Quantum AI Report

The convergence of Quantum with AI

Superconducting

Qubits built from superconducting circuits cooled to near absolute zero. The most industrially mature approach, favoured by IBM, Google, and Rigetti; fast gates, but short coherence times and demanding cryogenics.

128 stories

Source

DARPA Selects Atom Computing, Diraq, IBM, and IonQ for Final Stage C Verification under Quantum Benchmarking Initiative

DARPA has selected Atom Computing, Diraq, IBM, and IonQ to participate in Stage C of its Quantum Benchmarking Initiative. The four companies, whose platforms span neutral atoms, silicon spin qubits, superconducting transmons, and trapped ions, will undergo independent hands-on testing of fault-tolerant architectures. They join Microsoft and PsiQuantum, which were already in the final verification stage, as part of DARPA's goal of a utility-scale quantum machine by 2033.

OutlookPlausible

DARPA's Stage C verification could produce the first independent, cross-platform comparison of logical error rates and resource overhead across superconducting, trapped-ion, neutral-atom, and silicon-spin systems within two years.

othererror correctionneutral atomphotonicspin qubitsuperconductingtopologicaltrapped ionAtom ComputingDARPADiraqIBMIonQMicrosoftPsiQuantum
Source

Modular Neural Decoder for Surface Code Memory and Logic on Superconducting Processors

A preprint describes a modular graph-recurrent neural-network decoder for surface-code quantum error correction. The decoder is first trained on simulated noise, then fine-tuned using data from actual superconducting qubits. The work focuses on making machine-learning decoders portable across different devices and noise models.

OutlookPlausible

If fine-tuning transfers as intended, this could give error-correction teams a reusable decoder that adapts to new superconducting devices with brief hardware calibration rather than retraining from scratch.

error correctionalgorithms softwaresuperconducting
Source

Fermilab SQMS Center Identifies Microscopic Origins of Qubit Performance Variance Across Superconducting Transmons

Researchers at Fermilab's SQMS Center have traced significant variation in superconducting transmon coherence times to nanometre-scale differences in fabrication geometry and the presence of surface oxides. These microscopic features were associated with up to twofold swings in qubit performance. The resulting materials-level data is intended to feed device models and automated qubit calibration pipelines.

OutlookPlausible

ML-driven calibration pipelines could ingest these fabrication geometry and oxide signatures to flag low-coherence qubits early and adjust tuning strategies before standard characterization, reducing calibration overhead on multi-qubit devices.

superconductingalgorithms softwareFermilab SQMS Center
Source

Researchers Build Fast Qubit Reader Using Spiking Nets

Researchers built a qubit-state classifier that uses spiking neural networks to process superconducting qubit measurement signals as they stream in. The network identifies the qubit state before the complete readout trace is collected, lowering classification latency. The authors state the approach is fast enough for real-time quantum error correction and control.

OutlookPlausible

Within two years, stream-based spiking classifiers could be embedded in FPGA or neuromorphic coprocessors alongside superconducting qubit control stacks, enabling faster active reset and feedback without waiting for full readout digitization.

superconductingalgorithms softwareerror correction
Source

IBM gets PILOT agreement from IDA for $2.5 billion Quantum project

The Dutchess County Industrial Development Agency approved tax incentives for IBM's planned $2.5 billion expansion of its Poughkeepsie site. The facility would build and test fault-tolerant quantum computing systems, positioning it among the largest private quantum manufacturing investments in the US.

OutlookPlausible

IBM could bring online a dedicated fault-tolerant quantum hardware integration and test line at Poughkeepsie within two years, giving its superconducting quantum programme in-house capacity to iterate on error-corrected systems faster than through external research fabs.

superconductingerror correctionDutchess County Industrial Development AgencyIBM
Source

Quantum Machine Learning for Cybersecurity Applications: Simulation and Hardware Validation

A hybrid classical-quantum machine learning architecture compresses input features with a classical multilayer perceptron and then passes them to few-qubit quantum support vector machine and variational quantum circuit heads. The approach was tested on intrusion detection and spam classification datasets, including validation runs on IBM Quantum hardware.

OutlookPlausible

Within two years, security teams could pilot lightweight hybrid QSVM/VQC classifiers for anomaly detection on network flows, using classical feature compression to keep qubit counts within reach of existing noisy quantum devices.

superconductingalgorithms softwareIBM
Source

GlobalFoundries Finalizes $375 Million CHIPS Act Award for Onshore Quantum Semiconductor Fabrication

GlobalFoundries has finalized a $375 million award under the US CHIPS Act to expand domestic fabrication capacity for quantum computing chips. The funding targets the physical hardware layer that supports hybrid quantum-classical and AI-accelerated computing stacks, such as those being assembled by NVIDIA and IonQ.

OutlookPlausible

This could enable US quantum hardware developers to iterate on cryogenic control and readout chips with shorter domestic fab cycles, accelerating integration of quantum processors into hybrid AI-quantum stacks.

othercryogenics controlspin qubitsuperconductingGlobalFoundries
Source

IBM, Lockheed Martin Announce Swiss Quantum Innovation Hub at ETH Zurich, Anchored by Switzerland's First IBM Quantum Computer

IBM will install a Quantum System Two, its latest superconducting system with the Nighthawk processor, at the Swiss National Supercomputing Centre by the end of 2026. The installation is part of a defence-offset agreement with armasuisse and anchors a new innovation hub at ETH Zurich run with Lockheed Martin. The partners plan joint research in quantum sensing, additive manufacturing, and AI-adjacent algorithms.

OutlookPlausible

The Zurich hub could become an operational testbed where Lockheed Martin and Swiss researchers run defence-relevant quantum sensing and additive-manufacturing algorithms on IBM's Nighthawk processor without moving data or hardware across US export boundaries.

superconductingalgorithms softwarequantum sensingETH ZurichIBMLockheed Martinarmasuisse