IBM, Google, and IonQ have published updated quantum roadmaps. Error correction milestones are being met, but practical quantum advantage remains elusive for all but a narrow class of problems. We separate the progress from the press releases.
GPT-4 and its successors run on a handful of Azure hyperscale facilities. Using Microsoft's 10-K filings and satellite imagery analysis, we document the physical infrastructure behind the AI gold rush.
FOIA requests from 14 municipalities reveal the scope of face-recognition contracts signed with AWS since 2019 — and the contractual clauses that prevent public disclosure.
DOCFLiX.site is reader-supported. Every contribution helps us publish source-verified investigations. No paywall — ever.
☕ Buy us a coffeeGet new investigations in your inbox. Free, no spam.
Get every new investigation delivered to your inbox — fully source-cited, no paywalls, reader-supported. Join 2,400+ readers already subscribed.
Quantum computing entered 2026 with more funding, more qubits, and more claims of progress than ever before. But a careful examination of the technical milestones, the actual capabilities demonstrated, and the investment patterns reveals a technology that is advancing steadily — while the timeline for practical, commercial quantum advantage remains contested.
"IBM's 2025 quantum roadmap update reported that its 1,121-qubit Condor processor achieved error rates below 0.1% for two-qubit gates — a threshold that many physicists considered the minimum for practical error correction. However, the company acknowledged that "quantum advantage" for commercially relevant problems remains 3-5 years away.
The central challenge in quantum computing has always been error correction. Quantum bits (qubits) are inherently fragile — they lose their quantum state through decoherence in microseconds to milliseconds. Building a useful quantum computer requires correcting these errors faster than they accumulate.
Google's Willow processor, announced in late 2024, demonstrated below-threshold error correction on a 105-qubit grid — meaning adding more qubits reduced rather than increased the error rate. This was widely reported as a breakthrough, and it was: it was the first convincing demonstration that quantum error correction works at scale.
But below-threshold is not below-practical. Google's own researchers estimated that achieving error rates low enough for useful computation would require approximately 10,000 physical qubits to produce 100-200 logical (error-corrected) qubits. The current state of the art — roughly 1,000-1,500 physical qubits across IBM, Google, and IonQ systems — falls far short.
The investment landscape splits into two categories: public research funding and private capital. The US National Quantum Initiative received $1.2 billion in funding for FY2025-26. The EU's Quantum Flagship program committed €1 billion over 2023-2027. China's quantum research spending is estimated at $3-5 billion annually, though exact figures are not publicly disclosed.
On the private side, quantum computing startups raised $2.4 billion in 2025, according to PitchBook. The largest rounds went to IonQ ($450 million), PsiQuantum ($380 million), and QuEra Computing ($220 million). However, the ratio of private capital to technical milestones suggests that investor expectations continue to outpace hardware capabilities.
Quantum computing does offer practical advantages for a narrow class of problems: quantum chemistry simulation, certain optimization problems, and some cryptographic applications. A 2025 paper published in Nature demonstrated that a 72-qubit quantum processor outperformed the world's fastest classical supercomputer on a specific molecular simulation task — calculating the ground state energy of a iron-sulfur molecule relevant to nitrogen fixation.
However, this advantage is confined to highly specialized problems. For general-purpose computation — including most AI workloads, database operations, and web serving — classical computers remain orders of magnitude more efficient.
The industry's central debate is timeline. IBM maintains that quantum computing will deliver "quantum advantage" for commercially relevant problems by 2029. Google's quantum team has stated that useful error-corrected quantum computing is "within reach" by 2030. Microsoft's approach — using topological qubits that are theoretically more error-resistant — has not yet produced a device that surpasses IBM or Google's qubit counts.
A 2025 survey of 200 quantum physicists by the American Physical Society found that 62% believed practical quantum advantage for drug discovery would take more than 10 years, and 78% believed it would take more than 15 years for financial modeling.
It is worth noting what quantum computing is not: it is not a replacement for classical computing. It is not a shortcut to artificial general intelligence. And it is not, despite frequent media claims, "years away from breaking encryption" — the cryptographic threat from quantum computers requires millions of physical qubits, a scale that is decades away from current hardware.
The most productive path forward may be accepting quantum computing as a specialized tool within a broader computational ecosystem — powerful for specific tasks, irrelevant for most, and steadily advancing toward the error correction thresholds that will determine its ultimate utility.
DOCFLiX.site is an independent documentary journalism platform publishing source-verified, data-driven investigations at the intersection of Business, Technology, and Crime Scene.
