The Quantum Speedup Era:
Breakthroughs That Are Reshaping Computing

💡 Part 1: Jiuzhang 4.0 – The 10⁵⁴ Speedup

On May 13, 2026, a team led by Pan Jianwei and Lu Chaoyang at the University of Science and Technology of China published results in Nature that redefined what "quantum speedup" means. Their photonic quantum computer, Jiuzhang 4.0, achieved a speedup of 10⁵⁴ over the world's fastest classical supercomputer, El Capitan [Nature] [Xinhua].

To put that number in perspective: the fastest supercomputer on Earth would require 10⁴² years to solve the Gaussian Boson Sampling problem that Jiuzhang 4.0 completes in 25 microseconds [Ncsti]. That's a speedup so vast that "billions of billions of billions" doesn't even begin to cover it.

How did they achieve this? The team introduced a "programmable space-time hybrid encoding" architecture that allows photons to interfere simultaneously across both spatial and temporal dimensions [Ncsti]. This solved the "photon loss" bottleneck that had previously limited scaling [Xinhua]. The system handles 1024 squeezed-state inputs, 8176 modes, and manipulates 3050 photons—a 10× increase over Jiuzhang 3.0's 255 photons [Ncsti]. Since the computational state space grows exponentially with photon count, a 10× increase in photons yields a speedup far beyond 10× [Ncsti].

Jiuzhang 4.0 – Key Metrics
            ├── Photons: 3050 (vs. 255 in Jiuzhang 3.0)
            ├── Modes: 8176
            ├── Squeezed States: 1024
            ├── Speedup: 10^54 over El Capitan
            ├── Runtime: 25 microseconds
            └── Classical Equivalent: 10^42 years

China is now the only country to have achieved quantum supremacy in both photonic and superconducting qubit architectures [Ncsti]. The 2025 "Zuchongzhi 3.0" superconducting system had already set records, and Jiuzhang 4.0 now extends that dominance into the photonic domain [Baidu].

🔬 Part 2: Helios – 98 Qubits with Unprecedented Fidelity

On June 17, 2026, Nature published peer-reviewed results on Helios, a 98‑qubit trapped‑ion quantum computer developed by Quantinuum in collaboration with Sandia National Laboratories [Nature] [The Quantum Insider]. The system achieved single‑qubit fidelity of 99.9975% and two‑qubit fidelity of 99.921%, making it the largest and highest‑performing trapped‑ion quantum computer to date [The Quantum Insider].

What sets Helios apart is its reconfigurable architecture: any two qubits can be directly connected, enabling complex algorithms without the routing overhead that plagues fixed‑topology superconducting systems [Nature News]. This "all‑to‑all" connectivity is a game‑changer for fault‑tolerant quantum computing, as it dramatically reduces the number of SWAP gates required to execute algorithms [Nature News].

Sandia's benchmarking methods confirmed that Helios maintains accuracy while scaling to larger system sizes—something that has historically been difficult for trapped‑ion platforms [Nature News]. The results establish a clear path toward fault‑tolerant systems large enough to solve real‑world chemistry, materials science, and optimization problems [The Quantum Insider].

🧬 Part 3: Majorana 2 – A 1000‑Fold Leap in Reliability

At Microsoft Build 2026, the company unveiled Majorana 2, its second‑generation topological quantum processor [CNA]. The chip, designed with the help of agentic AI, replaces aluminum with a lead‑based superconducting material stack that more than doubles the topological gap protecting quantum states [The Quantum Insider].

The results are staggering: qubit coherence times improved from milliseconds to over 20 seconds—a 1,000‑fold improvement over the first generation [The Quantum Insider] [CNA]. Some configurations reached 60 seconds of qubit lifetime [CNA]. To put that in perspective: most quantum systems measure qubit lifetime in microseconds; Majorana 2 measures it in minutes [CNA].

This breakthrough has allowed Microsoft to cut its roadmap in half: the company now expects to deliver a scalable, commercially viable quantum computer by the end of 2029, rather than the previously projected 2032 [CNA]. Such a system, Microsoft claims, will be capable of tackling "complex problems in global health, food supply, sustainable development, and energy production" [CNA].

Majorana 2 – Reliability Milestones
            ├── Coherence Time: 20+ seconds (vs. milliseconds in Gen 1)
            ├── Improvement: 1,000×
            ├── Material Stack: Lead-based (vs. aluminum)
            ├── Topological Gap: 2× larger
            └── Commercial Target: 2029 (halved from previous roadmap)

💻 Part 4: HamSim – 1,269× Speedup on Classical Hardware

Not all speedups come from quantum hardware itself. On June 15, 2026, researchers submitted HamSim, a classical simulation framework that achieves extraordinary performance by exploiting diagonal sparsity in Hamiltonian matrices [arXiv].

Benchmarks on the HamLib suite show HamSim outperforming Qiskit‑Aer by 182× to 1,269× on optimization instances like TSP and MaxCut, and 4.8× to 841× on physical models such as TFIM and Heisenberg [arXiv]. On GPUs, it achieves up to 178× speedup for 12–16 qubit problems [arXiv].

Unlike traditional Trotterization, HamSim maintains near‑perfect fidelity without requiring exponential steps [arXiv]. This demonstrates that structure‑aware classical algorithms can still play a vital role in the quantum era—and may even outperform quantum hardware on certain workloads [arXiv].

🌐 Part 5: Borealis – Public Cloud Quantum Advantage

Xanadu has deployed its Borealis photonic quantum processor on public cloud infrastructure via Amazon Braket, marking the first time a programmable photonic system demonstrating quantum advantage has been opened to public access [Quantum Computing Report].

Borealis uses 216 squeezed‑state qubits generated as pulses of light entangled across three temporal dimensions within a continuous loop of optical fiber [Quantum Computing Report]. It completes the Gaussian Boson Sampling sequence in 36 microseconds; direct simulation on the world's fastest supercomputer using the best known classical algorithms would require approximately 9,000 years for a single equivalent sample [Quantum Computing Report].

This represents a 50‑million‑fold runtime acceleration over earlier, non‑programmable photonic sampling demonstrations, validating the scalability of time‑multiplexed optical architectures [Quantum Computing Report].

Borealis – Photonic Quantum Advantage
            ├── Qubits: 216 squeezed‑state photonic
            ├── Runtime: 36 microseconds
            ├── Classical Equivalent: 9,000 years
            ├── Speedup vs. Previous Photonic Systems: 50 million×
            └── Availability: Public cloud (Amazon Braket)

🔬 Part 6: Beyond the Headlines – Other Notable Advances

The past two months have seen a cascade of additional breakthroughs:

• Exponential Quantum Speedup (Simon's Problem) – Researchers demonstrated exponential speedup on IBM's 156‑qubit Boston and 120‑qubit Miami processors using constant‑depth compiled circuits [arXiv]. The circuits achieved sufficiently high fidelity to exhibit algorithmic quantum speedup without error suppression [arXiv].
• Quantum‑Classical AFQMC – Researchers combined IonQ Forte (24 qubits) with NVIDIA GPUs to model chemical reaction barriers, achieving a 656× time‑to‑solution improvement over prior state‑of‑the‑art [Physical Review Research]. The workflow simulated the oxidative addition step of the nickel‑catalyzed Suzuki–Miyaura reaction—the largest QC‑AFQMC experiment ever performed on quantum hardware [Physical Review Research].
• 50‑Qubit Simulation on Exascale – Europe's first exascale supercomputer achieved a 16.6‑fold speedup over previous records in simulating 50‑qubit universal quantum circuits, demonstrating near‑linear scalability [ScienceDirect].
• Fault‑Tolerant Resource Reduction – QuEra and Los Alamos National Laboratory published a new architecture that cuts physical resource requirements for early fault‑tolerant quantum simulation by orders of magnitude [QuEra].
• Logical Qubits Outperform Physical Qubits – Pasqal demonstrated an industry first: logical qubits (error‑corrected) outperforming physical qubits in solving differential equations on a neutral‑atom quantum computer [Pasqal].

🔮 Part 7: What It All Means

The quantum speedup breakthroughs of 2026 are not incremental. They represent an inflection point:

• Photonic quantum computing – Jiuzhang 4.0 and Borealis demonstrate that light‑based systems are not just viable but capable of unimaginable speedups. The 10⁵⁴ factor is so large it defies intuition—and it's real.
• Trapped‑ion systems – Helios proves that trapped ions can scale while maintaining extraordinary fidelity, making them a serious contender for fault‑tolerant computing.
• Topological qubits – Majorana 2's 1,000× reliability leap suggests that Microsoft's bet on topological protection may pay off sooner than anyone expected.
• Classical simulation – HamSim shows that classical algorithms are far from obsolete; they are evolving to exploit the same structural insights that make quantum computing powerful.

The race is no longer about whether quantum advantage is possible—it's about which modality will dominate, and when these systems become commercially available for chemistry, materials science, and optimization at scale [The Quantum Insider].