SK Hynix’s consideration of a potential public listing for Solidigm highlights how the artificial intelligence boom is creating value far beyond the processors that receive most of the industry’s attention. Solidigm, the US-based NAND flash memory and enterprise solid-state drive business owned by SK Hynix, is considering an initial public offering that could value the unit at as much as 150 billion US dollars. The company has already held meetings with investment banks over a potential offering that could take place as early as next year.

The reported valuation is striking because Solidigm operates in storage rather than the high-profile processor segment dominated by companies supplying artificial intelligence accelerators. But the rapid expansion of AI data centres is creating enormous demand for high-capacity storage capable of handling increasingly large datasets and workloads.

The proposed listing would therefore offer the market a direct way to value a storage business benefiting from the same infrastructure expansion that has driven demand for advanced computing.

AI Is Changing the Economics of Storage

Artificial intelligence systems require enormous quantities of data, and the infrastructure supporting them needs storage at several stages. Data centres must retain training datasets, model information, intermediate results and increasing volumes of information generated during inference. As AI workloads expand, storage becomes a more important part of the overall data-centre architecture.

Industry data show how quickly that shift is occurring. Enterprise solid-state drives accounted for 48% of global NAND shipments in the second quarter of 2026, nearly twice their share a year earlier, with AI inference workloads identified as a major driver. SK Hynix ranked second among suppliers, with a 22% shipment share, helped by strong growth in Solidigm’s shipments.

The change matters because enterprise storage traditionally served a wide range of corporate computing requirements. AI is now becoming a major force in determining where NAND capacity is allocated, with data-centre customers willing to pay for higher-capacity and higher-performance storage.

That gives Solidigm a potentially attractive position at a time when the economics of NAND memory are improving.

Solidigm Is More Than a Conventional Storage Company

Solidigm was created after SK Hynix acquired Intel’s NAND and solid-state drive business for approximately 9 billion US dollars in a transaction announced in 2020. The business was subsequently reorganised under the Solidigm name, creating a US-based storage company focused heavily on enterprise applications.

Its products are used in servers, cloud infrastructure and data centres, making it directly exposed to the capital spending cycle of large technology companies.

That exposure has become more valuable as hyperscalers build enormous AI computing systems. Storage requirements rise as the amount of data processed by those systems expands, and enterprise customers increasingly need high-capacity solid-state drives capable of delivering fast access to large datasets.

Solidigm’s focus on high-capacity enterprise storage therefore places it closer to the core of the AI infrastructure cycle than its consumer-facing profile might suggest.

The IPO Could Unlock a Standalone Valuation

For SK Hynix, a public listing could serve several purposes. It could raise substantial capital for Solidigm, provide an independent market valuation and potentially give the subsidiary greater flexibility to invest in capacity and technology.

The reported potential valuation of up to 150 billion US dollars would make the transaction one of the largest semiconductor listings if it goes ahead. Sources cited in the reporting said Solidigm could seek to raise about 15 billion US dollars, although the size and timing remain preliminary and could change with market conditions.

The distinction between a potential valuation and an eventual market value is important. Semiconductor valuations can change sharply with memory prices, supply conditions and expectations about technology demand. A company benefiting from a strong AI cycle can command a high valuation when investors expect demand to remain strong, but the same business can face pressure when memory supply expands faster than demand.

The IPO would therefore create both an opportunity and a market test for Solidigm.

The growth of enterprise storage is also changing the competitive structure of the NAND market. Samsung remains a major supplier, while SK Hynix has expanded its position through both its own NAND operations and Solidigm. Other competitors, including Kioxia, Micron and Chinese memory producers, are also investing in technologies aimed at data-centre demand.

Enterprise SSD revenue among the leading suppliers reached nearly 37.6 billion US dollars in the second quarter of 2026, with SK Hynix Group generating more than 8.6 billion US dollars. Solidigm’s high-capacity quad-level cell enterprise products have been an important part of that growth.

The shift toward data-centre storage can improve profitability because enterprise customers often value capacity, reliability and performance more than the lowest possible price. That can create better economics than commodity-oriented consumer storage.

However, the market remains cyclical. Memory manufacturers have historically experienced periods of oversupply followed by shortages, creating substantial fluctuations in prices and margins.

SK Hynix Is Expanding Across the AI Supply Chain

Solidigm’s possible listing also needs to be viewed alongside SK Hynix’s broader expansion in AI-related memory. The parent company is investing heavily in high-bandwidth memory, including a planned 4 billion US dollar facility in Indiana that is expected to begin advanced memory production later this decade. SK Hynix expects strong memory demand to continue through 2030.

The company is therefore building exposure across different parts of the AI memory ecosystem. High-bandwidth memory serves the computing side of AI systems, while enterprise solid-state drives address storage requirements.

A separate listing for Solidigm could make that distinction more visible to investors. Instead of valuing Solidigm simply as one component of SK Hynix’s wider business, the market would have an opportunity to assess the storage unit according to its own growth prospects.

The real question is whether AI demand is durable.

The potential IPO comes at a moment when investors are placing exceptionally high valuations on companies linked to artificial intelligence. That creates favourable conditions for companies with credible exposure to the sector, but it also increases the importance of demonstrating that growth is supported by actual customer demand.

Solidigm’s strongest argument is that enterprise storage demand is already increasing as AI workloads move from training into inference. Storage is required regardless of which particular AI model becomes dominant, because data has to be stored, accessed and moved through computing systems.

Yet the company remains exposed to the normal risks of the semiconductor cycle. New capacity, changing technology and fluctuations in data-centre investment could affect future margins.

For SK Hynix, the potential listing therefore represents more than a fundraising exercise. It could establish Solidigm as a separately valued AI infrastructure company at a time when storage is becoming one of the less visible but increasingly important components of the artificial intelligence economy.

The proposed IPO is still at an early stage, and its final structure, timing and valuation are not settled. But the fact that a storage business can potentially command such a large valuation illustrates how the economics of artificial intelligence are spreading beyond processors into the infrastructure required to store the enormous quantities of information on which AI systems depend.

(Adapted from TradingView.com)