Latest / Quickly Quantum / July 26, 2026 — Who Owns a Quantum Algorithm?
Transcript
- Today on Quickly Quantum: if a mathematical formula can't be patented, and a quantum algorithm is, underneath everything, a mathematical formula, then what is IBM actually buying when it folds in a company holding more than eleven hundred patents? That question has been sitting quietly under this week's biggest quantum deal, and almost nobody covering the deal itself stopped to ask it. So today we're doing something a little different — one subject, the whole episode: who actually owns a quantum algorithm, and can American patent law even answer that question right now? Welcome back to Quickly Quantum, your daily brief on the quantum frontier. It's Sunday, July 26, 2026. Let's get into it. Quick recap for anyone who missed Thursday and Friday: IBM is acquiring HRL Laboratories, the old Boeing and General Motors research shop, mainly for its silicon-spin qubits — a different way of building a quantum bit that could pair with IBM's existing superconducting chips. What didn't make the headlines is what's riding along in that deal: more than eleven hundred patents, built up over decades of defense and aerospace contract work. And that number is what makes an obscure corner of patent law suddenly very expensive. Here's the problem in plain English. The Supreme Court ruled, in a 2014 case called Alice versus CLS Bank, that you cannot patent an abstract idea — and a bare mathematical algorithm counts as an abstract idea, full stop. That doctrine goes back further, to a 1972 case, Gottschalk versus Benson, where the Court said you couldn't patent a method for converting decimal numbers into binary, because that's just math wearing a computer's clothes. Bilski versus Kappos in 2010 tightened the screws further, and Alice, in 2014, gave us the test courts still use today: does the claim add an 'inventive concept' on top of the abstract idea, or is it just the abstract idea dressed up in legal language? Now here's why that's a genuine crisis for quantum, and not just a lawyer's headache. Nearly every commercially valuable thing in this field — Shor's algorithm and its variants, error-correction codes like the surface code — is, at its core, a mathematical procedure. Courts have generally been comfortable patenting the hardware: a physical qubit, a cryostat, a control chip. Software-only claims, the actual algorithm, get brutalized. And the numbers show who's built around that reality already — IBM alone holds 4,388 patent families, more than one-point-eight times second-place Google's 2,385, with Microsoft, China's Origin Quantum, and D-Wave rounding out a top five that controls nearly half of all patent families in the field. Searching arXiv for 'quantum' turns up more than thirteen thousand results, which tells you how much of this field's actual innovation is happening in the open literature rather than behind a patent wall, and how hard that makes life for a patent examiner trying to figure out what's genuinely new. So when HRL's eleven hundred patents move into IBM's estate, the question isn't just how many patents — it's whether the underlying law can even tell you which of those patents deserve to exist. So let's walk the sides here, starting with the people who think this uncertainty is actively hurting the country. Senators Chris Coons and Thom Tillis, along with Congressman Kiley, have introduced — and reintroduced, three times now, in 2023, 2024, and again this year as the current bill in the Senate — the Patent Eligibility Restoration Act, which would legislatively overturn Alice's abstract-idea exclusion. Andrei Iancu, who runs the Council for Innovation Promotion, argues that Congress hasn't meaningfully touched what counts as patentable subject matter since seventeen ninety-three, the original Patent Act, and that leaves twenty-first century technologies like quantum, AI, and biotech without a clear framework to build on. Kiley's version of the argument is more competitive: American inventors, he contends, are working under tighter rules than rivals like China, who already allow broader software and algorithm patenting. The bill has bipartisan sponsorship. It's had hearings, including as recently as October of last year. And it has gone exactly nowhere three separate times. That's not a small detail — it's the whole story of this camp's frustration. Their argument isn't really about any one company or any one case; it's that capital-intensive technologies like quantum hardware need predictable IP protection to justify the kind of investment we're seeing right now, and right now, predictability is the one thing Section 101 doctrine cannot offer. A company deciding whether to spend real money developing a new error-correction scheme has no reliable way to know, going in, whether a patent examiner, or a federal judge two years later, will treat that scheme as a genuine technical advance or as dressed-up math. That's the reform camp's pitch: fix the statute, because courts keep applying it inconsistently, and every year Congress doesn't act is a year quantum companies are building patent strategy on sand. And it's worth noting who's cheering hardest for this bill — mostly large, patent-dense incumbents who'd benefit most from clearer, broader claims, which is exactly the kind of detail that makes the opposing camp nervous. Now, camp two doesn't think we need to wait for Congress at all — and their evidence is a single, fairly obscure ruling that quietly became a big deal in patent circles. Attorneys tracking Patent Office practice at firms like Goodwin, Foley Hoag, and Maschoff Brennan point to a February 2025 decision from the Patent Trial and Appeal Board, a case called Ex parte Yudong Cao. An examiner had rejected a patent application for a hybrid quantum-classical method — an algorithm that splits work between a quantum processor and a regular computer — for solving systems of linear equations, on the grounds that it was just an abstract mathematical idea. The Board reversed that rejection. Why? Because the claimed method specifically let noisy, limited quantum computers — the kind we actually have today, before full error correction — practically solve a useful class of problems they otherwise couldn't touch. The Board found that controlling the qubits according to specific circuit parameters was the actual focus of the invention, and that's what integrated the abstract math into what the law calls a 'practical application.' This might be the Patent Office's first real, on-the-record ruling that tells you something concrete about how quantum algorithms get treated, and the takeaway these attorneys draw is genuinely useful if you're an inventor: a wide range of quantum algorithms can clear Section 101, as long as you tie the claim clearly to how it improves a real quantum computer's ability to solve a specific problem, rather than just describing math in the abstract. Draft it right, and the eligibility door is open right now, without waiting on a bill that's died three times in Congress. But — and every attorney making this argument adds this caveat — one favorable PTAB decision does not settle a doctrine. A witness at a 2025 Senate hearing put it plainly: eligibility law is still, in that person's words, inconsistently applied by federal courts. A single win at the Board doesn't bind the Federal Circuit, and it certainly doesn't bind a district court hearing a validity challenge five years from now, once actual licensing revenue is on the line. So this camp's confidence comes with an asterisk: the path exists, but it's a path built one careful application at a time, not a settled highway. Camp three is skeptical of the entire push to loosen things up, and their worry has a name: the software patent thicket of the nineties and two-thousands, when vague, overbroad claims on 'a computer that does X' clogged the system and chilled competition instead of rewarding real invention. This camp draws on two precedents cited in a guest analysis published on The Quantum Insider back on July 18th — American Axle and Symantec — both cases where courts made clear that being useful, containing a technological advance, or even including some specific hardware in the claim is not, by itself, enough to make an invention eligible for a patent. Here's where that lands for quantum specifically. Take a surface code, one of the leading approaches to quantum error correction, where you spread the information from one reliable 'logical' qubit across many noisy physical qubits arranged in a grid, so errors can be caught and corrected without disturbing the actual data. The guest post's point is that a surface code implementation is, at its heart, centered on an algorithm that works the same way regardless of whose hardware you run it on, and that portability is exactly what should worry you if you're trying to argue it deserves patent protection. If the algorithm doesn't actually depend on the specific machine, the argument goes, then tacking on 'and you run it on a quantum computer' shouldn't be enough to convert an abstract mathematical scheme into a patentable invention, any more than running the same math on a classical laptop would be. This camp isn't arguing that quantum companies deserve zero patent protection, to be fair — that's a strawman worth knocking down before we move on. The argument is narrower: Section 101 scrutiny exists for a reason, and that reason doesn't evaporate just because the underlying tech is quantum instead of some other kind of software. Loosen the standard specifically to help quantum companies, and you loosen it for every marginal software patent behind it too. And there's a competitive angle buried in here that the reform camp doesn't love to discuss: broader eligibility tends to favor whoever already has the biggest, densest patent portfolio — which, as we're about to get into, is exactly the position IBM just spent a lot of money reinforcing. Which brings us to camp four — the people who'd tell you the legal theory is almost beside the point, because the practical strategy is just to buy your way past the uncertainty. Whatever doctrinal fog is hanging over Section 101, IBM's approach under CEO Arvind Krishna has been to acquire patent-dense research shops outright rather than litigate eligibility case by case. HRL brings more than eleven hundred patents and decades of expertise in silicon-spin qubits, quantum sensing, cryogenics, and control electronics into an IBM portfolio that was already the largest in the industry at 4,388 patent families. You don't need a settled legal doctrine if you simply own enough of the prior art and enough of the adjacent claims that nobody can build around you cheaply. Call it a moat, or call it — as some smaller quantum startups almost certainly will — a land grab that raises the cost of entering the field at exactly the moment it's getting commercially interesting. The fifth camp zooms out even further, to the geopolitical picture. Analysts at groups like ITIF and the US-China Economic and Security Review Commission point out that America's more restrictive eligibility doctrine, whatever its legal merits, might be ceding raw volume and standard-setting influence to China, whose patent office takes a more permissive approach to software and algorithm claims, backed by heavy state investment — China has claimed more than fifteen billion dollars in public quantum funding, far outpacing US public spending. Data compiled by China's own CAICT institute, through August of last year, put the US at just under fifty percent of global quantum computing patents and China at just over twenty-four percent, though other trackers dispute those exact shares, and definitions get slippery about what counts as 'quantum computing' patents versus 'quantum communication' patents, which is a China stronghold for different reasons entirely. So does a permissive Chinese patent office translate into an actual technological lead, or just a bigger filing cabinet? That's genuinely contested, and it's one of the open questions we don't have a clean answer to yet, but it's exactly the kind of asymmetry that makes Washington nervous every time this debate comes up. So where do I land on all this? Here's my read: IBM's HRL deal isn't really a bet that the patent doctrine will resolve in its favor — it's a bet that it doesn't have to. If you're the company holding the biggest pile of patents no matter which way Section 101 eventually breaks, uncertainty actually works for you. It's the small quantum startup with three patents and one promising algorithm who needs Congress to fix Alice, because they can't out-spend anyone in litigation and they can't out-file anyone at the Patent Office either. That asymmetry is the part of this story that doesn't get said out loud enough: patent uncertainty isn't a neutral tax on the whole industry, it falls hardest on exactly the companies that need protection most and can least afford ambiguity. We've talked before on this show about the strategic tension sitting under everything IBM does in quantum right now — the read here has been that IBM needs quantum to be the answer, and whether the market's actually buying that long-term is still an open question. This patent story doesn't resolve that tension, but it sharpens it. Locking up HRL's eleven hundred patents is IBM hedging its bets across two qubit technologies at once, silicon-spin and superconducting, while also hedging against a legal landscape it can't fully control. That's not necessarily cynical — genuine R&D consolidation and defensive land-grab aren't mutually exclusive, and I'd guess this deal is honestly both at once. Where I come down on the doctrine itself: Ex parte Yudong Cao is one data point, not a trend, and anyone telling you it settles quantum patent eligibility is overselling a single PTAB reversal. The skeptics citing American Axle and Symantec have the better read on the actual legal risk of loosening things too far — we've watched software patent fights clog courts before, and quantum shouldn't get a doctrinal carve-out just because the marketing is cooler than the average business-method patent. But the reform camp isn't wrong that three dead bills in three years is a genuinely bad look for a Congress that claims to care about American competitiveness in exactly this kind of technology. The open questions here are the ones that matter for the next year, not the next decade: does the Patent Eligibility Restoration Act ever get a floor vote, or does quantum patent law keep getting built one PTAB ruling at a time? Is IBM's HRL haul a genuine consolidation of research talent, or the first move in pricing smaller quantum companies out of a field they helped build? And if China really is racking up patent volume under looser rules, does that translate into technological leadership, or just a bigger filing cabinet with less inside it? None of those questions have honest answers yet, which is exactly why this is a story worth watching instead of one worth declaring finished. These are threads we'll keep pulling on, especially once we see what Congress does, if anything, with patent reform this session. If this kind of deeper dive is useful to you, follow the show wherever you're listening, and if you've got a few minutes, a rating genuinely helps more curious people find us. This has been Quickly Quantum, an AI-voiced podcast, created and built by a real human using today's cutting-edge technology. Nothing you heard on this show is financial advice. I'm Brian Lampert, and I'll catch you all tomorrow — take care!