BlackRock, Temasek and Nvidia are not treating quantum computing like a science project anymore. They are buying early positions in the infrastructure layer that could sit beside GPUs, not politely waiting for the field to become obvious.
Quantum computing has spent years as the technology everyone praised in public and discounted in private. Useful machines were always coming later. Commercial revenue was always thin. The stocks were always too easy to mock. Then the money started arriving in sizes that are harder to ignore.
The Wall Street Journal reported in September 2025 that PsiQuantum raised $1 billion in a Series E round led by BlackRock, Temasek and Baillie Gifford, with Nvidia's venture arm joining as a new investor. The round valued the Palo Alto company at $7 billion. PsiQuantum is not promising a modest lab machine. It is trying to build a one-million-qubit, fault-tolerant photonic quantum computer, with large systems planned for Brisbane by the end of 2027 and Chicago in 2028.
You don't have to believe every date in that plan to understand the signal. A $1 billion private round is not a grant for interesting physics. It is a claim on a future computing supply chain: chips made at GlobalFoundries in New York, photonic systems that need data-center-scale facilities, and software work that has to make those machines useful when they arrive.
Quantinuum then gave public-market investors their own test. MarketWatch reported that the Honeywell-backed company raised $1.68 billion in an upsized IPO on June 4, 2026, selling 28 million shares at $60 each and listing on Nasdaq under the ticker QNT. Its shares opened at $68 before closing at $60.38. The pricing valued the company at about $15.6 billion, even though Quantinuum generated only $30.9 million in 2025 revenue and lost $192.6 million.
That last comparison is the whole story in miniature. A $15 billion valuation on $31 million of revenue is not a normal software multiple with a little optimism sprinkled on top. It is a bet that fault-tolerant quantum computers become scarce, strategic machines, and that the companies with the best hardware, error correction and customer access get priced before the revenue turns up.
Frankly, that is exactly how frontier infrastructure gets funded. It looks excessive until it doesn't.
Nvidia is buying the bridge, not replacing the GPU
Nvidia's role is the easiest part to misunderstand. Jensen Huang said in January 2025 that very useful quantum computers were probably 15 to 30 years away, and quantum stocks fell hard after the remark. That sounded bearish. It wasn't a reason for Nvidia to sit still.
NVentures has since put money into PsiQuantum, QuEra and Quantinuum. Those are three different bets on three different hardware paths: photonic systems at PsiQuantum, neutral atoms at QuEra, and trapped ions at Quantinuum. If you are Nvidia, that is not panic. It is disciplined option buying.
The practical reason is error correction. Quantum processors are fragile, and useful logical qubits need classical systems constantly checking, decoding and correcting errors. That keeps GPUs in the room. Huang's skeptical timeline and Nvidia's investments can both be true because near-term quantum machines still need classical acceleration wrapped around them. Nvidia wants CUDA-Q, its quantum software stack, and its AI hardware tied into that work before someone else writes the standard.
PsiQuantum's collaboration with Nvidia points in the same direction. The companies are working on integrating quantum hardware with AI chips and developing quantum software and algorithms, according to the Journal. That is not a moonshot press release. It is plumbing. And plumbing is where platform companies usually make their money.
The real work is still error correction
For founders, the useful question is not whether quantum is real. That is too broad to help you. The better question is where the bottleneck sits today.
Right now, it sits in error correction, control systems, cryogenics, software tooling and hybrid workflows. Quantinuum's Helios trapped-ion processor has been used in research showing computations with between 48 and 94 logical qubits beyond break-even performance. QuEra-linked research has shown neutral-atom systems encoded with up to 96 logical qubits. Those are not mass-market computers, but they are the kind of milestones that change investor behavior because they move the conversation from physics demonstration to systems engineering.
Reuters reported in February 2026 that Finland's IQM Quantum Computers was preparing to list in the US at an initial valuation of about $1.8 billion. IQM builds superconducting quantum computers, the kind that depend on scarce dilution refrigerators and highly specialized hardware supply chains. That detail matters. The quantum race is not only about qubit counts. It is also about who can source the equipment, cool the systems, connect them to classical machines and keep them stable long enough for real workloads.
Temasek's interest belongs in that frame. Singapore's sovereign wealth ecosystem has backed quantum through direct exposure and through funds such as Quantonation's second fund, which closed at €220 million in February 2026 with support from Vertex Holdings, a Temasek subsidiary. Governments and sovereign investors have worked out that quantum is now a competitiveness question, not only a venture category.
You should still be skeptical of clean timelines. PsiQuantum's Brisbane machine is supposed to be online by the end of 2027. Quantinuum says it is working toward a commercial-scale, fully fault-tolerant quantum computer before the end of the decade. Those promises may slip. Big technical promises often do.
But the investment pattern has changed. BlackRock, Temasek and Nvidia are not betting that every quantum company wins. They are betting that the category has crossed from research curiosity into infrastructure formation. That is the signal worth tracking.
Also read: Rocket Lab is buying Iridium for $8 billion to build the vertically integrated space company Starlink forced everyone to consider • Coinbase halved its AI bill without restricting engineers and the playbook is worth stealing • China is turning humanoid robots into an industrial assembly line and the rest of the world is watching
