Aramco and SandboxAQ look at quantum AI for industry

Aramco named SandboxAQ as its collaborator on next-generation quantum AI within the set of technology agreements and memoranda it announced at LEAP 2026 in Riyadh, the one entry in that partner list that belongs to research rather than to production. The label needs unpacking, because quantum AI is used loosely and covers at least three different things. It can mean quantum computing applied to machine learning problems, which is largely theoretical and constrained by how few reliable qubits exist. It can mean classical algorithms borrowed from quantum physics and run on ordinary hardware — quantum-inspired methods, which are real today and sometimes very effective on optimisation problems. And it can mean the cryptographic consequence: the fact that a sufficiently capable quantum computer would break the public-key cryptography that currently protects almost everything, and that organisations therefore need to inventory and replace it well before such a machine exists. SandboxAQ, which was spun out of Alphabet in 2022, has built its business across that territory, combining AI with quantum-derived techniques and doing substantial work on the cryptographic migration problem. Neither Aramco nor SandboxAQ has published which strand this collaboration covers, and the difference is not academic. Quantum-inspired optimisation applied to, say, refinery scheduling or logistics is something an operator could put to work within a planning cycle. Post-quantum cryptography migration across an industrial control estate is a multi-year inventory-and-replace programme. Machine learning running on quantum hardware is a research relationship. The same two-word label covers all three. For an energy company the plausible near-term applications are the unglamorous ones. Optimisation is everywhere in the business: which crude to run, how to schedule maintenance across a fleet, how to route cargoes, how to configure a plant when feedstock changes. These are combinatorial problems where small percentage improvements are worth a great deal, and they are exactly the class of problem quantum-inspired methods target. Molecular simulation for catalysts and materials is the longer-horizon prize, and the one that would genuinely need quantum hardware. Saudi Arabia has been putting money into the underlying science for some time. KACST's National Center for Quantum Technologies signed a research collaboration with Pasqal on quantum-safe cryptography, pairing neutral-atom quantum processors with national research infrastructure. A corporate collaboration is a different kind of signal from a national research one: it suggests somebody has an application in mind, even if nobody has said what it is. The collaboration sits alongside the rest of Aramco's LEAP announcements, which include hardware collaborations with NVIDIA, Groq and Qualcomm , process control work with Yokogawa , and on the enterprise side a collaboration agreement with ServiceNow . The set spans cybersecurity, industrial AI, innovation, the localisation of critical technologies and national industrial capability. What has not been published is the scope, the value, the duration, the form of the agreement, or which of Aramco's businesses is involved. The evidence for this collaboration is a single line in a partner list, and it should be read as a statement of interest rather than a programme with milestones. Quantum announcements age badly for a reason: the gap between what is demonstrated in a paper and what runs in a refinery is wider here than anywhere else in computing. The thing to watch is whether the first published output from this collaboration is a cryptographic inventory, an optimisation result with a number attached, or nothing at all. TechScoop is reporting from LEAP 2026 in Riyadh.