Amazon Unveils Ocelot Quantum Chip With Significantly Better Error Correction

Amazon has introduced Ocelot, a prototype quantum computing chip developed under AWS that aims to improve quantum error correction.

Unlike traditional error-handling approaches that require extra qubits to stabilize computations, Ocelot uses bosonic qubits, which are engineered to naturally suppress errors. AWS claims this design could reduce the need for additional qubits by up to 90%, potentially making quantum computing more efficient and scalable. Amazon presents their findings in a research paper in the renowned scientific journal nature.

The introduction of Ocelot represents a shift for Amazon, which has previously focused on cloud-based quantum computing through AWS Braket. With this move, Amazon steps into direct competition with Google and Microsoft, two companies that have already made advances in quantum hardware.

Why Error Correction is Quantum Computing’s Biggest Challenge

Unlike conventional computers that use stable binary bits, quantum computers process information using qubits, which can exist in multiple states at once due to a principle known as quantum superposition.

This capability allows quantum machines to solve problems that would take classical computers an impractical amount of time. However, qubits are highly sensitive to environmental disturbances, making error correction a necessity.

Most existing quantum processors rely on logical qubits, which combine multiple physical qubits into a more stable computational unit. While this approach improves accuracy, it requires a significant increase in qubit count, making large-scale quantum computing highly complex.

Ocelot’s use of bosonic qubits, also called cat qubits, is designed to tackle this issue differently. These qubits store quantum information in electromagnetic oscillations rather than in individual particles, making them less prone to common errors like bit-flips and phase shifts.

Oskar Painter, AWS Quantum Hardware Director, explained the design approach, stating: “We didn’t take an existing architecture and then try to incorporate error correction afterwards. We selected our qubit and architecture with quantum error correction as the top requirement.”

If this strategy proves successful, Ocelot could enable quantum processors that require fewer qubits to maintain stability, reducing hardware costs and power consumption.

Competing with Google and Microsoft’s Quantum Technologies

Amazon’s quantum hardware ambitions put it up against Google’s Willow chip and Microsoft’s Majorana 1, both of which offer alternative methods for handling error correction.

Google’s Willow, unveiled in December 2024, demonstrated an exponential reduction in error rates by using logical qubits. By grouping multiple physical qubits into a single, stabilized structure, Google’s approach allowed its quantum system to complete a complex benchmark task in under five minutes—something that would take the Frontier supercomputer longer than the age of the universe to process.

Microsoft, meanwhile, has opted for a more experimental approach with topological qubits. The company claims that its Majorana 1 chip, introduced in February 2025, could eliminate the need for traditional error correction altogether. Unlike superconducting qubits, which require additional correction layers, topological qubits are theoretically resistant to decoherence from environmental noise.

However, skepticism remains. John Preskill, a theoretical physicist at the California Institute of Technology who is also involved with Amazon’s AWS Center for Quantum Computing commented about Majorana 1, that “There is no publicly available evidence that this test has been conducted successfully.”

While a study published in Nature supported some aspects of the theoretical basis for Majorana qubits, Preskill and other scientists remain uncertain whether they can be implemented at scale.

By focusing on a more immediate refinement of error correction rather than an entirely new qubit design, Amazon’s Ocelot could serve as a bridge between Google’s proven logical qubit systems and Microsoft’s unverified topological qubits.

Amazon’s Shift from Quantum Cloud Services to Hardware Development

Until now, Amazon’s involvement in quantum computing has primarily revolved around AWS Braket, a cloud service that provides researchers and businesses with access to third-party quantum processors from companies such as IonQ and Rigetti. This model allowed AWS to facilitate quantum computing research without directly developing its own hardware.

However, the introduction of Ocelot represents a deeper investment in quantum technology, signaling that Amazon no longer wants to be just a cloud intermediary but an active competitor in building next-generation quantum systems. While companies like Google and IBM have been developing quantum processors in-house for years, Amazon’s decision to shift into proprietary hardware suggests a long-term strategy to gain a foothold in quantum computing infrastructure.

Microsoft has pursued a similar transition through Microsoft Quantum, gradually shifting from offering quantum computing simulations to developing its own hardware. However, unlike Microsoft’s ambitious but unproven attempt with topological qubits, Amazon’s Ocelot focuses on refining existing quantum error correction methods. This places AWS in a more immediate position to compete with established quantum hardware manufacturers.

What Comes Next for Amazon’s Quantum Strategy?

Despite Ocelot’s introduction, AWS has not provided specific performance benchmarks or a roadmap for when—or if—the chip will become part of its commercial offerings. The fact that it is still in the prototype stage suggests that Amazon is conducting further research before committing to large-scale deployment.

Meanwhile, quantum research efforts across the industry continue to evolve. Google’s logical qubits have already demonstrated error-correction breakthroughs, and Microsoft remains focused on advancing DARPA-supported research into scalable quantum computing. IBM, while taking a different approach, has pushed forward with increasing raw qubit count, as seen with its Condor processor, which currently holds the record for the largest superconducting quantum chip.

The real question now is whether Ocelot’s bosonic qubit approach will prove more effective than existing quantum architectures. If AWS successfully integrates this technology into practical computing models, it could reduce the high cost of error correction and pave the way for broader adoption of quantum systems in areas such as pharmaceutical research, financial modeling, and materials science.

Real-World Implications of Quantum Error Correction Advances

The competition among Amazon, Google, Microsoft, and IBM is not just about theoretical performance—quantum error correction has real-world implications across multiple industries. If companies can stabilize quantum computations at scale, industries that require immense processing power will see transformative advancements.

One of the most immediate applications is drug discovery, where pharmaceutical companies are exploring quantum simulations to model complex molecular interactions at speeds far beyond classical computing capabilities. In cybersecurity, the rise of stable quantum computers could eventually pose a threat to existing encryption standards, leading to the need for post-quantum cryptographic solutions. Meanwhile, materials science could benefit from quantum simulations that help discover new superconductors, advanced battery chemistries, and novel industrial materials.

Amazon’s Ocelot might not be the final answer to quantum error correction, but it introduces a new variable into a field where competition is accelerating. With companies exploring different architectures—from Google’s logical qubits to Microsoft’s topological qubits—the race to scalable quantum computing is no longer theoretical. Whether Amazon’s hardware initiative will prove successful remains to be seen, but one thing is clear: AWS is no longer just providing access to quantum computing—it is now actively shaping its future.