GlobalFoundries: cash-rich and improving, but the price asks for 15% growth a year

GlobalFoundries ended 2025 with $1.16 billion of adjusted free cash flow, $4.0 billion of cash and securities, and, in 2026, its first dividend. This month it was named to operate two new New York fabs financed by Japan, and demand from data centers is surging. Yet the shares fell 9.1% on October 22 to $43.37, and they are 53% below their May high of $92.55. At this price the market is paying about 10 times this year's EBITDA for a business whose growth is mostly in a small part of its revenue.

At $43.37 on October 22, 2026, the shares trade 37% above my triangulated Base value of $31.59, a downside of 27.2% to Base. The Bull value of $53.68 is 23.8% above the price and the Bear value of $17.31 is 60.1% below it. The probability-weighted value of $33.54 is 22.7% below the price. The required margin of safety of 45% puts the maximum buy price at $17.37. The framework verdict is hold, do not add. Third-quarter results are expected on November 4, according to a data provider.

Valuation route: Technology hardware and semiconductor manufacturing, applied to a contract chip maker ("foundry") focused on mature and specialty processes. Economic classification: Standard Company (4/12), because the return on capital is well below 15% and free cash flow per share is falling in 2026. Two methods carry equal weight: mid-cycle EV to normalized EBITDA, built up from a sector base with adjustments, and a ten-year cash flow model. Reference checks: historical and peer multiples, and a reverse calculation of what the price assumes. Values are present fair-value estimates, not 12-month price targets.

What GlobalFoundries is

GlobalFoundries makes chips designed by other companies. It does not compete at the leading edge of manufacturing; it focuses on mature and specialty processes, including low-power silicon-on-insulator chips, silicon germanium for fast communications and silicon photonics, which moves data as light. It has plants in the United States (Malta, New York and Burlington, Vermont), Dresden in Germany and Singapore. In the second quarter, manufacturing services were about 89% of revenue and technology services, including processor intellectual property from the MIPS and ARC businesses, about 11%.

By end market in the second quarter, smart mobile devices were about 36% of revenue, home and industrial internet-of-things 19%, automotive 19% and communications infrastructure and data center 16%. The ten largest customers were about 63% of wafer shipment volume in 2025, down from 72% in 2023. Soitec supplied about 71% of the company's silicon-on-insulator wafer spend. The majority shareholder is Mubadala of Abu Dhabi, which sold 22 million shares in a block trade on May 27 and now holds roughly 73%.

The latest quarter

Communications infrastructure and data center revenue rose 62% on the year and 20% on the quarter. Automotive fell 10% on the year because of customer shipment timing, and smart mobile fell 6%. For the full year, call coverage says management now expects data center and communications infrastructure revenue up 50% to 60%, silicon photonics revenue more than doubled, automotive up by a low double digit percentage, home and industrial internet-of-things up 10% to 15% and smart mobile down by a low teens percentage. Price increases were put through in the second quarter, with the revenue effect starting in 2027.

Why the current earnings are not the whole story

The first reason is the gap between the two profit measures. In the second quarter, adjusted operating profit rose 16% on the year while IFRS operating profit fell 11%. The difference was $124 million, made of $86 million of share-based compensation, $31 million of amortization of acquired intangibles and $7 million of restructuring. Share-based compensation was 4.8% of revenue and about 15% of adjusted EBITDA, and it was $205 million in 2025. I treat it as a cost. In 2024 IFRS showed an operating loss of $214 million against an adjusted profit of $920 million; I did not verify the reasons for that gap.

The second is that cash flow turned in the second quarter. Adjusted free cash flow was $277 million a year earlier and negative $3 million this quarter, because capital spending rose from $159 million to $411 million. For 2025 free cash flow was $1.157 billion, 17% of revenue, on capital spending of $722 million, 10.6% of revenue. That is low for a business with $7.1 billion of plant and equipment and $1.3 billion a year of depreciation, and the second quarter suggests that spending is catching up. No full-year capital spending guidance was given in the sources I could read.

The third is accounting quality. The company's annual report for 2025 says that internal control over financial reporting was not effective at December 31, 2025, citing material weaknesses. Tax expense swings with currency revaluation of deferred taxes: $38 million of the first quarter's $81 million tax charge came from a German currency effect. Net income in the second quarter also included about $20 million of gains on equity investments.

Balance sheet, deals and shares

Cash and marketable securities are $3.30 billion, of which $946 million is non-current securities, against debt of $1.12 billion ($98 million due within a year) and non-current lease obligations of $495 million. Net of debt and leases the company holds about $1.7 billion. Interest cost is small, about $22 million a quarter, and the cash flow levels check was not triggered. Minority interests are $56 million. In the model I treat debt, leases and minority interests as claims, and hold back $250 million of cash as an operating reserve. Net financial claims are negative $1.38 billion, which adds to equity value.

On October 14 the company was named technology provider and operator of two new 300mm fabs at Malta, New York, with up to $21 billion of financing from Japan: about $10 billion for the first fab, due online in 2029, and about $10 billion for a second, to be decided by the end of 2028 depending on demand. Press coverage says the fabs would be jointly owned by the U.S. government and Japanese interests, and that GlobalFoundries would earn fees for at least 20 years, about $50 million next year and roughly $3 billion of adjusted free cash flow over ten years, according to reports of an investor call. Those fee figures are not in the company release. I treat the new fabs as outside the balance sheet and do not add the fees to the Base. A report that the deal adds 10 to 15 cents to 2027 earnings was also not in the release, and I ignore it.

In a separate agreement the U.S. Commerce Department will receive about 4% of the equity, taking the government's stake from 3% to 7%, according to the CEO as quoted in press coverage. I could not verify whether the shares are new or transferred, or at what price, so I treat them as new shares: about 23 million on top of 556 million diluted shares in the second quarter, 579 million in total. At that count the market value is $25.1 billion and enterprise value is about $23.7 billion, 10.1 times 2025 adjusted EBITDA of $2.357 billion and 9.8 times my estimate for 2026 of about $2.41 billion.

The company also reallocated $955 million of an existing CHIPS Act award toward silicon photonics, silicon germanium and packaging, and has a $300 million letter of intent from the Commerce Department for photonics and an expected $375 million grant for quantum manufacturing; neither is a final award. The annual report says the company plans to invest more than $16 billion over ten or more years at its New York and Vermont fabs, depending on government funding and market conditions.

Valuation: two methods, equal weight

The first method is mid-cycle EV to normalized EBITDA. The second is a ten-year cash flow model on cash flow after stock compensation, capital spending and working capital. Each runs through Bear, Base and Bull and each gets half the weight.

The Base normalized EBITDA of $2.35 billion is in line with 2025 ($2.357 billion) and 2% below my estimate for 2026. I do not assume that margins keep rising: a foundry with $7 billion of plants earns well in a good year and badly in a weak one, and the 2024 and 2025 EBITDA margins were 36.7% and 34.7% before share-based compensation.

Why 7.5x. The multiple is built up: a sector base of 7.0x for a capital-intensive mature-node foundry, plus 1.0x for growth in data center and silicon photonics, plus 1.0x for the balance sheet, plus 0.5x for supply agreements and long-term fee income, minus 1.0x for dilution (share-based compensation, the new government stake) and minus 1.0x for risk (a controlling shareholder holding about 73%, customer concentration, policy dependence and the internal control weakness). I add nothing for return on capital, which is below 10%. At 7.5x, the enterprise value of $17.6 billion on normalized free cash flow of about $1.2 billion is a 6.8% yield, and at a 9.5% discount rate that pays for perpetual growth of about 2.5%, the same as the Base assumes. That is a test the multiple passes.

The cash flow model starts with a second-half 2026 inflow of $150 million. Revenue grows 8%, 7%, 6%, 5%, 4.5%, 4%, 3.5%, 3.5%, 3% and 3% from $7.22 billion in 2026 to $11.5 billion in 2036. EBITDA after stock compensation is 29.5% of revenue and eases to 29%. Capital spending rises from 12% of revenue in 2027 to 14% from 2029, against depreciation of 17% falling to 14%: the 2025 level of 10.6% cannot last for a business with this asset base. Working capital takes 12 cents of each new revenue dollar. It uses a 9.5% discount rate, a 15% cash tax rate and 2.5% terminal growth. Terminal value is 51.2% of Base enterprise value.

Common basis and why the two methods differ

The gap is $1.43 billion of enterprise value, or $2.47 a share, and the divergence is 7.8%, well inside the framework's 20% limit. The two methods agree in the Base and Bull cases (1.9% apart in the Bull). In the Bear they are 25.1% apart, in the warning band, because capital spending stays at 14% of revenue while revenue falls.

Historical and peer cross-check

On trailing earnings the stock trades at about 33.9 times, and forward price to earnings is 20.0 times according to one data provider. On October 15, at $48.94, another provider put the trailing multiple at 38.2 times against a five-year median of 30.9 times; at today's price the trailing multiple is about 10% above that median. A PEG ratio is not meaningful because earnings growth depends on the adjustments, and free cash flow per share was $2.08 in 2025 and $0.41 in the first half of 2026.

UMC, the closest mature-node peer, trades at about half the multiple, with a larger net cash position and higher operating margin. The comparison is not like for like: GlobalFoundries reports adjusted EBITDA that adds back share-based compensation, and UMC's figure is a provider's. It does show that the market pays GlobalFoundries a premium for its U.S. and European footprint and for the data center growth. The shares fell 9.1% on October 22, the same day Coherent, Ciena and Fabrinet fell 11% to 12.5%; I found no company-specific cause, and one data service attributed the fall to a repricing of execution risk and capital intensity. Beta is 1.8.

Valuation triangulation

Divergence is 25.1% in the Bear, 7.8% in the Base and 1.9% in the Bull, so the Base and Bull are tightly supported and the Bear is a range of $15.14 to $19.48. Terminal value is 44.4%, 51.2% and 60.0% of enterprise value in the Bear, Base and Bull cash flow models.

What is the market already pricing in?

Run the cash flow model backwards and the price needs revenue to grow about 15% a year for four years from 2027 and fade after that, reaching roughly $14.3 billion in 2031, a compound rate of about 14.7% a year from 2026, against $9.7 billion in the Base (6.1% a year). Put differently, Base cash flows would have to be about 1.5 times higher every year. If the Base cash flows play out and the shares are bought at $43.37, the return on the enterprise is about 4.4% a year, short of the 9.5% required.

On multiples, at 7.5x the price implies mid-cycle EBITDA of about $3.2 billion, a third above the Base and above 2024, 2025 and my 2026 estimate. At 10x, the Bull multiple, it implies $2.37 billion, which is this year's level. The market is paying the Bull multiple for the current profit, or a normal multiple for a profit that has not happened yet.

Bear, base and bull scenarios

Bear: the handset and automotive recoveries fail, pricing gains do not stick, revenue falls 5% in 2027, EBITDA after stock compensation margins settle near 25%, capital spending stays at 14% of revenue and the market pays 5.5 times EBITDA. Base: data center and photonics growth offsets weak mobile, price increases lift 2027, and growth fades toward the pace of the economy. Bull: silicon photonics and communications infrastructure grow far beyond the guided rate, revenue grows 14% in 2027 and passes $13.7 billion by 2036, margins pass 32% and the market pays 10 times EBITDA.

The probability-weighted value is $33.54, 22.7% below the price.

Sensitivity: multiple method, value per share

Only one cell reaches the price: $2.7 billion of EBITDA at 9 times.

Sensitivity: cash flow model, value per share

Only the corner of a 7.5% discount rate and 3% terminal growth reaches the price.

How the buy price is set

Confidence C starts at a 35% margin of safety. Five points are added for the cyclicality of semiconductor demand and five for customer concentration, for 45%. A 45% discount to the $31.59 Base gives a maximum buy price of $17.37. Applied to each method separately, the same margin gives $16.70 for the cash flow model and $18.05 for the multiple method. Each risk is priced in one layer:

Dividend

The first quarterly dividend of $0.12 a share was paid on July 14, 2026, and another was declared for October 9. At $0.48 a year the yield is 1.1% and the cost is about $263 million. In 2025 free cash flow covered that more than four times, but in the first half of 2026 free cash flow was $230 million. The balance sheet can fund the dividend easily; it is not a reason to own the stock. The company bought back $400 million of shares in the first quarter, about $100 million remains under its authorization according to call coverage, and the new government stake will add to the share count.

Reasons to own GFS

• A strong balance sheet. $3.3 billion of cash and securities against $1.1 billion of debt, and a net cash position even after leases.
• Rising margins. Adjusted gross margin from 25.2% to 29.9% in a year, with 30.5% guided for the third quarter.
• Data center growth. Communications infrastructure and data center revenue up 62%, guided up 50% to 60% for the year, with silicon germanium capacity oversubscribed through 2027.
• Strategic position. One of few U.S. and European foundries, with a named role in a $21 billion Japan-funded project and government support.
• Pricing. Price increases across several technology corridors from the second quarter, with revenue effects from 2027.
• Cash generation in 2025. $1.16 billion of free cash flow, a 4.9% yield on enterprise value at today's price.

What could go wrong?

• Valuation. Only the Bull case clears the price, and the price needs about 15% revenue growth a year for four years.
• Capital spending. Quarterly spending rose from $159 million to $411 million in a year, and free cash flow was negative in the second quarter.
• Handsets and automotive. Smart mobile, 36% of revenue, is guided down by a low teens percentage and automotive was down 10% in the quarter.
• Controlling shareholder. Mubadala holds about 73% and sold 22 million shares in a block in May; further sales could weigh on the shares.
• Dilution. A new government stake of about 4%, share-based compensation at 4.8% of revenue and acquisitions paid in cash.
• Accounting. A material weakness in internal control at the end of 2025 and a wide gap between IFRS and adjusted profit.
• Policy and funding. The Malta project depends on Japanese financing and U.S. policy, and the company warns that funding could be delayed or withheld.
• Customer and supplier concentration. The top ten customers are 63% of volume and one supplier provides about 71% of silicon-on-insulator wafers.
• Missing data. Capital spending guidance, debt maturities and the largest customer's share were not in the sources I could read.

Management execution

Management has delivered on its guidance, lifted gross margin by about five points in a year, raised its 2026 growth outlook for data center and communications, completed acquisitions (MIPS, the Synopsys ARC processor business and Photeon's power team) to build a processor and design business, started a dividend and secured a large U.S.-Japan project. It also carries a weakness in financial reporting controls, reports adjusted profit that is far above IFRS profit, and has not given full-year capital spending guidance. I have not seen a call transcript, only call coverage. Management and capital allocation scores 6/10.

Stock Analyza scorecard

The valuation score is 25/100: the price is 37% above the Base and only the Bull case is above it.

Confidence score

The methods agree well, which lifts robustness. The grade is held back by reporting under IFRS with large adjustments, the internal control weakness and the missing capital spending and debt detail.

What I would watch from here

The next review point is third-quarter results, expected on November 4, 2026. Guidance is revenue of $1,885 million, plus or minus $25 million, adjusted gross margin of 30.5% and adjusted EPS of $0.51, plus or minus $0.05.

The three most important thesis breakers

1. Capital spending that keeps climbing while revenue growth stays in the single digits, ending the free cash flow that supports the balance sheet.
2. A weak handset and automotive cycle that outweighs data center growth, which is only 16% of revenue.
3. A loss of confidence in reporting, or a series of large share sales by the majority shareholder, while the price still discounts rapid growth.

Stress test and independent check

The whole model was recomputed separately in high-precision arithmetic and matched to the cent. Three judgments were tested because they move the answer. The share count matters: 556 million shares (no new government stake) gives a weighted Base of $32.90 and 600 million $30.48. Capital spending of two points of revenue in every year moves the cash flow value from $30.36 to $34.68 or $26.03, and margins of two points either way give $26.68 or $34.03. The second-half 2026 cash flow is not material: from negative $150 million to positive $400 million the value moves between $29.84 and $30.78. Treating leases as free instead of as debt adds about $0.85.

The strongest counter-argument is that the market is looking through 2026 to a company that will earn fees from two Japan-funded fabs, price its mature-node wafers higher, and sell silicon photonics into a long-lasting optical build-out, none of which is in my Base, and that a 9.5% discount rate is high for a company with net cash. At a 7.5% discount rate and the Bull operating path the cash flow value is $64.65, well above the price. That is a legitimate view, and it needs the Bull scenario and a low required return together.

Only the combination of the Bull operating path and a discount rate of 7.5% clears the price on the cash flow model. Main data limits: capital spending guidance for 2026 and later, debt maturities, current lease obligations, the largest customer's share, the terms of the government equity, the fee terms of the Malta project and the details of the internal control weakness were not available to me, and the third quarter has not been reported.

Final verdict: hold at $43.37, do not add

GlobalFoundries is in better financial shape than most chip manufacturers and has real momentum in data center and photonics. But the price already pays the Bull multiple for current earnings, and the two methods agree that the Base value is about $32. The cash on the balance sheet and the free cash flow of 2025 explain why the stock should not fall back to the Bear value, and the price explains why I would not add.

Verdict: HOLD at $43.37, do not add. Triangulated Base value $31.59, downside 27.2%, Confidence C (15/25). Required margin of safety 45%, maximum buy price $17.37.

For an existing holder, the downside to Base is at the edge of what the framework tolerates for a hold, and the Bull case is only 24% above the price, so the balance of risk is not favorable. New money waits for $17.37 or below, or for evidence that capital spending is under control, free cash flow stays positive and the reporting weakness is fixed.

The main thesis breaker is capital spending that keeps rising after the price increases and data center growth fade.

Sources

Primary sources include GlobalFoundries' second-quarter 2026 results release, the fourth-quarter and full-year 2025 results release, the first-quarter 2026 results release, the October 14 announcement of the Malta fabs and the fiscal 2025 annual report on Form 20-F. Earnings call coverage, press reports on the Malta deal and the government stake, Mubadala's share sale, peer and historical multiples and the market price come from third-party sources and data providers.

The reference market price is $43.37, the October 22, 2026 regular session close. An intraday price of about $44.2 on October 23 is within 5% of it. Balance sheet figures are as of June 30, 2026.