The Quantum Computing Conference kicks off today: IONQ rises nearly 12% in pre-market, MPWR rises over 8%, Nasdaq hits a new high; AI power demand gains attention again

Today (September 23 to 25), the global quantum computing event Quantum World Congress 2026 will officially kick off in College Park, Maryland, with industry giants such as IonQ, D-Wave, Rigetti, Microsoft, and IBM in attendance. However, before the curtain rises, the secondary market has already ignited a frenzy: after hours yesterday, the quantum computing sector collectively surged, with IonQ jumping over 12%; this morning, the enthusiasm not only persisted, but IonQ continued to rise nearly 12%, while the simulation chip and power management giant Monolithic Power Systems (MPWR) also surged over 8%.

On one side is quantum computing, which is still in a technological breakthrough phase, and on the other side is the power chips that supply data centers. These seemingly disparate tracks reflect the market's re-evaluation of the bottlenecks in computing power and energy for the next generation of AI infrastructure.

I. The Night Before the Conference: Funds Rush, IONQ Shows Independent Market Trend

Data from after-hours trading on September 22 shows that the quantum computing sector experienced a very clear step-like surge:

|------------|--------|----------|----------|----------|------------------| | Asset | Ticker | Normal Close | After-Hours Quote | After-Hours Increase | Core Driving Factor | | IonQ | IONQ | $40.74 | $45.74 | +12.27% | Major Breakthrough: Released real-time quantum error correction decoder | | Rigetti | RGTI | $16.52 | $17.45 | +5.62% | Pre-conference sentiment spillover, combined with institutional buy ratings | | D-Wave | QBTS | $17.56 | $18.48 | +5.24% | Favorable quantum annealing route | | Quantinuum | QNT | $51.39 | $53.01 | +3.15% | Sector valuation linkage repair |

Note 1: The data in the table is taken from the after-hours trading period (Eastern Time 16:00--20:00). After-hours trading has lower volume and liquidity, and prices can be easily influenced by large single orders, not representing the opening price the next day. The pre-market increases quoted in the text (IONQ +11.75%, MPWR +8.55%) are based on the pre-market trading period on September 23 (Eastern Time 04:00--09:30), calculated using the previous trading day's normal closing price as the benchmark.

II. What is Quantum Computing? Where are the Industry Pain Points?

Quantum computing is not an incremental upgrade of traditional computers, but a structural transformation occurring at the foundation of computing power. The computing power architecture of classical computers is built on binary classical bits—while the number of states that n bits can encode is actually the same as that of quantum bits, the difference is that it can only occupy one state at any given moment. The distinction of quantum bits is that they can allow these states to coexist at the same time and amplify the probability of the correct answer through interference. If we use a book as a metaphor, classical bits are like a large book where you can only open one page at a time; quantum bits are like all pages being opened simultaneously, with each page able to interfere with the others.

The industry pain point that lies between theory and usable computing power is noise. Quantum bits are extremely delicate; even slight disturbances in temperature or electromagnetic fields can cause the superposition state to collapse instantly, leading to rapid accumulation of errors during computation, ultimately producing meaningless noise. The only solution to noise in the industry is quantum error correction: encoding hundreds or thousands of physical bits into a highly reliable "logical bit" and issuing correction instructions in real-time through an external classical computer. However, this has triggered a fatal time lag contradiction: the coherence time of quantum bits is only on the microsecond level; if the computing power of classical decoders cannot keep up, symptom data will accumulate rapidly, forcing the quantum computer to stop and wait, and with every microsecond of waiting, environmental noise creates new errors.

Previously, the industry's crude solution was to frantically pile on computing power, leading directly to an almost paradoxical "cost curse": for every additional logical bit, the cost and power consumption of classical hardware skyrockets nearly exponentially. The deepest concern in the capital market is precisely this—before quantum computers can commercialize, they may first be dragged down by their own massive, expensive, and power-hungry classical logistics system.

III. The Comeback of a Single CPU: What Did IonQ Do Right?

IonQ announced after hours yesterday that it has successfully developed and verified what it claims to be the world's first end-to-end real-time quantum error correction decoder, with the most notable breakthrough being that it can be driven by a single ordinary commercial CPU, achieving zero slowdown in quantum computing. Following this news, IONQ's stock price jumped over 12% in after-hours trading.

Setting aside the hardcore technical parameters, what does this breakthrough actually mean? If we compare fault-tolerant quantum computing to a high-speed racing car, then quantum error correction is its braking and steering system. In the past, to prevent the racing car from losing control due to noise, the industry had to attach a whole heavy server cluster for real-time computation—before the car could go fast, the logistics drag had already brought it down. What IonQ brings is precisely the key to uprooting this "logistics nightmare."

Key data in the announcement also supports this point: the decoder uses a dual-decoder architecture, completing verification on a circuit with up to 408 logical quantum bits and 88 memory blocks, processing over 31.5 million quantum operations, while under standard operational noise, the additional overhead is only 0.02%. It should be noted that this verification was completed on a simulated circuit, not yet a complete fault-tolerant computation running end-to-end on actual physical hardware—it is a key intermediate milestone, not the endpoint.

IV. AI Agents Are Cornering CPUs, How Can Quantum Break Through?

From Meta's Muse to recent strategic statements from various cloud vendors, the industry is sending a very clear signal: the battlefield of AI is fully shifting from large model training to the ecological landing of AI Agent entities. In the past arms race of large models, the focus of computing power was entirely on GPUs; however, Agents with autonomous planning and execution capabilities need CPUs the most.

According to a report by TrendForce, in the workflow of AI Agents, tool processing and logical orchestration account for 90.6% of the total system latency. Faced with complex logic that involves task decomposition, code execution, and multi-round state transmission—characterized by dense branching and strong serial dependencies—the parallel computing advantages of GPUs vanish, and the computing load is entirely placed on the shoulders of CPUs. Recently, Intel and AMD have both aggressively raised prices due to the shortage of data center CPUs, which is the most direct reflection of this shift in computing power structure in the supply chain.

It is precisely against this backdrop that the strategic role of quantum computing has been endowed with a new problem-solving logic. It is never meant to replace CPUs; rather, it aims to take over "black hole" tasks that are extremely CPU-intensive, such as combinatorial optimization and molecular simulation, thereby freeing up precious CPU computing power for the orchestration and tool execution of Agents.

This is also why IonQ's "single CPU + 0.02% extremely low overhead" solution is most appealing to the market: it means that when quantum computing is integrated into data centers, it will not only avoid competing for the currently scarce CPU and power resources but can also serve as a co-processor with a very low barrier to entry, enabling smooth and seamless deployment. This step significantly alleviates the core concerns of cloud vendors and enterprise customers regarding commercial implementation.

V. The Hard Constraint of Power: The Background of MPWR's 8% Pre-Market Surge

On the same day, Monolithic Power Systems (MPWR) surged 8.55% in pre-market trading, representing another pressing hard constraint—power and energy management.

Reports indicate that MPWR recently raised its earnings expectations, significantly increasing the lower limit of growth for its enterprise data business in 2026 from 85% to 130%. Previously, the Q2 financial report showed that this segment's revenue reached $380.6 million, a year-on-year surge of 164.3%, with revenue accounting for 38.8%; meanwhile, the company has received its first orders for DDR5 memory components and has begun sampling high-voltage AC-to-DC products for 800V data center architectures.

The core logic driving this strong growth is that as the power consumption of AI servers skyrockets geometrically, traditional data centers are overwhelmed and are being forced to migrate from the previous 12V/48V power supply to an 800V DC high-voltage architecture. This is not a gradual small fix, but a "blood transfusion-level" reconstruction of the entire power supply chain components. When the ultimate bottleneck of AI expansion shifts from "can we get GPUs" to "can the data center supply power," the power management chips at the core of power supply naturally gain significant pricing power.

This also vividly underscores the value of IonQ's extremely low power consumption solution—in a world where power has become an absolute hard constraint, any computing power solution that "does not consume extra power" inherently carries high commercial value.

VI. The Mainline Logic Behind the Nasdaq's New High: Recalculating AI's Three Accounts

On September 22 (yesterday), the Nasdaq Composite Index rose 0.45% to 27,244.28 points, setting a new historical high. Behind the index's new high, the trajectory of capital's game shows a clear rotation of sectors:

• During the opening phase on the 22nd: Storage and hard drive giants like Micron, Seagate, and SanDisk led the gains, as the complex long context and state caching of Agents were overwhelming memory bandwidth;

• After hours on the 22nd: IONQ drove a pulse-like surge in quantum computing, delineating a cost reduction curve for fault-tolerant computing with single CPU error correction;

• Pre-market on the 23rd: MPWR strengthened, announcing the importance of power supply and high-voltage architecture with better-than-expected guidance.

These three seemingly unrelated market movements actually point to the same core: Wall Street is comprehensively recalculating the three accounts of AI infrastructure—memory account, power account, and cost account. Any segment that can provide a definitive solution for enhancing storage bandwidth (storage sector), reducing system power consumption and costs (IonQ's zero-cost error correction), and addressing high-voltage power supply bottlenecks (MPWR power management) is receiving heightened attention from the capital market.

Disclaimer | All market quotes and data referenced in this article are as of September 23, 2026, at 12:00 PM (Eastern Time). Price changes, company announcements, and market fluctuations occurring thereafter are not included in this article. This article is compiled and analyzed based on publicly available market information, company announcements, and third-party research reports, involving a large amount of predictive content and comprehensive judgments on the reasons for individual stock price movements. The related attributions are speculations made by the author based on public information and do not represent the official statements of listed companies or any institutions. The actual commercialization pace of quantum computing technology, the sustainability of AI infrastructure capital expenditures, and the degree to which related companies meet performance guidance all carry significant uncertainty: IonQ's current decoder validation is completed on a simulated circuit and is not equivalent to running a complete fault-tolerant computation on actual physical hardware end-to-end; the Quantum World Congress lasts for three days (September 23 to 25), and the anticipated rise before the conference often faces pressure to realize during or after the meeting. Furthermore, the pre-market and after-hours gain data referenced in this article are from inconsistent statistical periods (as explained in Note 1), where liquidity is far lower than during normal trading hours, making prices susceptible to large single orders, and actual opening and closing performances may differ significantly from this. Everything is subject to the official published information and actual market trends. This article does not constitute any investment advice or offer, and the mentioned individual stocks do not represent any buy or sell recommendations. The author is not a licensed investment advisor; investors should independently assess based on their financial situation and risk tolerance and bear all risks arising therefrom. The market carries risks, and decisions should be made cautiously.