Quantum Computing: The U.S. bets on 9 companies as industry turning point arrives

In 1947, the Danish royal family awarded Niels Bohr a knighthood.

This pioneer of quantum mechanics designed a very special family crest: instead of a lion, crown, or shield, it features a Taiji diagram. Around the pattern, an inscription in Latin reads: Contraria sunt complementa, meaning "Opposites are complementary."

This was one of Bohr’s most important ideas: electrons behave both as particles and waves; light exhibits both particle and wave properties. These seemingly conflicting descriptions are not mutually exclusive but together describe the same world.

Interestingly, a hundred years later, when we revisit quantum computing, we still cannot escape this Taiji symbol. Quantum computing isn’t about making old computers faster; it’s about acknowledging that the fundamental nature of the world isn’t strictly black or white, nor 0 or 1. It’s more like a gray area, flowing and full of possibilities between 0 and 1.

For a long time, quantum computing was considered a science far removed from everyday reality. It has Nobel-level physics foundations, countless papers, and laboratory breakthroughs, but it still seemed distant from ordinary life and market pricing, always shrouded in a layer of fog.

Now, the situation has changed.

On May 21, 2026, the U.S. Department of Commerce announced: under the CHIPS and Science Act, it signed memoranda of understanding with 9 quantum-related companies to provide $2.01B in federal incentives. As a condition, the U.S. government would hold minority, non-controlling equity stakes in these companies.

This marks another strategic move by the U.S. government into a key industry, following investments in Intel, rare earths, lithium, and other critical sectors. The impact extends beyond the surge in quantum company stock prices; more importantly, the U.S. has officially listed quantum computing as a national industry to be prioritized, not just a future technology.

With private capital and state investment both pouring in, and the U.S. government participating via equity, quantum computing is no longer just cutting-edge research in labs but has become a new industry that investors must understand:

• How will it change the real world?
• Who is leading the key technological pathways?
• Which companies are already on the scene?

1. What is quantum computing?

1. Limitations of classical computing

Before discussing quantum computing, it’s essential to understand the current classical computing landscape—from personal PCs to supercomputers—which forms the foundation of our entire world.

The smallest unit of a classical computer is called a bit, which can only be 0 or 1. Think of it like a switch: either on or off.

A photo, a video, a bank transfer, or an AI model—all can ultimately be broken down into vast sequences of 0s and 1s.

For example, the word "Apple" on your computer isn’t directly "understood" as text. It’s first broken into characters: A, p, p, l, e. Each character has a code; in early ASCII, A corresponds to 65, which in binary is 01000001; p corresponds to 112, which is 01110000. So, the word "Apple" becomes a string of 0s and 1s at the lowest level. Then, the computer uses font files to know how each letter should look, and pixel data to decide which dots are lit or dark, and what colors to display. Only then do we see the complete "Apple" on the screen.

Thus, classical computers don’t understand text, images, or videos in themselves. They merely translate everything into 0s and 1s, then process these at extremely high speeds. The modern digital world relies on this "clumsy" approach. It’s powerful: over the past decades, all internet, smartphones, games, cloud computing, and AI have been built on this binary foundation.

But the bit has its limits. Some problems aren’t just about "speed," but about the sheer number of possibilities—so vast that even with all the computing power on Earth, classical computers can’t solve them in a realistic timeframe. For example, a 100-bit binary password has 2^100 possibilities. Using a top-tier personal computer to brute-force it, even in a simplified hash scenario, would take about 1.8 trillion years.

If the password length increases to 128 bits, even with the fastest supercomputer (like El Capitan), and assuming one attempt per operation, it would take roughly 60 trillion years—about 430 times the age of the universe (138 billion years).

For 256 bits, the time needed would be 1.45 × 10^41 times the universe’s age—an unimaginably long span, far beyond the universe’s lifespan.

Humans’ continued speed-up in chip performance can’t solve these exponential problems.

Classical computers typically have two strategies for such exponential growth issues:

• Try brute force until it’s impossible to wait;
• Use approximate algorithms for certain problems, accepting results that are "good enough" but not necessarily optimal.

Thus, humanity has been seeking a paradigm shift in computation for decades.

2. The astonishing promise of quantum computing

The fundamental unit of a quantum computer isn’t called a bit but a qubit, or quantum bit. Unlike classical bits, which are either 0 or 1, qubits can exist in a superposition of both states before measurement.

This sounds strange. To illustrate, imagine drawing two cards: a king and a queen. You place one face down on the table without revealing it. It’s either a king or a queen, but once you look, it’s definitely one or the other.

In contrast, a superposition is like the card being both king and queen simultaneously until you observe it. Only when you look does it "collapse" into one state—either king or queen. This phenomenon is so counterintuitive that it can be frightening: our observation actually influences the outcome. It challenges our understanding of reality.

Of course, this example is simplified. In quantum mechanics, "observation" isn’t just "looking" but involves measurement devices and environmental interactions that change the system, producing different results.

A classical bit is deterministic: it’s either 0 or 1.

A qubit is probabilistic: before measurement, it exists in a superposition, and only after observation do we find it as 0 or 1.

In classical computing, two bits can only be in four states at once: 00, 01, 10, or 11.

In quantum computing, two qubits in superposition can represent all four states simultaneously, with certain probabilities.

• Three qubits can represent 8 states.
• Ten qubits, 1024 states.
• Fifty qubits, about ten million billion states.
• Three hundred qubits, more states than the atoms in the observable universe.

How does this quantum property translate into computation? It requires quantum algorithms that amplify the probability of correct answers while diminishing incorrect ones, so that when measured, the right answer is more likely to appear.

A simple analogy: classical computers are like searching in the dark, trying one path at a time among a million. If one path is wrong, you backtrack and try another.

Quantum computers, however, create a wave interference pattern among all paths, effectively "smoothing" the search and amplifying the correct solutions.

Quantum computing fundamentally differs from classical:

Classical computing tries solutions step-by-step. Quantum computing uses superposition, interference, and probability amplification.

This is its core distinction.

• Classical computers, even if faster, still operate on binary logic.
• Quantum computers leverage the laws of the microscopic world: superposition, interference, and measurement.

For password cracking, classical computers try each possibility individually. Quantum computers, in some scenarios, can evaluate many possibilities simultaneously, potentially offering shortcuts.

Quantum computing is more akin to natural "theology": classical simulation of phenomena like storms is approximate and computationally expensive, but quantum computing, being a part of nature itself, can more directly access the underlying rules. Richard Feynman famously said: "Nature isn’t classical, after all. When you want to make a simulation of nature, you’d better make it quantum mechanical."

The universe is fundamentally quantum. Humanity will eventually need a machine that operates according to quantum laws to understand and compute this quantum world.

3. How will quantum change the world?

Quantum computing isn’t a panacea. For everyday tasks—watching videos, spreadsheets, gaming, training large models—classical computers remain optimal. Quantum computers won’t be faster for these; they might even be slower.

Its true value lies in specific problems: those with enormous state spaces, where solutions are hidden among astronomically many possibilities, and where the problem’s structure can be exploited via quantum interference. In such cases, the speedup isn’t just 2x or 10x but can leap from "impossible to compute" to "feasible."

Three main problem types stand out:

First, cryptography

Today’s internet security—online banking, messaging, government communications—relies heavily on RSA, ECC, and other public-key cryptosystems. In 1994, Peter Shor at Bell Labs proposed the Shor algorithm. It proved that a sufficiently large, fault-tolerant quantum computer could factor large numbers efficiently, breaking RSA encryption much faster than classical algorithms.

This is the so-called Q-Day, or "quantum apocalypse."

Once powerful quantum computers arrive, many current encrypted communications—financial data, government secrets—could be decrypted.

Even more alarming: "record now, decrypt later." Attackers could store encrypted data today, then decrypt it in the future once quantum computers are available—rendering current security measures obsolete.

This poses a huge threat: our entire digital infrastructure depends on cryptography. When quantum arrives, the security foundations must be rebuilt in advance.

Second, molecular simulation

In 1981, physicist Richard Feynman proposed quantum simulation as a primary motivation for quantum computing. Molecules are governed by quantum mechanics; simulating their electronic structure is exponentially hard for classical computers.

Quantum computers, being quantum systems themselves, can naturally simulate other quantum systems. This could revolutionize drug discovery, new materials, batteries, and catalysts by providing more accurate models of molecules and reactions.

If successful, it could dramatically reduce the time and cost of discovering new medicines, materials, and energy solutions.

Third, combinatorial optimization

Many real-world problems—logistics routes, chip wiring, flight scheduling, investment portfolios, manufacturing schedules—are about finding the best solution among countless options.

For example, the Traveling Salesman Problem: a delivery driver must visit multiple locations once and return home, minimizing total travel distance. As the number of locations grows, the possible routes explode combinatorially—20 locations yield trillions of options; 30 locations, over 10^30.

Classical algorithms struggle with such scale. Quantum algorithms could, through superposition and interference, increase the probability of finding better solutions more efficiently.

In summary, quantum computing isn’t meant to replace smartphones or GPUs but to solve specific, hard problems in cryptography, chemistry, materials, finance, and defense—areas that underpin the entire digital and physical world.

4. The key milestones in quantum computing’s leap

Qubits are fragile: temperature fluctuations, electromagnetic noise, mechanical vibrations all cause errors. To make quantum computers practical, engineers must combine many physical qubits into a more stable "logical qubit."

A critical concept is the error correction threshold. Imagine many people copying a text: if everyone makes too many mistakes, the overall copy is unreliable. But if errors are rare, multiple copies can be combined to correct mistakes.

Quantum error correction works similarly.

When the error rate of physical qubits exceeds a certain threshold, adding more qubits only introduces more noise, making the system worse.
When the error rate is below this threshold, adding qubits allows for error correction, creating more stable logical qubits. The larger the system, the lower the logical error rate.

This is called crossing the error correction threshold—a phase transition from "more qubits, more noise" to "more qubits, more stability."

The first successful crossing occurred in December 2024, when Google’s Willow chip achieved an error suppression factor Λ = 2.14, meaning that increasing the code distance by 2 roughly halved the logical error rate. Since then, companies like Quantinuum, Zuchongzhi 3.2, and QuEra have crossed similar thresholds via different approaches.

Once this threshold is crossed, the discussion shifts from "Can we build it?" to "When will it be built?"

The next year or so is critical.

Part Two: The rapid advance of quantum

Since Google’s Willow release about a year and a half ago, many significant developments have occurred.

Clear structural inflection point!

1. Private and policy capital are both betting heavily

Market data is more direct.

QED-C reports that by the end of 2025, global public funding commitments to the quantum industry reached $56.7 billion. In the same year, venture capital investment in quantum startups totaled $4.9 billion, with U.S.-based companies securing $2.7 billion, nearly 60% growth over 2024’s $1.7 billion.

These figures are before the U.S. government’s $2 billion investment announced on May 21.

Over the past five years, private funding for quantum companies mainly supported fundamental research. The $2 billion announced in May is different: it’s for industry infrastructure. IBM received $1 billion to build America’s first dedicated quantum wafer fab; GlobalFoundries got $375 million for low-temperature CMOS control chips and packaging lines, and on the same day, established a Quantum Technology Solutions division to handle foundry orders.

These two companies received $60k, accounting for 68% of the total. The remaining $638 million was distributed among seven companies working on different approaches, with six receiving $100 million each, and Diraq getting $38 million.

2. What impact does this have on AI revolution?

The answer goes back to Feynman’s 1981 insight: classical computers can never accurately simulate quantum systems because they are governed by fundamentally different physics.

AI, especially large models, is essentially an engineered statistical inference system. It learns patterns in human language, images, videos, but cannot physically solve quantum problems faster than classical computers. GPT-5 can tell you roughly what a molecule looks like, but can’t precisely compute its electron cloud distribution—an inherently quantum mechanical problem.

AI excels at "statistical pattern extraction," while quantum computing aims at "physical simulation." They are different domains, each with its own limits and applications. Breakthroughs in drug discovery, energy, materials, and cryptography will require machines that are physically isomorphic to quantum systems—not just faster GPUs.

That’s why, on May 21, IBM announced a $1 billion investment to build a foundry, not another AI data center.

3. Time is pressing for everyone

First, opportunity. If quantum computing becomes practical between 2029 and 2033, whoever controls the upstream supply chain—chip fabrication, key materials, control systems—will have a ten-year window of dominance. This is a strategic opportunity comparable to TSMC or ASML. Entrepreneurs, investors, and nations must act now.

Second, threat. If any country reaches Q-Day—"quantum apocalypse"—first, it could break the strongest encryption, rendering current internet security obsolete overnight. Past encrypted data, if stored now, could be decrypted later. This could threaten banks, secret communications, missile and nuclear command systems, and more.

The U.S. investment isn’t just "subsidy" but a bet and a defense.

4. The three stages of industry development

After crossing the inflection point, who will win? Predicting the future is hard, but we can reduce uncertainty with a logical framework. The development will likely proceed in three phases:

First, validation. The first company to demonstrate a machine that outperforms classical computers on a real problem wins a ticket to the industry. Companies like IBM, Google, Quantinuum, IonQ are competing here. This moment will be akin to the advent of ChatGPT—except that, for you reading this, it’s a moment to start preparing mentally.

Second, specialization. Quantum computing will initially target high-value, narrow problems: drug discovery, materials simulation, chemical reactions, cryptography, financial optimization, national defense. These problems are "narrow but high-value." Success here will be the "GPT moment" for applied quantum.

Third, platform. If the number of logical qubits continues to grow, error rates decline, and software ecosystems mature, quantum computing will evolve from a "specialized machine" into a computing platform. It will become a cloud service with development tools, algorithms, and industry solutions—similar to today’s AI industry—offering countless opportunities.

Monitoring the development steps and key players is more important than obsessing over short-term price movements.

Who are the players on the table?

Like the AI industry, quantum computing will be layered. I broadly categorize it into three levels:

1. Hardware manufacturing layer

This is akin to AI’s compute infrastructure: quantum chips, wafers, packaging, cryogenic systems, control chips, lasers, photonic devices, dilution refrigerators, etc. It determines whether quantum can move from lab to industry. Companies like IBM, GlobalFoundries, SkyWater, Origin Quantum, Diraq are highly relevant here.

Unlike traditional chips, quantum hardware lacks a unified approach. There are multiple competing technologies—superconducting, ion traps, neutral atoms, photonics, silicon spins, topological qubits—and it’s still uncertain which will prevail. The core question is: What physical platform will produce the most stable, controllable, and cost-effective qubits?

• Superconducting: uses superconducting circuits cooled to near absolute zero; represented by IBM, Google, Rigetti, Origin Quantum.
• Ion traps: trap single ions with lasers; represented by Quantinuum, IonQ.
• Neutral atoms: use laser tweezers to arrange atoms; QuEra, Atom Computing, Infleqtion.
• Photonic: use light particles in integrated photonic circuits; PsiQuantum, Xanadu.
• Silicon spin: use electron spins in silicon; Diraq, Intel.
• Topological: aim to create inherently error-resistant qubits; mainly Microsoft, still in research.

The choice of hardware approach is fundamental but belongs to the hardware and manufacturing layer, not a separate category.

2. Software and algorithms layer

Having hardware alone isn’t enough. Like NVIDIA’s CUDA for GPUs, quantum computing needs programming frameworks, compilers, error correction software, industry-specific algorithms, and cloud access. Companies like IBM (Qiskit), Quantinuum, IonQ, and others are competing here.

3. Application layer

This is the most distant from maturity but with the greatest potential. Fields like new drugs, materials, batteries, finance, cryptography, and defense will each have their stories.

However, this layer is prone to hype. The phrase "future applications in medicine, materials, finance, defense" sounds promising but doesn’t mean revenue today.

From an investment perspective, the key questions are: Are there real customers? Are they paying continuously? Is quantum necessary for this?

But this layer is still early.

How should we value quantum companies?

In reality, most pure-play quantum companies are overvalued by traditional metrics. Price-to-sales ratios of dozens or hundreds are common. Revenues are often only a few tens of millions of dollars, yet market caps can reach billions or hundreds of billions. From a mature company’s perspective, this looks like a bubble.

But calling it a bubble oversimplifies. Early-stage "hard tech" valuations are not about current profits but about future industry positioning. During the boom, many companies will rise; after the tide recedes, only a few will grow into giants.

Valuation is very challenging. From an investor’s view, the priority is risk mitigation and capital preservation, which suggests a two-layer logic:

1. Focus on whether the core business is solid

This applies mainly to IBM and GlobalFoundries.

IBM’s quantum division, even if it fails, won’t go to zero. It has software, consulting, mainframes, hybrid cloud, enterprise clients, and free cash flow. Quantum is a long-term call option.

Valuation should be: core business cash flow as the floor, quantum business as the ceiling.

These companies may not be the fastest risers, but they offer stability: investors don’t need to worry about bankruptcy before the next funding round.

This is crucial in hard tech. Many great innovations fail not because of physics but because of cash flow issues. Similarly, GigaDevice is a foundry; its quantum business is an extension. If demand for quantum control chips, low-temperature CMOS, or advanced packaging materializes, they will benefit. If the industry stalls, they still have their core foundry business.

Such companies are best valued as "core business + quantum option."

2. How much is the option worth?

This applies to IonQ, Quantinuum, D-Wave, Rigetti, Infleqtion.

Valuation here depends less on current earnings and more on whether their technological route can succeed.

Investors should consider or monitor:

• Does their chosen approach have a physical advantage?
• Can the company survive to the next critical milestone?
• Are there real customers?
• Are technical metrics improving?
• Is the valuation overly optimistic about future growth?

Great industries don’t always mean great investments. Overpaying can take years to recover. The hardest part of quantum investing now is: you may pick the right direction but buy at the wrong price or wrong target.

The report here covers IBM and GigaDevice, the two most promising quantum stocks, with updates to follow.

With this, the landscape of quantum computing is roughly clear: some are building machines, some are refining foundations, some are developing software, and some are waiting for applications to explode. Some will become the next infrastructure layer; others will fade away.

Classical computing built the digital world of the past; quantum reminds us that the universe’s fundamental layer is older and deeper than 0 and 1.

It’s not here yet, but it will come—inevitably, in the most natural way consistent with the laws of creation.