Compare Memory Stocks by Valuation, Capacity & Demand
This guide shows you how to compare memory chip stocks (DRAM, NAND, and related semiconductor companies) using three key lenses: valuation, production capacity, and customer demand. You’ll learn which metrics to pull, how to weigh them, and how to avoid common traps. Before you start, have access to a financial data source (e.g., your broker’s research tools, Yahoo Finance, or company filings) and the latest quarterly earnings reports for the companies you’re comparing.
What You’ll Need
- Access to financial data (Yahoo Finance, MarketWatch, or your brokerage research section)
- Latest 10-Q or 10-K filings for each company (available on SEC EDGAR or company IR pages)
- A spreadsheet or notepad to track metrics
- Basic understanding of P/E, EV/EBITDA, and revenue growth
Step-by-Step
- List the memory stocks you want to compare. Focus on pure-play memory makers (e.g., SK Hynix, Micron, Kioxia) and maybe one or two diversified names (Samsung, Intel) if you want a broader view. Keep the list to 3–5 names so the comparison stays manageable.
- Pull valuation multiples for each stock. For each ticker, note the trailing P/E (price-to-earnings) and forward P/E (based on next year’s consensus EPS). Also grab EV/EBITDA (enterprise value divided by earnings before interest, taxes, depreciation, and amortization). Write these in a table.
- Normalize for the memory cycle. Memory prices are cyclical—they boom and bust. A low P/E might mean the stock is cheap, or it might mean earnings are about to crash. Check where we are in the cycle: are memory prices rising (shortage) or falling (oversupply)? If prices are peaking, a low P/E can be a warning, not a bargain. Adjust your interpretation accordingly.
- Estimate production capacity. Capacity is about how many wafers a company can process per month. Look for “wafer starts per month� or “capacity in wafer equivalents� in the company’s quarterly report or investor presentation. If not disclosed, use revenue per employee as a rough proxy, but that’s less precise. Write down each company’s capacity in millions of wafer starts per month.
- Calculate capacity per dollar of market cap. Divide market cap by monthly wafer capacity. Example: If Micron has a $100B market cap and 1.5M wafer starts/month, that’s $66,667 per wafer start. Lower is better—you’re getting more capacity per dollar. Do this for all companies and rank them.
- Assess customer demand. Look at end-market demand drivers: data centers, PCs, smartphones, automotive, and industrial. Check recent earnings calls for commentary on “bit growth� or “demand visibility.� Also look at days of inventory (DIO) from the balance sheet—if DIO is rising, demand may be softening. If DIO is falling, customers are buying faster than the company can ship.
- Compare revenue growth and gross margin. Pull the latest quarter’s year-over-year revenue growth and gross margin for each company. Memory companies with strong pricing power show high gross margins (above 40% in good times). Compare these to the valuation multiples you gathered in step 2.
- Rank on a composite score. Create a simple score: give each stock a score of 1–5 for valuation attractiveness (lower P/E = higher score), 1–5 for capacity efficiency (lower market cap per wafer = higher score), and 1–5 for demand momentum (higher revenue growth and lower DIO = higher score). Sum the scores for each stock. The highest total is your top pick based on this framework.
- Double-check with qualitative factors. Look at technology leadership (e.g., who’s shipping the latest DDR5 or 3D NAND layers), customer concentration (one big customer is risky), and capital expenditure plans. A company with huge capacity but outdated tech may not be as attractive.
Common Mistakes
- Using trailing P/E in a cyclical downturn. When memory prices crash, earnings collapse, making the P/E look artificially high. Conversely, at the peak, earnings are inflated, making the P/E look cheap. Always use forward P/E and check cycle position.
- Ignoring debt. Two companies can have the same P/E, but one has twice the debt. Use EV/EBITDA instead of P/E to account for debt.
- Mixing capacity definitions. Some companies report capacity in “bit shipments� (total data stored), others in wafer starts. Convert everything to the same unit—wafer starts per month is the most common.
- Overweighting capacity alone. A company with huge capacity but weak demand will sit on inventory. Capacity only matters if customers are buying.
Tips
- Use a 5-year view of memory prices (DRAM and NAND average selling prices) to understand where you are in the cycle. If prices have been rising for 6+ quarters, be cautious.
- Check the “days of inventory� trend over the last 4 quarters. A consistent decline usually signals strong demand.
- Look at the “bit growth� guidance—memory companies often provide expected growth in total bits shipped. Higher bit growth with stable pricing is a good sign.
- For capacity, use the most recent quarter’s data, as memory companies add capacity quickly.
- Cross-check your findings with at least one analyst report (free summaries are fine) to see if you’re missing a major factor like a new fab coming online.
Troubleshooting
What if you can’t find capacity data? Some companies don’t disclose wafer starts. In that case, use “capital expenditure per revenue dollar� as a proxy—higher capex intensity often means more capacity being built. Or use total revenue from memory products as a rough size indicator.
What if the valuation looks absurdly low? Check if the company has one-time charges or gains that skewed earnings. Look at “non-GAAP� earnings, which exclude those items, and compare that P/E instead.
What if two stocks tie on your composite score? Break the tie by looking at free cash flow yield (FCF divided by market cap). Higher FCF yield means the company is generating more cash relative to its price, which is a good tiebreaker.
What if demand data is contradictory? For example, revenue is up but DIO is also up. That could mean the company is building inventory ahead of a new product launch. Read the earnings call transcript to see management’s explanation before making a judgment.
Frequently Asked Questions
Q: Why is P/E not enough for memory stocks? A: Memory is cyclical. Earnings swing wildly with memory prices, so a P/E based on current earnings can mislead. Use forward P/E and EV/EBITDA to get a clearer picture.
Q: What is “capacity� exactly? A: It’s the maximum amount of memory chips a company can produce in a given period, usually measured in wafer starts per month or bit shipments. It shows how much supply the company can bring to market.
Q: How do I know if demand is strong? A: Look at revenue growth, days of inventory (falling = strong demand), and commentary from management about order backlogs or customer commitments. Also monitor end-market trends like data center spending.
Q: Should I compare memory stocks to other tech stocks? A: Only if you’re building a broader portfolio. For a memory-specific comparison, stick to memory names because their cycles are similar.
Q: What is a good market cap per wafer start? A: There’s no universal number, but you can compare across companies. If one is significantly lower, it suggests you’re getting more capacity for your money. Historical averages for major memory makers range from $50,000 to $100,000 per wafer start, but this varies with cycle.
Q: How often should I update this comparison? A: Quarterly, after each earnings report. Memory fundamentals shift fast, so a comparison from six months ago is likely outdated.
Related
Method: the standard formula for this calculation, rounded as each result states. See how the calculators work or report an error.