Sandisk: cheap on next year's profit, expensive on a normal one

Sandisk earned $11.4 billion in the fiscal year that ended July 3, 2026, and the shares trade at 6.9 times consensus profit for the year after. In the four fiscal years before that, the company added up to a net loss of $3.4 billion. The first number makes the stock look cheap. The second is why a cheap-looking multiple on peak profit is not enough.

At the October 12, 2026 close the shares trade 54.7% above my triangulated Base value of $962, a downside of 35.4%. The Bear value of $464 is 68.8% below the price and the Bull value of $1,664 is 11.9% above it. Confidence is grade D (10/25), which carries a required margin of safety (MOS) of 65%. That is above the framework's 60% ceiling, so I issue no maximum buy price and no buy zone. The stock fell 6.0% on October 12 on Chinese memory competition worries and is 36.8% below its 52-week high of $2,354.39, yet it is still 12.3 times its 52-week low of $120.96. The market has already paid for most of the good news. The question is how long the good news lasts.

Valuation route: Semiconductors and ecosystem, subtype MEMORY (NAND flash, fabless-style product company with a joint venture manufacturing partner). Economic classification: Standard (4/12). Primary method: mid-cycle EV/EBITDA on the average of fiscal 2027 to 2031. Secondary method: five-year free cash flow model with a terminal value. Cross-check: peak-year EBITDA multiple. Values are present fair value estimates, not 12 month price targets. All figures are in US dollars, and fiscal years end in late June or early July.

What Sandisk is

Sandisk makes NAND flash, the memory that stores data in phones, laptops, memory cards and data center solid-state drives (SSDs). It was part of Western Digital until the spin-off in February 2025. It does not run its own wafer factories. Its flash comes from Flash Ventures, three joint ventures with Kioxia that operate eight fabs in Japan. Sandisk holds 49.9% of each, takes about half the output at cost plus a small markup, and pays half of the ventures' fixed costs whether or not it takes the wafers. The ventures run until the end of 2034, after the two older ones were extended in January 2026.

Fiscal fourth-quarter revenue was $8,965 million, of which Edge (phones, PCs and client SSDs) was $5,432 million, Datacenter $2,977 million and Consumer $556 million. Datacenter was only $213 million a year earlier. Management says about one-third of the quarter's growth came from higher volume and two-thirds from higher prices. No customer was more than 10% of revenue in fiscal 2026, but the top ten were 44%.

The latest quarter

Two details matter. First, the guide implies pricing has stopped accelerating: revenue is expected to grow 15% to 20% with a flat gross margin and "modest price increases", after a quarter in which price did most of the work. Second, GAAP net income for the year includes $808 million of gains on investments (the cash flow statement), so non-GAAP EPS of $39.25 is the cleaner read of the quarter. The shares fell after the August 5 report even though the quarter beat expectations, and the call summary I read points to investors asking whether margins have peaked.

Why one year of earnings cannot be used

For a memory company the framework asks for mid-cycle economics, not peak or trough earnings. The table shows why. The figures for fiscal 2022 to 2025 are Sandisk's carve-out history from when it was part of Western Digital.

Free cash flow in fiscal 2022 to 2025 added up to -$786 million. Fiscal 2026 alone produced $11,494 million. Over the five years net income averaged $1.6 billion a year and free cash flow $2.1 billion a year. In the last downturn, fiscal 2023, revenue fell 38% to about $6.1 billion, average selling prices per gigabyte fell 39% and gross margin dropped from 33.3% to 7.1%. A business with that record cannot be valued by capitalizing its best year. I value it on the average of a five-year path that starts at the peak and ends at a settled level, and I test that path against a cash flow model.

Cash flow quality passes with a warning. Free cash flow was 101% of net income, but it includes $2.75 billion that the company itself strips out of adjusted free cash flow (customer prepayments under the new contracts), a $1.37 billion increase in income taxes owed that will be paid later, and a $3.64 billion rise in receivables that offsets part of it. Receivables were $4.7 billion at year end, about 48 days of fourth-quarter revenue. Stock-based compensation (SBC, pay in shares) was $232 million, 2% of free cash flow, and sits inside the GAAP operating income I use.

Balance sheet and shares

There is no financial debt. Cash was $4,762 million at July 3, and buybacks since then have used a large part of it: the share count fell from about 149 million to 146,419,001 by August 7 (10-K cover page). Most equipment spending appears to sit inside Flash Ventures, not on Sandisk's own capital spending line. Management put gross capital spending at $562 million in the fourth quarter (6.3% of revenue) and expects about 6% of revenue in fiscal 2027. Sandisk pays for it through wafer prices and by guaranteeing 50% of the ventures' lease obligations, and it must fund its half of new capital spending if the ventures' own cash flow is short. It also owes Kioxia $1.2 billion of collaboration payments over 2026 to 2029, and it holds an equity stake in Nanya Technology (bought in March 2026, reported at NT$31 billion).

The company's own first-quarter guide uses about 155 million diluted shares, so the estimate is consistent. Basic shares rose only 0.4% from 145.78 million at June 27, 2025, but diluted shares rose 8% (145 million to 157 million) largely because the treasury method adds more shares at a higher price. StockAnalysis shows a market value of $217.8 billion on basic shares. On economic shares it is $229.6 billion.

I could not confirm what sits in the $3.4 billion of other liabilities, which probably includes taxes owed, the Kioxia obligation and some customer deposits, so I count all of it as debt-like. Excluding it adds about $22 per share to the Base. Operating lease liabilities of $201 million are excluded because rent is already in operating costs. Working capital for the cash flow model starts at $5.5 billion (receivables plus inventory less payables and accruals), which is 15.4% of annualized fourth-quarter revenue.

Do the new contracts change the cycle?

This is the strongest argument for the bulls, and it deserves a straight answer. Sandisk has signed what it calls New Business Model (NBM) agreements with eight customers, five of them since April. They run up to five years with a weighted average above four years. Management says they carry at least $93.9 billion of revenue at floor pricing, and that remaining performance obligations were $59.8 billion at quarter end and $91.1 billion including two agreements signed after it. They are expected to cover more than half of fiscal 2027 bits and about two-thirds of fiscal 2028 bits. Management also says it expects NBM margins around 80%, with "attractive margins even at floor pricing".

Spread over four years, $93.9 billion is about $23 billion of revenue a year before any upside, and it is the reason my Base never lets revenue fall below that level before fiscal 2031. Three things limit how much protection it gives. The 10-K says the agreements carry stated volume commitments from both sides and financial guarantees that offset part of lost revenue if a customer fails to perform. The guarantees total $16.5 billion, which is 17.6% of the floor revenue, mostly provided by third-party financial institutions. In the parts of the 10-K I read, I did not find take-or-pay terms or the price-floor mechanism. And the floors were set at today's prices, so they protect revenue only for as long as customers are willing and able to buy. The contracts also tie a large share of bits to a few buyers, who are exposed to the same AI spending cycle as Sandisk.

Primary valuation: mid-cycle EV/EBITDA

Enterprise value (EV) is market value plus debt minus cash. For a cyclical business the framework asks for a multiple of mid-cycle EBITDA (earnings before interest, taxes, depreciation and amortization), not of the peak. I define mid-cycle as the average EBITDA of a five-year path, fiscal 2027 to 2031, that starts at today's peak, falls as new supply and normal price declines arrive, and settles at a level above history because of mix and the contracts. I apply 7x to that average in all three cases, add the net bridge of -$1.8 billion, and divide by 154.4 million shares. Holding the multiple fixed means the cases differ only in their operating paths, so the same risk is not counted twice. The 7x is my assumption for a cyclical manufacturer at mid-cycle earnings, and the grid below tests it.

The Base path is anchored on four facts: the first-quarter guide ($10.55 billion at the midpoint, 78% operating margin); fiscal 2027 consensus revenue of $49.1 billion (my Base is 2% below it, built from the guide plus sequential growth of 9%, 9% and 8%); the $23 billion annual contract floor; and the fiscal 2023 precedent of a 39% price drop.

Fiscal 2027 operating income of $37.4 billion gives GAAP earnings of about $207 a share, below the consensus non-GAAP figure of $214.82. Revenue stays above the contract floor until fiscal 2031, when contracts begin to roll off. The gross margin path is consistent with the contracts: if floor-priced bits ($23 billion a year at an assumed 80% margin) are two-thirds of fiscal 2029 revenue and the rest is sold at a 30% margin, the blend is about 64%, within four points of the Base. The settled fiscal 2031 margin of 47% is above the 33% of fiscal 2022 because of the higher share of data center drives and the contract structure. That premium is my estimate and is tested by the robustness table.

Each value is average EBITDA times 7, plus the -$1.8 billion bridge, divided by 154.4 million shares. The primary Base is $978. The Bear case revenue of $43.0 billion in fiscal 2027 is close to the low end of the analyst range ($43.4 billion), because the first quarter is guided and the year is largely set. The Bear then repeats the fiscal 2023 pattern, with contracts renegotiated lower and gross margin back near its 2022 level of 33%. The Bull has contracts fully honored, supply short through 2028 and a fiscal 2031 gross margin of 62%.

Sensitivity: value per share

The price of $1,487.33 needs about $33 billion of average EBITDA at 7x, or about $26 billion at 9x. Neither is close to the Base.

Secondary valuation: five-year free cash flow model

The secondary method discounts the cash flows of the same five years plus a terminal value. It uses tax of 15% in fiscal 2027 and 2028 (the company's guide) and 17% after, depreciation of $150 million, Sandisk's own capital spending at 0.9% of revenue (fiscal 2026 was 0.9%), working capital at 15.4% of revenue and SBC inside operating costs. The discount rate is 11% and terminal growth is 2% in all three cases, applied to fiscal 2031 after-tax operating income. The valuation date is October 12, 2026. A falling revenue path releases working capital, which helps the cash flows in the later years.

The secondary Base is $772. In the Base, about 42% of the enterprise value is cash that arrives in the next two years, which makes the method less dependent on the terminal assumption than usual and very sensitive to how fast prices fall. The Bull value rests on the terminal value for 54% of its total.

Sensitivity of the Base model: value per share

Peak-multiple cross-check and growth diagnostics

No reliable history of through-cycle multiples exists for Sandisk (it has traded alone only since February 2025), and I did not build a peer table of Micron, SK hynix, Samsung and Kioxia. That is a stated limitation. In place of a historical multiple I use the convention that cyclical memory companies trade at a low multiple of peak-year EBITDA, because the market expects the profit to fall. I apply my own assumptions to fiscal 2027 EBITDA: 3.5x in the Bear, 5.0x in the Base and 6.5x in the Bull. That gives $726, $1,205 and $1,728. This is a convention check on the same operating assumptions, not independent evidence, and the multiples differ by case because the multiple is meant to carry the expected decline after the peak. Today the stock trades at 6.2x my Base fiscal 2027 EBITDA, between my Base and Bull multiples.

The growth valuation cross-check is a diagnostic only. The forward P/E is 6.9x on consensus fiscal 2027 non-GAAP EPS of $214.82, which is 203% above fiscal 2026 non-GAAP EPS of $70.88, so the PEG ratio is about 0.03. That supports nothing. Fiscal 2025 non-GAAP EPS was $2.99, the growth comes from a price spike, and the GAAP trailing P/E of 20.2 includes investment gains. Growth durability: cyclical rebound. Cyclical distortion: high. Flags: PEG_NOT_MEANINGFUL, CYCLICAL_REBOUND, ACCOUNTING_BASIS_MISMATCH. A low PEG during a memory pricing recovery cannot support fair value by itself, and This check does not set fair value here.

The free cash flow growth quality check says a growth rate is not meaningful, because free cash flow was -$120 million in fiscal 2025 and $11,494 million in fiscal 2026 (absolute change +$11.6 billion). Free cash flow per share was about $74 on 155 million diluted shares, or $56 on the company's adjusted figure, with a free cash flow margin of 56.8% and cash conversion of 101%. The current free cash flow yield is 5.0% on reported and 3.8% on adjusted free cash flow. The forward yield looks high, 12.9% on my Base fiscal 2027 free cash flow and 10.9% on fiscal 2028, but both years are the peak of the path and fall to 3.1% by fiscal 2031. Basic shares are up 0.4% in 13 months. Consensus fiscal 2027 EPS ranges from $186.97 to $238.36 across 21 analysts, which I rate medium reliability, and the revision trend was not available to me. Bridge of the growth: higher prices (two-thirds of the fourth-quarter increase), higher volume, an 84.6% gross margin, $2.75 billion of prepayments and deferred taxes, offset by $3.6 billion of receivables. Stress test: cutting operating income in fiscal 2029 to 2031 by 20% gives a triangulated Base of $897 and by 40% gives $832. Distortion flags: low base, cyclical rebound, working capital, capital spending timing (equipment sits in the joint ventures). Classification: cyclical rebound, diagnostic only.

Valuation triangulation

The primary and secondary Base values differ by 23.5% (measured against their midpoint), a pass with a warning (the line is 20% to 30%). The gap comes from timing. The primary method pays a multiple on the whole path, including its decline, while the cash flow model discounts each year. The peak-multiple check and the cash flow model differ by 43.8% (against their midpoint), which is why I give the check only 20%. I kept the 50/30/20 weights I use elsewhere on this site. The unrounded values are $464.30, $961.61 and $1,663.95.

What is the market already pricing in?

At $1,487.33 and 154.4 million economic shares, market value is about $229.6 billion. After the -$1.8 billion bridge, enterprise value is about $231.4 billion, which is 18.4 times trailing EBITDA of $12.5 billion and 6.2 times my Base fiscal 2027 EBITDA. At 7x the price requires average EBITDA of about $33.1 billion a year over the cycle, 52% above my Base of $21.8 billion. The Sandisk of the last five years averaged net income of $1.6 billion.

The cash flow view is stricter. My Base cash flows are worth the current price only at a discount rate of about 5.3%. At 11%, the price needs my Base operating income path to be 1.9 times larger in every year, an average of $41.6 billion against my $21.7 billion. Put another way, at an 11% discount rate and 2% growth, the price is the present value of about $20.8 billion of free cash flow every year forever (operating income of $25.1 billion), which is 70% of my fiscal 2027 free cash flow. A buyer today needs five things to be true: contract prices keep rising through the next quarters, the NBM floors hold in practice and are renewed at similar economics after fiscal 2030, new supply from China and from the incumbents does not close the shortage, AI storage demand grows faster than supply for years, and gross margin settles far above any level Sandisk has seen in a normal year.

Bear, base and bull scenarios

I assign no probabilities. Fiscal 2027 EPS is close to consensus in every case, which is the point: next year is largely locked in, and the cases are about what comes after. The Bull is already generous, with five years of at least $24 billion of operating income and a gross margin that never falls below 62%. The price sits 11.9% below it.

How the buy price is set

A Confidence D score carries a base margin of safety of 50% or more, and I start at 50%. I add 10 points for operating leverage and cyclicality (gross margin went from 84.6% to the 7% seen in fiscal 2023 in a single downturn) and 5 points for concentration (a single manufacturing partner, Kioxia, and top-ten customers at 44% of revenue). The valuation divergence modifier does not apply, since the primary and secondary methods differ by 23.5%. No positive modifier applies, because a net cash balance sheet is not enough when predictability is low. The required MOS is 65%. Above 60%, the framework prefers a NO-BUY or WATCHLIST verdict to a pseudo-precise bargain price. No maximum buy price and no buy zone are issued. For orientation only, a 60% discount to the Base would give $384.64. That is not the real 65% requirement and must not be used as an automatic order or price alert.

Dividend and capital return

Sandisk pays no dividend, and management says buybacks are its preferred way to return cash. It repurchased $5.2 billion of shares in fiscal 2026, $4.5 billion of it in the fourth quarter at about $1,595 a share, and the remaining authorization of $15.5 billion is 6.7% of the current market value. Buybacks add value only when they are done below intrinsic value. At my Base of $962, shares bought at $1,595 are a transfer from the cash the remaining holders own to those who sold.

Reasons to own SNDK

• Contracted demand. Remaining performance obligations of $91.1 billion including recent signings, eight NBM customers, more than half of fiscal 2027 bits under multi-year agreements and a weighted average term above four years.
• Margins and pricing. The first-quarter guide is $10.3 billion to $10.8 billion of revenue at an 83% to 85% gross margin, and a third-party forecast quoted on October 12 expects NAND contract prices to rise another 15% to 20% in October to December.
• A clean balance sheet. No financial debt, $4.8 billion of cash at year end and $15.5 billion of buyback capacity.
• Data center mix. Datacenter revenue was $2,977 million in the fourth quarter, up 103% from the third, and management expects data center to be about half of the NAND market in 2026, up from about 30% in 2025.
• Little capital of its own. Cash capital spending was $177 million for the year, and the manufacturing venture runs to 2034.
• The multiple on next year's numbers. 6.9 times consensus fiscal 2027 EPS, and about 8.3 times the annualized first-quarter guide of $45.

What could go wrong?

• Thesis breaker: the price cycle turns. NAND is a commodity. In the last downturn revenue fell 38%, gross margin fell to 7% and the company lost $2.1 billion. On October 12 the stock fell 6.0% on reports of progress at China's CXMT and softer flash pricing, and a quarter in which volume has to offset lower prices would change the story.
• Contract enforceability. Guarantees cover 17.6% of floor revenue, I did not find take-or-pay terms in the 10-K excerpts I read, and the floors were set at today's prices. If prices fall far below the floors, renegotiation is more likely than payment.
• Partner and structure. Sandisk pays half of the Flash Ventures fixed costs whatever it takes, guarantees 50% of their lease obligations, owes Kioxia $1.2 billion through 2029 and buys flash in yen.
• Quality of cash. $2.75 billion of fiscal 2026 free cash flow came from prepayments and other items the company excludes from its adjusted figure. Receivables rose to $4.7 billion and inventory rose 20.6% in a quarter in which revenue rose 51%, to about 178 days of cost of sales from 158.
• Capital allocation at the top. $4.5 billion was spent on buybacks in one quarter at about $1,595, and $14 billion more was authorized in August.
• Customer concentration and demand. The top ten customers were 44% of fiscal 2026 revenue, and demand from AI data centers is itself cyclical. Management's own forecast of a NAND market near $500 billion in calendar 2027, after about $300 billion in 2026, is a demand projection that depends on prices staying high.

Management execution

Management has delivered what it guided since the spin-off: the fiscal fourth quarter beat expectations, fiscal 2026 debt of $1.9 billion was repaid, adjusted free cash flow is disclosed next to the reported figure, and the company signed multi-year contracts that change how the business sells. On the other side, the record is only about six quarters long, so the framework's three-year minimum is not met, and the fiscal 2026 buybacks were made at prices well above my value. Basic shares were flat. Assessment: 11/20, good but unproven. The next test is whether the first and second fiscal 2027 quarters hold the 83% to 85% gross margin as price increases fade.

Stock Analyza scorecard

Revenue grew about 20% a year from fiscal 2022 to 2026 (four years, with fiscal 2022 derived from the reported 38% decline), which earns the growth point even though it is measured into a peak. Without it the score is 2 and the class is still Standard. Return on invested capital averaged 21% over five years but only because of fiscal 2026 (104%), after -10%, -4% and 5% in the three years before it. Unit economics score zero because a gross margin that ranged from 7% to 85% is not stable. Recurring economics get one point because contracts cover about half of fiscal 2027 bits. Character change: positive but unproven. A spot-price commodity supplier is moving to multi-year contracts, which should narrow the range of outcomes, but the first contract year has not been tested by a price decline, and the framework lets a positive change lift the verdict only after evidence. Business quality is 52/100: moat and pricing power 9/20, return versus cost of capital 8/20, balance sheet 10/15, earnings and cash quality 8/15, growth and runway 9/15, management and capital allocation 5/10, dilution and governance 3/5. The valuation score is 13/100: discount to Base 2/40, protection against the Bear case 1/20, agreement between methods 5/15, historical valuation 1/10 (price-to-book is 14.6 against 0.7 a year ago), reverse valuation 2/10, data and model quality 2/5.

Confidence score

The company's release is current, and I read the 10-K cover page, the joint venture description and the risk factors on SEC.gov. The financial statement notes (remaining performance obligations, taxes, awards, the other liabilities) came from an aggregator and a call summary. Predictability is low because the business is a commodity cycle. Robustness is low because the three methods range from $772 to $1,205 for the Base, and the Bull is 3.6 times the Bear.

What I would watch from here

Today's readings: green on non-GAAP gross margin (84.6%), on contract prices (a third-party forecast of 15% to 20% increases in October to December), on the NBM backlog ($91.1 billion) and on economic shares (about 154.4 million). Yellow on inventory (about 178 days, up from 158) and on industry supply (the CXMT news). Red on price: $1,487.33 is 1.55 times my Base. The next review point is the fiscal first-quarter report on October 29, 2026. Consensus for it is about $10.64 billion of revenue and EPS of $46.22 (Benzinga).

The three most important thesis breakers

1. The second-quarter outlook shows revenue flat or down and gross margin below 80%, which means contract prices have rolled over and the Bear path starts early.
2. Any customer shortfall, renegotiation or guarantee claim under the NBM agreements, which would show the floors do not hold when prices fall.
3. New supply, from Chinese producers or the incumbents, large enough to end the shortage in calendar 2027.

The opposite result would show my Base is too low. A second-quarter outlook above $12 billion at an 84% gross margin, NBM floors raised or extended beyond fiscal 2030, and price increases that continue through the first half of 2027 would move the value toward my Bull case.

Stress test and independent check

I recomputed the bridge, the three primary values, the three cash flow models, the peak-multiple check, the triangulation, the sensitivity tables, the yields and the reverse valuation with a second, separately written calculation, and I checked the reported inputs against the fiscal fourth-quarter release, the call summary, the 10-K and the market price. The results matched to the dollar. As a check on the operating assumptions, the Q1 guide of $45 EPS on 155 million shares equals $7.0 billion of net income, and my first-quarter operating income of $8.3 billion after 15% tax gives $7.0 billion. The remaining weaknesses are in the inputs, not the arithmetic.

Every check on the Base lands below the price, and the most generous reasonable combination I could build gives $1,329, 10.6% below it. That is why the verdict is a reduce and not a hold. The price is not out of reach, though: it sits 11.9% below my Bull case of $1,664, and the Bull case is a coherent outcome.

The strongest counter-thesis, for the bulls: demand exceeds supply, 91 billion dollars of contracts are signed, the margin is 84% and the first-quarter guide implies a $42 billion run rate before further price increases. Management says it aims to avoid the 2023 boom and bust, supply stays on allocation beyond calendar 2027 and demand grows faster than supply. If that is right, the Bull case is the base case. The response is that this is possible and it is in my Bull, but the price already pays 6.2 times Base fiscal 2027 EBITDA, the contracts are only partly guaranteed, the Bear case is 69% below the price, and the stock needs about $33 billion of average EBITDA a year, against a five-year average net income of $1.6 billion.

Audit flags: I read the 10-K cover page, joint venture description and risk factors on SEC.gov, but the financial statement notes came from an aggregator, so remaining performance obligations, the composition of other liabilities, tax payable and award counts rest on the call summary and estimates. I read the earnings call through a summary page, not the full transcript. The cash spent on post-quarter buybacks is my estimate. Consensus for fiscal 2028 and later is behind a paywall, so no outer-year consensus is used. The multiples (7x, and 3.5x to 6.5x for the cross-check), the settled gross margin, the price path after fiscal 2027 and the 15.4% working capital ratio are my assumptions. The fiscal 2022 revenue figure is derived from a reported decline, and fiscal 2022 to 2025 are carve-out figures. I did not build a peer table, and I did not value the Nanya stake beyond its carrying value. Treat the result as conditional, because several critical inputs are estimated and the methods differ by 23.5% to 43.8%.

Final verdict: reduce at $1,487.33

Sandisk has contracts, a record margin and a balance sheet with no debt. It also has a history in which its profit went from $1.1 billion to a $2.1 billion loss in a single year and back to $11.4 billion. The shares at $1,487.33 pay for roughly 70% of fiscal 2027's cash flow to repeat every year forever. My Base assumes the contracts mostly hold, and it still comes out 35% below the price.

Verdict: REDUCE at $1,487.33. Triangulated Base value $962, downside 35.4%, Confidence D (10/25). The required margin of safety of 65% exceeds the framework's practical ceiling, so no maximum buy price is issued.

The main risk to this verdict is that the shortage lasts longer than I assume, the contracts are renewed at similar economics and gross margin stays near 80%, which would put the value near my Bull case of $1,664. New money belongs on the watchlist. The part of this story that has been confirmed is the profit. The part that has not been confirmed is how much of it is permanent.

Sources

Primary sources include Sandisk's fiscal fourth-quarter 2026 earnings release (August 5, 2026), the Form 10-K for the year ended July 3, 2026 and the fiscal fourth-quarter earnings call transcript. The October 12, 2026 close of $1,487.33, the 52-week range, the multi-year financial statements and consensus estimates come from StockAnalysis (forecast page) and are labeled as estimates where they are third-party figures. Market context comes from Benzinga and The Motley Fool (fiscal 2023 downturn figures and the TrendForce price forecast). The Nanya investment is described in TrendForce.

The reference market price is the October 12, 2026 regular session close at 4:00 PM EDT. Balance sheet figures are as of July 3, 2026, and the common share count is as of August 7, 2026. The economic share count, the post-quarter buyback cash, the scenario paths, the multiples, the working capital ratio and the robustness tests are analytical estimates, not company guidance.