Quantum computing stocks are rising quickly, but investors will need to be strapped in for a long ride.

Model of the quantum processor computer.

Market reaction has been fierce in the wake of Google-parent Alphabet’s Dec. 9 announcement of its new Willow quantum computing chip. Alphabet shares are up 7% since the reveal, but more of the price action has been focused on three much smaller companies with a combined market capitalization of under $8 billion, Quantum Computing, Rigetti Computing, and D-Wave Quantum, which are up by an average of 159%.

Another player in quantum computing is IBM, with its shares down 4% since Google announced Willow. At least one analyst thinks Big Blue is being overlooked.

“IBM thinks it is building the most comprehensive quantum solution in the world,” writes Ben Reitzes of Melius Research in a research note. “And they seem right despite some big and small competitors.”

John Preskill, director of the Institute for Quantum Information and Matter at the California Institute of Technology told Barron’s that IBM has “an excellent team and are making impressive progress on scaling up their hardware and improving its performance.”

But investors should temper their excitement. While Willow represents an advance in technology with its reduced error rates, the promise of quantum computing is still years off.

“There’s good enough, and then there’s practical,” Quantum Circuits Chief Scientist Rob Schoelkopf told Barron’s. “For useful computing, we need vastly lower error rates.”

Market research firm IDC estimates that the total commercial sales for quantum computing will be $8.9 billion in 2028. That’s equal to 16 days of iPhone sales, or 13 days of Google ad sales.

And even that modest number may be too optimistic. “I am skeptical but would be delighted to be proved wrong,” physicist Steven Girvin of Yale University’s Quantum Institute told Barron’s. “Progress is very rapid at the moment, but the goal is still quite a distance off.”

Even with its bullish note, Melius was clear that this isn’t a project that will bring a revenue boost soon.

“We forecast quantum could result in multiple billions of dollars in revenues and profits in the 2030s for IBM with a ‘best guess’ at materiality in 2029,” Melius wrote to clients.

Preskill speculated “quantum computers that can run a variety of applications with business value could be a couple of decades away.”

Quantum represents the next big step in computing. The computers we have today are based on binary math. Chips are composed of billions of tiny transistors, each representing a “bit” of data which has only two states—on or off. Quantum computers have “qubits” which can have multiple states, and they can process data much faster as a result, significantly improving performance in certain tasks like complex cryptography, optimizations and simulations, at least in theory.

In Google’s test of quantum memory—storing and retrieving data—Willow completed a benchmark test in under five minutes that would take a current supercomputer 10 septillion years. This shows the potential for quantum computing on certain types of tasks.

But there remain limitations, primarily the unacceptably high error rate. Though Google Willow shows progress in this regard, there are still big hurdles for all the players to overcome before quantum computing becomes a commercial product, and it is unclear how long that will take.

“I will keep saying ‘10 years from today’ until I am proven wrong,” Girvin said. “Just as the inventors of the transistor, laser and atomic clock could not even foresee what these devices would be used for outside the laboratory, I don’t think we can yet foresee all the uses of this new technology.”