Xanadu Quantum Technologies (XNDU) delivered a third-quarter report that missed on every line the market can measure. Meanwhile, the progress management wants investors to focus on cannot yet be measured in dollars. The company posted a loss of $0.11 per share against a consensus estimate of -$0.09, a 22.2% miss. Revenue of $1.31 million fell 53.9% short of the $2.83 million consensus. That is the central tension of the quarter. Partnerships, government funding and component-level loss reduction all improved meaningfully. Revenue, however, shrank again and spending climbed faster than many investors likely expected.

The quality of the headline deserves scrutiny in both directions. The company reported 32% year-over-year revenue growth, from $987,000 to $1.305 million, driven by DARPA Quantum Benchmarking Initiative work and partnership engagements. The sequential picture is worse: revenue slid from $2.8 million in Q1 to $1.5 million in Q2 and $1.3 million in Q3. On the bottom line, the per-share loss looks far smaller than the year-ago $0.36. That comparison is distorted by the reverse recapitalization, which lifted weighted shares from roughly 55.8 million to 304.9 million. In absolute terms, the net loss widened to $32.2 million from $20.3 million. The adjusted EBITDA loss grew to $26.3 million from $15.4 million a year ago and $21.3 million in Q2. R&D expense rose 67% to $25.6 million, up from $19.7 million in Q2. G&A more than doubled to $9.2 million, including stock-based compensation that jumped to $5.5 million from under $1 million. Some cushioning came from $1.5 million of other operating income and $2.3 million of interest income on the post-listing cash pile.

The call made clear that the spending is deliberate rather than a loss of control. Headcount reached 361, up from 306 in Q2 and 277 in Q1. Silicon nitride wafer runs have doubled year-to-date, and thin-film lithium niobate runs rose 15% sequentially. The resulting technical gains were concrete:

• Electro-optic switch loss fell to a record 7 millidecibels from roughly 20 a year ago.
• Homodyne detector loss improved twofold to 77 millidecibels.
• Fiber delay line loss fell more than threefold year-to-date.

Management reaffirmed its roadmap of up to 200 logical qubits by 2029, up to 500 by 2030, and more than 1,000 by 2031. Management's tone was noticeably more confident than last quarter, and the company kept the faster cadence of design cycles that the CEO framed as the attack on loss.

The catalyst list lengthened considerably. The headline partnerships include:

• **GlobalFoundries:** a multi-year partnership to produce silicon nitride chips and SNSPD detectors on a 300mm line.
• **ASML:** a lithography collaboration aimed at waveguide loss, the company's largest expected loss source.
• **Bluefors:** a cryogenics partnership.
• **AMD:** the launch of Backline, with sub-3-microsecond classical round trips.

Canada's roughly $140 million agreement for the Inception facility moved from "final discussions" in Q2 to a signed deal. DARPA provisionally approved up to $10 million for a QBI extension. The company also booked its first commercial software revenue through RBC, Booz Allen and a JSR study.

Bears still have legitimate ammunition. The Canadian money is structured as a non-dilutive loan and is reimbursed as qualifying spend occurs, not received upfront. DARPA Stage C commentary softened from "looking good" in Q2 to no new update. Management will now report the aggregate system-loss metric only annually, which reduces visibility. There is also a timing discrepancy. The release references a year-end 2026 system loss target of 10.2, while the call clarified that the milestone means the beginning of 2027. That is a tight, near-term checkpoint the company will be judged against.

The balance sheet buys time but not indefinitely. Cash fell to $286.5 million from $312.8 million, and capex rose to $8.4 million. Management guided Q4 costs and capex higher still. The company did not draw on its synthetic ATM during the quarter, but $232.8 million remains available for opportunistic use, so dilution is an ongoing consideration.

Investor sentiment entered the report modestly negative at -0.23, suggesting expectations were already guarded. The stock's AVWAP trend reading also remains negative. Both readings fit a market that has been reluctant to pay up for roadmap milestones while near-term financials deteriorate.

The bottom line is that Xanadu's quarter splits cleanly into two stories. On the reported numbers, it was a clear miss, with shrinking sequential revenue, widening losses and accelerating burn. On execution against its technical and funding agenda, it was arguably the company's strongest quarter as a public company. For now, the investment case rests almost entirely on the beginning-of-2027 loss milestone, the DARPA Stage C decision and how much equity the company needs to fund the build-out.