The paper states that none of these infidelities are fundamentally limited and are likely able to be improved further .

For context, typical two-qubit gate fidelities for competing superconducting systems from Google and IBM are reported in the range of 99.5% to 99.7%, meaning Helios operates with roughly 10–100× lower error rates .

QCCD Architecture with Rotatable Ion Storage Ring

Helios is built on a quantum charge-coupled device (QCCD) architecture implemented on a 2D surface electrode trap . Unlike superconducting qubits that remain fixed in place, Helios physically transports ions between spatially separated memory and logic zones — analogous to how bits flow between cache, data buses, and logic units in a classical CPU .

Key architectural features:

• ¹³⁷Ba⁺ hyperfine qubits: Helios uses individual barium ions as information carriers, a choice that enables improved quantum operation error rates with a more scalable laser architecture .
• Rotatable ion storage ring: All-to-all connectivity is enabled by a ring-shaped structure that connects two quantum operation regions via a junction, allowing any qubit to interact with any other qubit .
• Tuning-fork-style trap junction: The junction routes ions to any gate zone, enabling arbitrary qubit pair interactions .
• Average depth-1 time of 55 ms per circuit layer across all 98 qubits .

This transport-based approach yields reconfigurable, flexible connectivity without the wiring overhead of fixed-coupling architectures.

Scalable Error-Corrected Logical Qubits

A major result of the Nature paper is the demonstration of multiple scales of logical qubit encoding on Helios:

• 94 logical qubits with error detection and 48 logical qubits with full error correction were extracted from the 98 physical qubits, achieving better-than-break-even performance — meaning the encoded logical qubits outperformed their unencoded physical counterparts .
• With 50 error-detected logical qubits, Helios ran the largest encoded simulation of quantum magnetism to date .
• The system achieved a 2:1 encoding rate for error-corrected logical qubits (two physical qubits per logical qubit), a ratio thought impossible only a few years ago .

The Nature paper used mid-circuit measurements (nondestructive readout), a capability essential for active error correction, and Sandia contributed a new benchmarking methodology specifically to measure the performance of these measurements .

In a separate technical report published the same day, Mitsubishi Electric evaluated the Quantum Fourier Transform on up to 12 logical qubits encoded in the Steane code on Helios .

Independent Validation by Sandia National Laboratories

Sandia National Laboratories independently evaluated and certified the Helios system, supplying new benchmarking tools and co-authoring the paper . Sandia's Quantum Performance Laboratory has a long-standing Cooperative Research and Development Agreement with Quantinuum, most recently renewed in May 2026 .

The public sources contain no mention of Europe's JUPITER exascale supercomputer validating Helios. This claim is uncorroborated by the Nature paper or related news coverage — any association between JUPITER and Helios validation remains unsupported by the available evidence.

Broader Implications for Quantum Computing

The Helios results carry several implications for the field:

• Trapped-ion technology reaches 100-qubit parity with superconducting circuits: Helios is the first 100-qubit-class trapped-ion quantum computer and demonstrates that trapped-ion systems can achieve superconducting-circuit-scale qubit counts while maintaining 10–100× lower error rates .
• Beyond classical simulation: Random circuit sampling benchmarks confirm that Helios operates well beyond the reach of classical simulation .
• Predictive error model: Component-level infidelities are predictive of system-level performance, meaning the error model is reliable and scalable .
• Supports DOE fault-tolerant quantum computing goals: The results mark a "new frontier of fidelity and complexity for quantum computers," aligning with the U.S. Department of Energy's objectives .

Quantinuum CEO Dr. Rajeeb Hazara characterized Helios as "a seamless fusion of hardware and software, creating a platform for discovery unlike any other" .

The paper states that none of these infidelities are fundamentally limited and are likely able to be improved further .

For context, typical two-qubit gate fidelities for competing superconducting systems from Google and IBM are reported in the range of 99.5% to 99.7%, meaning Helios operates with roughly 10–100× lower error rates .

QCCD Architecture with Rotatable Ion Storage Ring

Helios is built on a quantum charge-coupled device (QCCD) architecture implemented on a 2D surface electrode trap . Unlike superconducting qubits that remain fixed in place, Helios physically transports ions between spatially separated memory and logic zones — analogous to how bits flow between cache, data buses, and logic units in a classical CPU .

Key architectural features:

• ¹³⁷Ba⁺ hyperfine qubits: Helios uses individual barium ions as information carriers, a choice that enables improved quantum operation error rates with a more scalable laser architecture .
• Rotatable ion storage ring: All-to-all connectivity is enabled by a ring-shaped structure that connects two quantum operation regions via a junction, allowing any qubit to interact with any other qubit .
• Tuning-fork-style trap junction: The junction routes ions to any gate zone, enabling arbitrary qubit pair interactions .
• Average depth-1 time of 55 ms per circuit layer across all 98 qubits .

This transport-based approach yields reconfigurable, flexible connectivity without the wiring overhead of fixed-coupling architectures.

Scalable Error-Corrected Logical Qubits

A major result of the Nature paper is the demonstration of multiple scales of logical qubit encoding on Helios:

• 94 logical qubits with error detection and 48 logical qubits with full error correction were extracted from the 98 physical qubits, achieving better-than-break-even performance — meaning the encoded logical qubits outperformed their unencoded physical counterparts .
• With 50 error-detected logical qubits, Helios ran the largest encoded simulation of quantum magnetism to date .
• The system achieved a 2:1 encoding rate for error-corrected logical qubits (two physical qubits per logical qubit), a ratio thought impossible only a few years ago .

The Nature paper used mid-circuit measurements (nondestructive readout), a capability essential for active error correction, and Sandia contributed a new benchmarking methodology specifically to measure the performance of these measurements .

In a separate technical report published the same day, Mitsubishi Electric evaluated the Quantum Fourier Transform on up to 12 logical qubits encoded in the Steane code on Helios .

Independent Validation by Sandia National Laboratories

Sandia National Laboratories independently evaluated and certified the Helios system, supplying new benchmarking tools and co-authoring the paper . Sandia's Quantum Performance Laboratory has a long-standing Cooperative Research and Development Agreement with Quantinuum, most recently renewed in May 2026 .

The public sources contain no mention of Europe's JUPITER exascale supercomputer validating Helios. This claim is uncorroborated by the Nature paper or related news coverage — any association between JUPITER and Helios validation remains unsupported by the available evidence.

Broader Implications for Quantum Computing

The Helios results carry several implications for the field:

• Trapped-ion technology reaches 100-qubit parity with superconducting circuits: Helios is the first 100-qubit-class trapped-ion quantum computer and demonstrates that trapped-ion systems can achieve superconducting-circuit-scale qubit counts while maintaining 10–100× lower error rates .
• Beyond classical simulation: Random circuit sampling benchmarks confirm that Helios operates well beyond the reach of classical simulation .
• Predictive error model: Component-level infidelities are predictive of system-level performance, meaning the error model is reliable and scalable .
• Supports DOE fault-tolerant quantum computing goals: The results mark a "new frontier of fidelity and complexity for quantum computers," aligning with the U.S. Department of Energy's objectives .

Quantinuum CEO Dr. Rajeeb Hazara characterized Helios as "a seamless fusion of hardware and software, creating a platform for discovery unlike any other" .