U.S. Memory Chip Boom: Why SK Hynix Could Build a New American NAND Factory

Educational research only — not investment advice.

Memory chip stocks are back in focus as AI demand pushes semiconductor companies to expand production closer to U.S. customers.

SK hynix subsidiary Solidigm is considering building a NAND flash-memory factory in the United States, with upstate New York emerging as a leading location.

No final investment decision has been made. But the proposal shows how AI demand, supply shortages and semiconductor policy are reshaping the global memory industry.

What Is Solidigm Planning?

Solidigm is SK hynix’s U.S.-based NAND business.

According to Reuters, the company is studying several options for a new American manufacturing site.

A U.S. factory would be separate from SK hynix’s ongoing discussions with Intel about potentially producing memory chips at Intel’s Ohio facilities.

That means SK hynix is exploring multiple ways to expand its U.S. manufacturing presence.

What Is NAND Memory?

NAND is the type of memory used to store data.

It is found in:

  • smartphones
  • laptops
  • solid-state drives
  • servers
  • data centers

DRAM, by comparison, provides the working memory processors use while performing tasks.

AI systems need both.

High-bandwidth memory, or HBM, receives most of the attention because it sits beside advanced AI processors.

But AI data centers also need enormous amounts of storage.

That is increasing demand for NAND products used in enterprise SSDs.

Why Is AI Creating More NAND Demand?

AI models generate and process huge amounts of data.

That information has to be stored somewhere.

As companies build more AI data centers, they need more:

GPUs + HBM + DRAM + NAND storage

Strong AI-server demand has contributed to a broader global memory shortage that industry executives expect could persist through at least 2027. Memory manufacturers have also prioritized investment in higher-value DRAM and HBM, limiting new NAND capacity.

That creates a simple supply problem:

AI demand rises → manufacturers prioritize advanced memory → NAND supply stays tight → memory prices strengthen

Why Build in the United States?

There are several strategic reasons.

Reduce dependence on China

Solidigm currently relies on its NAND manufacturing facility in Dalian, China.

A U.S. factory would diversify production and reduce dependence on a single manufacturing location.

Avoid trade and export risks

Semiconductors have become increasingly important in U.S.–China trade policy.

Producing NAND inside the United States could reduce exposure to tariffs and restrictions affecting semiconductor equipment or cross-border supply chains.

Move closer to AI customers

The United States is home to many of the world’s largest AI and cloud-computing companies.

Building closer to customers can create a more resilient supply chain.

SK hynix is already taking this approach with its more than $4 billion Indiana facility, which is expected to begin volume production of next-generation HBM4E products in 2029.

Why This Matters for Memory Chip Stocks

The memory industry is highly cyclical.

When supply becomes scarce, prices rise.

That can improve:

  • revenue
  • profit margins
  • factory utilization
  • cash flow

For companies such as SK hynix, Samsung and Micron, today’s AI-driven shortage can therefore be financially attractive.

SK hynix shares rose 6.4% on September 18, outperforming the broader Korean market, as investors digested the latest U.S. expansion reports.

But shortages also encourage companies to build more factories.

That creates the industry’s traditional risk:

shortage → higher prices → more investment → more supply → lower prices

Investors therefore need to watch both demand and future capacity.

China Is Expanding Too

The competition is not limited to South Korea and the United States.

Chinese memory producer CXMT is preparing to enter NAND flash memory, expanding beyond its traditional DRAM business.

That would put it into competition with Samsung, SK hynix, Micron and Chinese NAND leader YMTC.

China’s expansion matters because additional capacity could eventually reduce global shortages.

It also shows that memory chips are becoming increasingly strategic.

The industry is no longer driven only by normal consumer electronics cycles.

It is now influenced by:

AI investment + national industrial policy + supply-chain security

Why a U.S. Factory Is Not Guaranteed

There are still major obstacles.

Semiconductor manufacturing is expensive.

Reuters reports that SK hynix is concerned about the higher cost of producing chips in the United States compared with South Korea. The company also faces competing political pressure from Washington and Seoul over where future semiconductor investment should take place.

And semiconductor factories take years to build.

By the time a new NAND plant begins production, today’s shortage may look very different.

That makes long-term demand assumptions crucial.

What Should Investors Watch?

The most useful signals are NAND prices, AI data-center spending, memory shortages, new factory announcements, SK hynix capacity and Chinese semiconductor expansion.

The bigger story is straightforward:

AI is changing more than the GPU market.

It is increasing demand throughout the memory and storage supply chain.

If Solidigm moves ahead with a U.S. NAND factory, it would be another sign that semiconductor companies increasingly see American manufacturing as strategically important.

But for memory chip stocks, the long-term question remains the same:

Will AI demand grow faster than new memory supply?

As long as the answer remains yes, pricing power could remain strong.

If capacity eventually catches up, the memory cycle could turn again.

Track Semiconductor Trends With TradingSimuLab

TradingSimuLab’s Trend Detector and Macro tools help users study changing market trends, momentum and broader economic conditions across supported assets.

For more quantitative market research and educational trading tools, sign up to TradingSimuLab.

TradingSimuLab is for educational and research purposes only and does not provide investment advice.

Continue exploring TradingSimuLab.

  • Moving Average 10 Explained: What MA10 Shows in TrendAnalysis

    The 10-period moving average (MA10) is a short-term trend reference that smooths recent price action and helps show whether price is trading above, below, or repeatedly crossing its nearby trend. On a daily chart, MA10 usually represents the most recent 10 trading sessions. Its main purpose is simple: Is short-term price action holding above an…

  • Monte Carlo Simulation in Trading

    Monte Carlo simulation helps traders and investors study many possible market outcomes instead of relying on one forecast. Rather than asking: “Where will this asset be in the future?” Monte Carlo analysis asks: “Across many simulated paths, what range of returns, drawdowns and downside outcomes could occur?” Inside TradingSimuLab, Monte Carlo-style analysis powers Risk Simulation,…

  • Monte Carlo Simulation in Trading

    Monte Carlo simulation is a way to study many possible market paths instead of relying on one forecast. In trading and investment risk analysis, it can help answer questions such as: TradingSimuLab uses Monte Carlo-style path analysis inside Risk Simulation to provide context around expected return, probability of gain, simulated ranges, VaR, CVaR, maximum drawdown…

  • Max Drawdown Explained

    Maximum drawdown is one of the simplest ways to understand how painful an investment path can become. A portfolio can finish with a positive return and still experience a severe decline along the way. That is what maximum drawdown, often shortened to max drawdown or MDD, measures. It answers: What was the largest peak-to-trough decline…

  • Macro Scenario Payoff Table Explained

    TradingSimuLab’s Macro Scenario Payoff Table connects the broader macro outlook with the historical behavior of the selected asset. It answers three questions: How likely is each macro scenario? How did this asset historically perform after similar macro conditions? How much does each scenario contribute to Macro Expected Value? This is important because a weak macro…

  • Macro Net Score and Confidence Explained

    TradingSimuLab’s Macro Net Score and Model Confidence answer two different questions: Net Macro Score: Does the current macro backdrop lean constructive, defensive, or mixed? Model Confidence: How clear and internally consistent is that macro read? The distinction matters. A macro outlook can be positive but uncertain. It can also be negative with relatively high confidence…

  • Macro Model Workflow With Risk, Trend and Timing

    A macro outlook is useful, but it should not make the entire market decision. TradingSimuLab uses the Macro Model as the 12-month backdrop layer of a broader five-model research workflow. The process is designed to answer five different questions: The purpose is not to make five models produce the same answer. It is to identify…

  • Macro Model Explained: How to Read Net Score, 12-Month Outlook and Scenario Probabilities

    TradingSimuLab’s Macro Model is the long-horizon context layer of the five-model framework. It is designed to answer: Does the broader 12-month market backdrop look constructive, defensive, or mixed? Instead of relying on one economic indicator, the model combines broader macro and market context and summarizes the result through several outputs: The Macro Model is deliberately…

  • Macro Expected Value Explained

    Macro Expected Value, or Macro EV, is TradingSimuLab’s probability-weighted estimate of how an asset historically behaved across the Macro Model’s possible scenarios. In simple terms: Macro EV combines how likely each macro scenario appears with the asset’s historical payoff after similar model-defined conditions. It answers: If several macro outcomes remain possible, what does the probability-weighted…