PsiQuantum

Learning Objectives

• Understand what makes PsiQuantum's silicon-photonic approach to quantum computing distinct from the more common superconducting and trapped-ion designs.
• Recognize why PsiQuantum is skipping the noisy near-term era and aiming straight for a million-qubit, fault-tolerant machine.
• Learn how to honestly assess a long-horizon infrastructure bet — including its funding, government backing, and the still-distant connection between quantum computing and AI.

What Is PsiQuantum?

PsiQuantum is a private company headquartered in Palo Alto, California, founded in 2016 by a team that includes Jeremy O'Brien and Terry Rudolph. While most quantum-computing companies build qubits from superconducting circuits or trapped ions, PsiQuantum has placed a different bet: it builds qubits from light itself, using silicon photonics.

The company's central thesis is unusual. Rather than racing to demonstrate small, noisy machines first, PsiQuantum is aiming directly at a fault-tolerant, error-corrected machine with roughly one million qubits — the scale most researchers believe is required to solve genuinely useful problems. Its argument is that the only credible path to that scale is to manufacture quantum chips the same way the world already makes conventional computer chips: at high volume, in existing semiconductor foundries.

The Photonic Bet

PsiQuantum's qubits are made from photons — individual particles of light — that travel through tiny waveguides etched onto silicon chips. Its photonic chip is called Omega, and the company reports gate fidelity of 99.999 percent on key components, a figure that matters enormously because error correction only becomes practical when the underlying operations are extremely reliable.

PsiQuantum's chips are fabricated at GlobalFoundries, a commercial semiconductor manufacturer. This is the heart of the bet: by tying quantum hardware to an existing industrial supply chain, the company hopes to reach the million-qubit scale that hand-built laboratory machines cannot.

Funding and Government Backing

PsiQuantum is the best-funded pure-play quantum company. It raised a one billion dollar Series E in September 2025 at a valuation of roughly seven billion dollars, drawing investors that include BlackRock, Temasek, Baillie Gifford, and NVIDIA's venture arm, NVentures. It also secured a roughly one hundred million dollar letter of intent through the US Commerce Department's CHIPS program in May 2026, and is building government-backed data centers in Brisbane, Australia, and in Chicago.

The scale of this backing reflects the size of the wager. A million-qubit machine is a multi-year, capital-intensive undertaking, and the involvement of sovereign wealth funds and government programs signals that quantum computing is increasingly viewed as strategic infrastructure rather than a pure science project.

The AI Angle

The connection between PsiQuantum and AI is indirect and, honestly, still far off. The most concrete link is financial: NVIDIA's venture arm is an investor, which reflects the broader industry interest in how advanced computing platforms might one day complement each other.

The deeper theme is that fault-tolerant quantum machines could eventually accelerate certain kinds of simulation — modeling molecules, materials, and physical systems — that are extraordinarily expensive for classical computers, including the GPU clusters that power today's AI. If that promise is realized, quantum simulation could feed better data and better models into scientific AI workflows. But this remains a long-term possibility. Quantum-for-AI is not a near-term reality, and it would be misleading to suggest that PsiQuantum changes anything about how AI is built or used today.

How to Follow It

There is no public cloud, software development kit, or trial to sign up for. The way to engage with PsiQuantum is to read its published research and roadmap statements and to track milestone announcements over time.

Strengths

• A genuinely differentiated approach — silicon photonics — that sidesteps the deep cryogenic refrigeration required by superconducting designs.
• A manufacturing strategy tied to existing commercial foundries, which is the most credible known path to extreme qubit counts.
• Strong reported component fidelity (99.999 percent on key operations), which is foundational to practical error correction.
• The best funding position among pure-play quantum companies, with deep-pocketed and sovereign-backed investors.
• Government partnerships and data-center construction that signal long-term institutional commitment.

Limitations & Considerations

• No product, no cloud access, and nothing to use today — this is a research and infrastructure bet, not a tool.
• The million-qubit, fault-tolerant target is still years away, with useful machines aimed at around 2027 and beyond — and timelines in quantum computing routinely slip.
• Skipping the near-term noisy era means there are fewer interim demonstrations to validate progress against competitors.
• The AI connection is speculative and long-horizon; nothing here affects how AI systems are trained or deployed in the present.

Best Use Cases

This page is for understanding and tracking PsiQuantum, not for using it. The table below frames where attention is best spent.

Getting Started

1. Read PsiQuantum's published overview and roadmap statements on its website to understand the million-qubit thesis in the company's own words.
2. Study the silicon-photonics concept — why particles of light, fabricated on standard wafers, are the company's chosen path.
3. Track milestone announcements over time, treating timeline claims with healthy skepticism given how often quantum schedules shift.
4. Watch the broader field — including superconducting and trapped-ion efforts — to keep PsiQuantum's progress in context rather than in isolation.

Key Takeaways

• PsiQuantum is a private, pre-product company betting that the path to useful quantum computing runs through silicon photonics — qubits made from particles of light.
• Its central wager is to skip the noisy near-term era and build directly toward a million-qubit, fault-tolerant machine, manufactured at scale in commercial foundries like GlobalFoundries.
• The company is the best-funded pure-play in quantum, raising one billion dollars in 2025 at roughly a seven billion dollar valuation, with government backing and data centers under construction in Brisbane and Chicago.
• The AI connection is indirect and long-horizon — NVIDIA's venture arm is an investor, and fault-tolerant machines could one day aid scientific simulation, but quantum-for-AI remains far off.
• There is no product to use today; this is a forward look at a roadmap targeted around 2027 and beyond, best treated as something to understand and follow.