Solidigm's IPO Could Create an AI-Storage Options Benchmark

The IPO headline is a valuation event before it is an options event#

Reuters reported on September 25 that SK Hynix's U.S. subsidiary Solidigm is considering a U.S. initial public offering as early as 2027 that could raise about $15 billion and value the business at as much as $150 billion. The company held meetings with investment banks competing for potential roles, but Reuters also emphasized that the plans remain early and that the size, valuation and timing could change.

That distinction matters for options research.

Solidigm is not yet a listed underlying, so there is no Solidigm option surface to inspect. The useful question is what the proposed listing could make observable in markets that already exist.

The direct parent is SK Hynix, whose U.S.-listed ADR, SKHY, now has exchange-listed options. The closest U.S. public storage comparison is Sandisk, whose business is heavily exposed to NAND flash and whose data-center segment has expanded rapidly with AI infrastructure demand. Micron adds a broader memory comparison, while CoreWeave provides a customer-side view because it has a multi-year agreement for priority access to Solidigm enterprise SSD capacity.

Those exposures answer different questions. The proposed IPO becomes interesting when it starts turning Solidigm from an internal SK Hynix asset into a separately valued AI-storage business that can be compared with public peers.

A $150 billion figure is not yet a comparable valuation#

The reported upper-end valuation is large enough to attract attention, but it is not yet a valuation multiple.

Reuters said Solidigm could be valued at up to $150 billion and could raise about $15 billion. The same report noted that the figure would exceed many recent semiconductor IPO valuations and would give investors a standalone price for a business that has historically sat inside SK Hynix.

What is missing is just as important.

There is no public IPO filing yet that establishes Solidigm's current revenue, operating income, cash flow, customer concentration, capital intensity or the financial contribution of its enterprise SSD business under the structure that would actually list.

Without those disclosures, comparing $150 billion directly with the equity value of Sandisk, Micron or SK Hynix would mix a preliminary transaction aspiration with fully disclosed public companies.

That means the first genuine valuation catalyst is not the Reuters headline. It is the disclosure that makes the proposed valuation measurable.

An eventual registration statement could reveal whether investors are being asked to value Solidigm mainly as a cyclical NAND producer, a higher-value enterprise SSD supplier, an AI infrastructure platform, or some combination of those businesses.

That classification could matter to listed options because each narrative implies a different set of public comparables.

Sandisk is the cleanest U.S. storage comparison, but not a neutral one#

Sandisk is more useful than a generic semiconductor ETF for one reason: its current business already shows how strongly AI data-center demand can change the economics of flash storage.

For fiscal 2026, Sandisk reported $20.25 billion of revenue. Its data-center revenue reached $5.15 billion, up 437% from the prior year, and in the fiscal fourth quarter alone data-center revenue was $2.98 billion. Management described data center as a key growth pillar.

That makes SNDK a more direct public-market reference for Solidigm than a broad chip index.

It is still not a clean valuation control.

Solidigm emphasizes enterprise SSD density, performance and energy efficiency for AI and cloud infrastructure. Sandisk has its own product mix, manufacturing partnerships, customer agreements and pricing cycle. An IPO filing would be needed before the two companies could be compared on consistent revenue, margin or cash-flow metrics.

The current SNDK options surface also shows that the public storage market is already carrying substantial uncertainty. On September 23, Options Skew Analytics measured SNDK 30-day at-the-money implied volatility at 73.34%. The source had only 96 usable historical sessions, so it did not assign a one-year percentile. Its 90-day volatility was only modestly above the 30-day reading, with a term-slope ratio of 1.029.

That is a useful pre-headline baseline, not evidence that investors had anticipated the Solidigm IPO.

SNDK had its own powerful company-specific drivers, including rapid data-center growth, NAND pricing changes and recent index-related activity. The correct future comparison is therefore not whether SNDK volatility is high in absolute terms. It is whether a verified Solidigm valuation milestone changes SNDK's volatility relative to its own prior curve and to broader semiconductor controls.

SKHY measures the parent-company problem, not the same thing as SNDK#

SKHY provides the most direct listed exposure because Solidigm remains an SK Hynix subsidiary.

The economics of an IPO could reach SK Hynix in several ways.

A listing could establish a visible market value for an asset currently embedded inside the parent. It could bring outside capital into the subsidiary. It could reduce SK Hynix's percentage ownership if new investors receive shares. It could provide funds for Solidigm expansion or allow the parent to recover part of its earlier investment.

Those channels do not all point in the same direction, and Reuters has not reported a final ownership structure or use of proceeds.

That is why SKHY options should not be treated as a simple referendum on a $150 billion valuation.

The timing problem is equally important. The reported IPO could occur in 2027, and no registration statement has been filed publicly. SKHY already has listed expirations extending well beyond the nearest weekly contract, but a far-dated option still contains every other risk affecting SK Hynix: HBM demand, DRAM and NAND pricing, capital spending, U.S. manufacturing, trade policy, earnings and the wider memory cycle.

The research task is to identify when Solidigm becomes large enough and specific enough to separate from those other variables.

Micron and CoreWeave help define what Solidigm actually represents#

Micron belongs in the research universe because it is a major U.S. memory company, but it is not the cleanest Solidigm proxy.

Its economics span DRAM, HBM and NAND, and its near-term options are heavily affected by its own earnings and AI-memory cycle. A Solidigm IPO could contribute information about enterprise SSD and NAND valuation without changing the variables that dominate Micron's HBM business.

That makes MU more useful as a memory-sector control than as the central article subject.

CoreWeave offers a different type of evidence.

In August, CoreWeave announced a multi-year strategic agreement giving it priority access to Solidigm enterprise SSD capacity for its AI cloud platform. The agreement shows that Solidigm is connected to a scaled AI infrastructure customer through storage rather than through GPU compute alone.

CRWV is therefore useful for validating the demand channel: AI infrastructure requires data storage to scale alongside compute. It is not a direct valuation comparable because a cloud platform and an SSD supplier capture different economics.

Together, these markets define three distinct layers:

SKHY measures the parent and capital-allocation relationship.

SNDK measures public-market valuation of a close storage business.

CRWV helps establish that enterprise storage is part of the AI infrastructure stack rather than a standalone NAND narrative.

A broad semiconductor ETF can then serve as a control for sector-wide volatility.

The IPO has several clocks, and options should not collapse them#

A possible 2027 listing is not one event.

The first clock is adviser selection. Reuters reported that Solidigm has already held bank pitch meetings. That shows process activity but does not determine transaction terms.

The second clock is formal filing. A confidential submission, if one occurs, would advance the process without necessarily making the financial statements public.

The third clock is public disclosure. A registration statement would expose the financial information required for real peer comparison.

The fourth clock is valuation discovery. An indicated price range, updated share count and ownership structure would turn the headline valuation into something closer to a measurable equity value.

The fifth clock is the listing itself.

Those stages can fall into different SKHY and SNDK option expirations. Treating a 2027 IPO as if it were a single September 2026 catalyst would create a horizon mismatch.

The more useful options study is whether volatility migrates toward expirations that actually contain new information.

The New York factory adds a capital-intensity test#

Reuters separately reported on September 18 that Solidigm is considering a U.S. NAND factory, with upstate New York among the leading candidates. No decision has been made, and the project could take years before producing memory.

That proposal changes the interpretation of the IPO.

A listing could be more than a valuation exercise if outside capital is eventually used to fund a more expensive U.S. manufacturing footprint. It could also expose public investors directly to the economics of moving NAND capacity away from dependence on Solidigm's existing Dalian manufacturing base.

The factory question therefore links IPO valuation to capital intensity.

A public filing that combines high AI-storage growth expectations with large future manufacturing requirements would give investors a much richer basis for comparing Solidigm with Sandisk and other memory companies. The relationship between growth, margins and capital spending would matter more than the headline enterprise value by itself.

For options, that disclosure could create a different volatility problem from the IPO date. A valuation headline is one event. A newly quantified capital program is another.

The clean test begins when financial disclosure arrives#

The strongest current finding is not that a $150 billion Solidigm value should alter SNDK or SKHY options immediately.

It is that the proposed listing could create a new public benchmark for AI storage at a time when the storage layer is becoming more economically visible.

The next useful observation is therefore specific.

When Solidigm releases enough financial information to compare its enterprise SSD business with Sandisk and other memory companies, preserve the pre-disclosure SKHY and SNDK volatility curves. Then compare the change in longer-dated implied volatility, term structure and skew with a semiconductor-sector control.

If SKHY changes more than the peer group, the parent-company ownership and capital-allocation channel deserves closer examination.

If SNDK changes more than SKHY while broad semiconductor volatility remains comparatively stable, the market may be treating the disclosure primarily as a new valuation benchmark for AI storage.

If the whole memory group moves together, the Solidigm event may be arriving alongside a wider NAND or AI-infrastructure repricing that cannot be attributed to the IPO alone.

The research opportunity is therefore not the $150 billion number itself. It is the moment when Solidigm's filing makes that number testable against public storage economics and gives listed options a new benchmark to react to.