Sandisk (SNDK 5.69%) is one of several semiconductor companies that has benefited from a severe supply shortage in memory chips driven by demand for artificial intelligence. The stock has advanced 1,050% since being spun off from Western Digital in February 2025.
Meanwhile, D-Wave Quantum (QBTS 5.71%) is one of several pure-play quantum computing companies that has captivated investors. The stock has advanced 1,900% since January 2023.
In the third quarter, Citadel Advisors (a hedge fund run by billionaire Ken Griffin) sold nearly 2 million shares of Sandisk, cutting its position 97%. Citadel also bought stock in D-Wave Quantum. Neither position is/was very large, but the trades are still noteworthy because Citadel is the most profitable hedge fund in history.
Here's what investors should know about Sandisk and D-Wave Quantum.
Image source: Getty Images.
Sandisk: The stock Citadel sold in the third quarter
Sandisk is a semiconductor company that develops data storage solutions based on NAND flash memory for edge devices (computers, smartphones) and data centers. Central to its business is a joint venture with Japanese manufacturer Kioxia. Both companies realize cost efficiencies and supply chain security by sharing R&D expenses and capital expenditures related to designing and producing memory chips in-house.
Sandisk also has another important advantage in its vertically integrated business model. The company not only manufactures memory wafers through its joint venture, but also it packages wafers and integrates chips into final products like solid state drives (SSDs). That lets Sandisk optimize the performance and reliability of its storage products in ways that other (less vertically integrated) suppliers cannot.
Importantly, while Sandisk's non-GAAP earnings fell 33% in the first quarter, the company expects adjusted earnings to jump 160% in the second quarter because of an unprecedented supply shortage in memory chips created by insatiable demand for artificial intelligence (AI) infrastructure. Experts generally think the supply shortage will persist through 2026, which portends strong growth for several quarters to come.
Indeed, Wall Street estimates Sandisk's adjusted earnings will increase at 259% annually through the fiscal year ending in June 2027. Under different circumstances, that would make the current valuation of 170 times earnings look reasonable. But the memory chip industry is notoriously cyclical, and current supply constraints probably mean we are nearing a peak.
So what? Once memory chip supply begins to outpace demand, Sandisk will quickly lose pricing power and the market will probably afford the stock a much lower price-to-earnings multiple. In other words, shares could fall (perhaps sharply) at some point in the future. Of course, no one knows when that day will come, but Ken Griffin's decision to sell makes sense in that context.
NASDAQ: SNDK
Key Data Points
D-Wave Quantum: The stock Citadel bought in the third quarter
D-Wave develops superconducting quantum computing systems. The company is best known for quantum annealers, though it also develops gate-based machines. Annealers are designed for optimization problems, whereas gate-based quantum computers (also known as universal systems) can run any quantum algorithm.
Importantly, while gate-based machines will ultimately be able to solve more problems, annealing systems are currently easier to scale because they are less sensitive to errors. That informed D-Wave's decision to focus on annealing systems, allowing it to become the first company to commercialized quantum computers in 2011, and the first to offer cloud-based quantum services in 2018.
That first-mover advantage means D-Wave has customer relationships that could make it the go-to option as quantum computing technology matures. Furthermore, as the only company that builds annealers and gate-based machines, D-Wave has an edge over pure-play competitors because its ability to monetize niche quantum services in the near term should help fund R&D of universal systems in the long term.
CEO Alan Baratz has made comments to this effect: D-Wave systems can already solve problems beyond the scope of the most advanced supercomputers. While that is true, it is somewhat misleading. D-Wave systems will not be useful to most enterprises for years. In fact, Grand View Research estimates quantum computing sales will total just $4 billion in 2030, meaning the market (at that point) will still be about 100 times smaller than the AI market is today (i.e., AI sales totaled $390 billion in 2025).
Despite mainstream adoption being several years away, D-Wave stock currently trades at 347 times sales. That is absurdly expensive, even when analysts estimate the company's sales will increase at 70% annually through 2027. This stock is in bubble territory, and I think Ken Griffin added a very small position to capitalize on the momentum, not because he has any lasting conviction in the company. It is too early to predict long-term winners and losers with high confidence.
