borealis produced in 36 millionths of a second. The task, known as Gaussian boson sampling, was chosen because it is hard for ordinary computers, and the random instance Borealis ran answers no practical question. Weedbrook framed it the same way, calling it “a very tough math problem” of the kind that “you start off with”.

None of this makes him wrong about the destination. Quantum hardware needs more capital than private rounds supply, and Xanadu now has a large cash pile, $312.8 million at the end of June, to spend on getting there. What it changes is who bears the wait. A founder with a quarter of the votes can afford ten years. A shareholder who bought at the facility’s prices may not have that luxury.

Asked what drives him, Weedbrook told a story about a visitor from the future who promises he will become a billionaire if he sells shoes. He would refuse, he said, “because I don’t care about shoes.” It is a fine answer from a man who has spent a decade on one very hard problem, and it is hard to doubt that he means it. His shareholders were promised nothing that certain, and their clock runs faster.

This feature draws on a fireside chat at nrth 2026 in Toronto, in which BNN Bloomberg’s Christine Li interviewed Christian Weedbrook. Watch the full conversation.

More like this