QDC#11: Hot Quantum Summer

GM frens, this is the Quantum Doom Clock with Colton Dillion and Richard Carback, the founders of Quip Network, building the world’s shared quantum computer.

When we last wrote, we highlighted the algorithm that could break ECDSA with fewer than 500,000 physical qubits by the research team out of Google, Ethereum, and Stanford with the details hidden by a zero knowledge proof (ZKP). In particular, we had criticized this as "more marketing than substance" with skepticism about both the efficacy of a ZKP, and the fidelity of the approach more generally. We were right:

• Researcher André Schrottenloher trivially bypassed the ZKP by combining existing research.
• There are now even more improved circuits, courtesy of EigenLab AI research improvement approach over at ECDSA Fail, featuring a 40% improvement as of this writing.

Yet again we are a couple weeks late on this update. We had to sort through over 4000 articles to find over 500 publications with actual new updates across them, so you will notice that this is a very beefy update post. The amount of new high signal information is becoming difficult for our team to parse, therefore expect changes in the form of slightly more frequent updates to keep things tractable. The theme of this one is just the overwhelming amount of advances across many of the different areas we track, and we think this will be the trend for the summer.

The Narrowing Case for Skepticism

Scott Aaronson, a quantum pioneer who is also a popular skeptic of quantum claims, is now beating a different drum. He now explains that recent experimental results in two-qubit gate fidelity are closing the gap toward practical error correction. This progress undermines previous skeptical arguments that correlated noise would prevent large-scale computation. The hardware advantages are structural, occurring when algorithms leverage specific mathematical properties like entanglement, not algorithmic! Scott is now sounding the alarm:

"if quantum computers start breaking cryptography a few years from now, don’t you dare come to this blog and tell me that I failed to warn you. This post is your warning."

Dr. David Gunnarsson argues that the industry is now waiting for a single, high-impact commercial application to trigger mass adoption. As we have been fond of saying recently, the focus is shifting from technical metrics toward demonstrating commercial value. We expect to see more high-profile examples in the coming year as the latest generation hardware starts to roll out.

Impressive Next Generation Hardware Announcements

Microsoft researchers released a preprint describing Majorana 2, and now claims their quantum chip can maintain qubit information for over 20 seconds. Everyone is still very skeptical of these claims, and so are we. If true, these advances are absolutely transformative.

Chinese groups also had two major announcements:

1. Origin Quantum launched the Wukong-180, their fourth-generation superconducting quantum computer with 180 computational qubits.
2. The University of Science and Technology announced the Jiuzhang 4.0 photonic quantum computer prototype, which can solve Gaussian boson sampling problem in just 25 microseconds.

We’re excited to be getting cloud access to the Wukong, which is in line with current offerings by IBM et al. We expect to see confirmations with random circuit sampling and other experiments.

Real World Use Cases

A number of real world use case announcements were released. We are breaking this one down by sector in listicle format for you:

Physics:

1. Alice & Bob is collaborating with Los Alamos National Laboratory and GE Vernova to simulate the magnetic properties of new materials. Powerful permanent magnets that do not require rare earth elements eliminates reliance on a brittle part of the global supply chain.
2. Researchers at the Niels Bohr Institute engineered quantum dots that emit single photons compatible with existing fiber optic networks.

Biomedical:

1. IBM, RIKEN, and Cleveland Clinic successfully simulated a protein consisting of 12,635 atoms using a hybrid quantum-classical approach. This large-scale simulation represents a significant milestone for drug discovery and life sciences research.
2. Qubit Pharmaceuticals and the Centre for Quantum Technologies have partnered to develop and test quantum algorithms for molecular simulation. Researchers will use Quantinuum hardware to determine if these methods provide a practical advantage in drug development over classical techniques.

Finance:

IonQ developed a hybrid quantum-classical process to solve complex portfolio optimization problems. This method utilizes a 100-physical-qubit system to demonstrate advantages in risk minimization and return maximization.

Tech:

1. Cisco developed a universal quantum switch to route information between different systems.
2. Multiverse Computing developed a method to reduce AI uncertainty by integrating quantum circuit blocks into large language models.
3. HSBC, Haiqu, and academic researchers developed a method to encode complex probability distributions into shallow quantum circuits.

This is a lot of concrete developments over a short period of time.

Benchmarking

A number of benchmarking models are coming out, and many are being used to optimize the hardware. The Johns Hopkins team created a noise-modeling framework with sevenfold greater predictive accuracy for superconducting processors. Leonardo Placidi and colleagues analyzed how qubit connectivity impacts computational complexity, finding that sparse architectures can increase circuit depth by 30%. Such connectivity-aware frameworks are essential for determining when quantum hardware can truly surpass classical limits.

We also saw several optimizations. For example, Qruise automated the calibration of a 21-qubit QPU to drastically reduce bring-up time. AQT achieved a Quantum Volume of 32768 on its LYNX trapped-ion system. Pasqal demonstrated that an error-detecting code can reduce average errors by over 50% during machine learning workflows.

Frameworks and datasets for benchmarks are also becoming prominent. Alice & Bob introduced a five-criteria framework to standardize logical qubit measurements across different hardware. Sudip Vhaduri and colleagues evaluated the efficiency of quantum machine learning models on the MNIST dataset. Lastly, researchers developed a standardized metric to quantify energy efficiency by comparing algorithmic performance to hardware power consumption.

Algorithmic Advancements

We saw a number of quantum simulation and algorithmic advances. Researchers at the Jülich Supercomputing Centre and NVIDIA fully simulated a 50-qubit universal quantum computer using the JUPITER exascale supercomputer. This achievement surpasses previous simulation records and provides a high-fidelity benchmark for testing algorithms before they reach physical hardware. We had believed simulation at this scale was not possible, and going much further is likely intractable unless you introduce approximation to the approach.

Speaking of approximation approaches, physicists at the Center for Computational Quantum Physics and Boston University used tensor networks to simulate complex quantum systems on a personal laptop. Aalto University researcher also developed a quantum-inspired algorithm to model complex quasicrystals using tensor networks. Such work is always characterized as challenging assertions of quantum supremacy, but our attitude is that we needed to understand quantum to get to this point, so there is an intrinsic value to the entire space when you see such developments.

It is also simply not correct to look at such a result and assume there is no quantum supremacy. For example, Pasqal demonstrated that logical qubits outperform physical qubits when solving differential equations.

Lastly, we are also seeing a lot of resource optimization work. Quanscient and Haiqu developed a One-Step Simplified Lattice Boltzmann Method algorithm to reduce qubit requirements for fluid simulations. QuEra Computing researchers proposed a new architecture to reduce resource overhead in fault-tolerant quantum computing, and Nvidia released an Ising AI model suite to assist with quantum processor calibration and error correction.

Business News

A few major business announcements stand out. IBM and the U.S. Department of Commerce will launch Anderon, a dedicated 300-millimeter quantum chip foundry based in Albany, New York. This initiative is supported by a $1 billion CHIPS Act incentive and a $1 billion investment from IBM to accelerate the engineering of fault-tolerant systems. The foundry aims to provide advanced wafer fabrication for multiple hardware vendors, securing American leadership in quantum manufacturing.

D-Wave Quantum has unveiled a strategic roadmap to develop commercial, fault-tolerant gate-model systems. The company targets a 100-logical-qubit (NOT physical) architecture by 2032 using a specialized superconducting dual-rail qubit design. This plan focuses on reducing the physical qubit overhead required for error correction through hardware-level detection.

Other notable developments:

• Quantinuum is preparing for a major Nasdaq IPO that could value the firm as high as $20 billion.
• IonQ achieved a technological milestone by completing its first sale of a 256-qubit system.
• Rigetti Computing announced plans to invest $100 million in the United Kingdom to deploy a 1,000-qubit system.
• Microsoft invested over €134 million in a specialized research laboratory in Denmark to drive commercial use.

---

The Quantum Doom Clock is brought to you by Richard Carback and Colton Dillion, the cofounders of Quip Network.

The Quantum Doom Clock is a monthly mailing list that summarizes news for Quantum Computing and its effects on the cryptography and cryptocurrency spaces. We do not sell your e-mail.