I Got Rich Off NVIDIA. This Is Even Bigger.

At CES 2025 in Las Vegas, Jensen Huang told the world that useful quantum computers were 15 to 30 years away. The market reaction was instant and brutal: Ion Q dropped 39% in one day, Rigetti fell 45%, and D-Wave crashed 36%. One sentence from one CEO cratered an entire class of stocks.

Then something strange happened. Two months later at Nvidia GTC 2025, Jensen changed his tune, saying quantum was closer than he thought and hosting the first-ever Nvidia Quantum Day. Three months after that in Paris, he said quantum was reaching an inflection point and would solve real problems in the coming years. Those comments fueled a massive bubble in quantum stocks late last year.

Now the bubble is behind us. Two weeks ago, Nvidia shipped Icing, an open-source software stack that splits problems between GPUs and quantum computers, with Ion Q named as an early adopter. In ten years of investing in Nvidia, the host has never seen Jensen change his stance on any technology this fast.

Nvidia isn't building a quantum chip. Instead, they're building the infrastructure between everyone else's quantum chips and their own GPUs. Late last year, Nvidia announced NVQLink, a super-fast connection that lets GPUs send instructions and read results in just a few millionths of a second — fast enough to watch qubits, correct errors, and adjust on the fly before decoherence kicks in.

Two weeks ago, they released Icing, an open-source AI toolkit that helps quantum chips fix their own errors in real time and speeds up calibration from days to hours. As an investor, this made one thing crystal clear: Nvidia doesn't think quantum computers will replace GPUs, or else they wouldn't build systems to support them. Instead, they see quantum plugging into current data center infrastructure for very specific workloads. The future will be hybrid computing — a mix of CPUs, GPUs, and QPUs all working together.

Ion Q is the closest thing to being the Nvidia of quantum computing. They use trapped-ion systems: charged atoms held in place by electromagnetic fields and controlled with lasers. This gives Ion Q some of the highest-quality qubits in the industry — slower, but with fewer errors and longer coherence times. Critically, trapped-ion systems are the easiest to network together.

Last week, Ion Q proved this by announcing the first photonic interconnect between two quantum processors. The path from 100 qubits to 10,000 is no longer science fiction. Ion Q reported $62 million in revenue last quarter, up 429% year-over-year, and $130 million for the full year — the only public quantum company to cross $100 million. Their backlog grew to $370 million, almost three times last year's revenue, and they're guiding for $235 million in 2026, implying 80% growth.

They have $3.3 billion in cash and short-term investments, a massive war chest for a company this small. But they reported an operating loss and expect larger losses in 2026 as they invest in scale. For investors, Ion Q has the biggest backlog and strongest balance sheet, but you're still betting on an unprofitable, highly volatile stock with real execution risk.

If Jensen Huang is right that quantum will work with GPUs instead of replacing them, Rigetti is set up to win big. It's already plugged into Nvidia's NVQLink and CUDAQ stack for hybrid workloads, meaning any enterprise wanting Nvidia's stack plus quantum will run on Rigetti hardware. Rigetti's superconducting systems use tiny electrical circuits cooled near absolute zero, running about 4,000 times faster than Ion Q's trapped ions — but with so many more errors that they're only about 10 times faster in practice.

Rigetti is the only way to invest directly in superconducting quantum without buying IBM or Google and all their other business units. They have the first 100-plus-qubit processor available on AWS, giving millions of developers instant access. But their Q4 2025 revenue was just $1.9 million, down 18% year-over-year, and full-year revenue was $7.1 million. Rigetti is still pre-commercial.

They finished the year with about $590 million in cash and burn about $20 million per quarter, giving them roughly 7.5 years of runway. Don't mistake lack of revenue for lack of funding. For investors, Rigetti is the smallest, highest-risk, highest-reward bet on Nvidia's quantum future.

D-Wave is the only company with a commercial product that customers pay for today — not pilots, not research partnerships. Their quantum annealing systems are specialized machines built to solve optimization problems like scheduling, routing, and materials design. Companies like Volkswagen, Google, NASA, and Lockheed Martin already pay to run workloads on D-Wave systems.

In January 2025, D-Wave closed a $550 million acquisition of Quantum Circuits, Inc., giving them a gate-based quantum platform on top of their annealing business. Gate-based superconducting is more general-purpose, complementing D-Wave's specialized systems. For investors, D-Wave has two different quantum architectures under one roof — they don't have to win the war to survive it, only one needs to work out long term.

Last quarter, D-Wave's revenue was $2.8 million, up 19% year-over-year, but full-year 2025 revenue came in at $24.6 million, up 179%. In January and February 2026 alone, they reported $32.8 million in new bookings — already more than their entire 2025 revenue in just two months. They have $630 million on the balance sheet and burn around $70 million per year, giving them almost nine years of cash runway. That optionality is the unique thing about D-Wave.