After the optics boom: which profits become cash?

Selling more optical modules does not guarantee faster cash returns. Operating cash at China’s two leading module suppliers was less than a quarter of first-half attributable profit. Coherent still relied on external funding for expansion, while Lumentum’s results required adjustments for debt-related effects. Shared demand does not justify a common valuation.

Shared demand, different cash-flow profiles

High-speed connectivity demand has reached suppliers’ earnings, but the four companies offer shareholders different cash flows. Innolight and Eoptolink rely heavily on modules; Coherent spans components and systems; Lumentum’s net income also reflects capital-structure effects. Over the next 12–18 months, investment, collections and changes in share count can separate shareholder returns even under strong demand.

The clearest warning comes from cash flow. First-half operating cash flow at both Chinese module suppliers was less than one-quarter of net profit attributable to shareholders. Coherent's full-year operating cash flow also fell short of its capacity investment. Lumentum generated comparatively stronger operating cash flow, but a one-off debt transaction severely distorted reported net income. The income statement shows what products can earn; the cash flow statement shows how much funding growth requires upfront; changes in the share count determine how many shares participate in the resulting earnings.

Sources: Zhongji Innolight interim report, pp. 8 and 60; Eoptolink interim report, pp. 7 and 44; Coherent earnings release, Tables 2 and 4; Lumentum earnings release; and Lumentum 10-K cash flow statement. Chinese filings report in yuan, converted here to RMB billion; US release figures in USD million are divided by 1,000 to obtain USD billion. Cash capital expenditure for the Chinese companies includes purchases and construction of fixed assets, intangible assets and other long-term assets; the US figures cover cash purchases of property, plant and equipment. These expenditure measures are not free cash flow to the firm.

Half-year and full-year amounts cannot form a meaningful league table. The US companies' June fiscal years include the preceding calendar year's second half, whereas the Chinese interim reports cover only the current year's first half. During a rapid production ramp, that timing difference can materially affect product mix and margins. The comparison below therefore focuses on capital tied up in growth and how profits are earned. It does not double first-half results to produce a full-year forecast.

Modules and lasers earn margins differently

Module margins depend on mix and manufacturing execution. Innolight reported a first-half optical-transceiver margin of 46.59%, and Eoptolink an optical-interconnect margin of 48.46%. These product measures cannot be ranked directly against consolidated adjusted margins. High-speed mix, yield, materials and acceptance all affect the outcome.Innolight, pages 21 and 26; Eoptolink, pages 12–14

Module suppliers can benefit from upgrades in two ways: higher-specification products increase value per module, while manufacturing experience reduces the cost of producing a given specification. Customers, however, also capture cost savings through annual price negotiations. Even with revenue still growing, a larger contribution from lower-margin mature products, higher prices for scarce components or poor yields on new production lines can absorb the profit from incremental sales. Extrapolating net income from revenue growth assumes those pressures away.

Coherent generated FY2026 Datacenter & Communications revenue of USD 5.2746 billion and Industrial revenue of USD 1.8436 billion. Lumentum's full-year Components and Systems revenue was USD 2.0056 billion and USD 1.0084 billion, respectively. Neither is simply a pure-play laser company. Lumentum's fourth-quarter gross margin was 47.4% under GAAP and 50.4% on an adjusted basis, illustrating how accounting adjustments alone can change a cross-company ranking. Coherent segment revenue, Table 5; Lumentum product revenue and margin tables.

The more important question is where the profit accrues. If laser supply is constrained and customers lack qualified alternatives, component suppliers can raise prices. Module manufacturers bear the cost pressure if they cannot pass those increases through. Vertical integration can reduce dependence on external suppliers, but requires upfront investment in wafer, packaging and testing capacity. Broader product coverage is therefore both a competitive advantage and a capital commitment.

For the same reason, price-to-earnings multiples cannot be compared mechanically across component manufacturers, module suppliers and chip designers. The comparison must first establish which production stages a company owns and which it outsources, how much equipment and inventory new orders require, and whether research spending already covers the next product generation. An asset-light model does not eliminate capital costs: outsourcing embeds some of the supplier's required investment return in procurement prices. This analysis does not force additional companies into a valuation ranking that could mistake differences in business scope for a valuation discount.

China’s module leaders must collect the growth

First-half operating cash flow was 13.2% of attributable net income at Innolight and 21.5% at Eoptolink. After the long-term asset purchases shown in the table, the balances were negative RMB3.002 billion and positive RMB353 million. This tests cash coverage of those investments; it does not adjust minority claims, leases or investment transactions and is not distributable profit.

Fast-growing companies pay for materials and equipment before collecting from customers, so a short-term cash shortfall can coexist with genuine orders. The question is whether the funding requirement scales with revenue or becomes progressively heavier with each round of growth. When profits rise but cash lags, shareholders must wait longer or accept additional financing. For companies already earning high gross margins, better collections can matter more to valuation than another percentage point of gross margin.

Eoptolink's supplemental cash flow disclosure shows RMB 4.489 billion absorbed by inventory growth and RMB 5.754 billion by increases in operating receivables, partly offset by RMB 4.186 billion released through higher operating payables. Together, these three items absorbed approximately RMB 6.056 billion. “Operating receivables” is broader than trade receivables; the entire amount cannot be characterized as overdue customer payments. Eoptolink interim report, pp. 110–111; original amounts in yuan.

The implication is specific: Eoptolink is funding growth upfront, with supplier credit covering only part of the requirement. The constructive interpretation is that inventory accumulation supports delivery against new orders. The cautious interpretation is that some inventory may not keep pace with product transitions. Subsequent sales, collections and impairments will distinguish the two; management's demand commentary alone cannot. Even after products are sold, expanding payment terms can erode returns on capital.

Zhongji Innolight faces greater current cash spending pressure, but one half-year is insufficient to establish weaker long-term earnings quality. Capacity payments and customer settlements both depend on timing. A better comparison follows a rolling four-quarter period and tracks when capital spending begins to generate revenue. If new lines are slow to deliver products that meet requirements, spending initially justified as preparation for growth can become a low-return asset base.

For neither Chinese company has this analysis obtained a current end-customer revenue concentration table on a basis comparable with the US annual reports. Overseas revenue share is not a substitute for customer concentration. Modules may pass through equipment vendors or contract manufacturers before reaching a cloud provider, and one direct customer name need not represent one ultimate source of demand. Without that mapping, a longer customer list or wider geographical coverage does not establish lower risk.

Payment terms deserve their own valuation scrutiny. Consider a simplified stress test excluding tax: assume sales are spread evenly across the first half's 181 days and all customer payments are delayed by 20 days. Additional receivables funding would be approximately half-year revenue × 20 / 181. Zhongji Innolight and Eoptolink would need to advance an additional RMB 4.616 billion and RMB 2.310 billion, respectively—more than either reported in first-half net operating cash flow. Actual collection periods, seasonality and contractual terms differ, so this is not a forecast of receivables balances. It demonstrates that settlement delays can materially alter financing needs even with no bad debts. Strong customer credit and growth that needs no funding are entirely different propositions.

US suppliers also face financing claims

Coherent’s full-year operating cash flow less property and equipment spending was approximately negative $1.0234 billion. It also raised about $1.999 billion through common equity. The annual report principally attributed weaker operating cash flow to growth-related inventory investment. External equity helped finance expansion that operations had not yet fully funded.Coherent 10-K

Subtracting cash and short-term investments from the carrying value of period-end debt gives Coherent simplified net debt of USD 1.2352 billion: 0.0079 + 3.2143 − 1.1620 − 0.8250 = 1.2352, all in USD billion. This calculation does not offset restricted cash against debt and excludes operating leases and noncontrolling interests. It is not a complete enterprise-value adjustment. Coherent earnings release, Table 3; original USD million figures divided by 1,000 here.

On the same simplified basis, Lumentum held net cash of USD 1.1010 billion, but it also received nearly USD 2 billion from convertible preferred-stock issuance during the year. Its fourth-quarter debt equitization generated a USD 7.7566 billion noncash loss on debt extinguishment, the principal reason for the full-year GAAP loss. A negative price-to-earnings ratio cannot usefully assess core profitability here. Equally, removing the loss does not justify treating the capital structure as unchanged. Lumentum earnings release and reconciliation; 10-K financing cash flows.

Lumentum's full-year operating cash flow less cash equipment expenditure left USD 0.3001 billion, indicating stronger current internal funding coverage than Coherent. Growth is not free, however. Preferred-stock proceeds and cash generated by operations are different funding sources, and the former carries an equity claim. Adjusted earnings may also exclude stock-based compensation, but an economic assessment must still recognize its cost. Adding the expense back while holding the share count fixed ignores that burden.

Business expansion has not automatically reduced concentration. Coherent's two major customers accounted for 20% and 12% of FY2026 revenue; Lumentum's two major end customers accounted for 26.6% and 15.0%. The definitions differ, preventing a precise ranking of concentration risk. Even so, these disclosures challenge the assumption that broader product coverage necessarily produces diversification. Coherent 10-K, Major Customers; Lumentum 10-K, Customers.

Large customers provide scale but can also change acceptance requirements, supplier allocations and settlement arrangements. A demand slowdown can therefore put pressure on gross margins and working capital simultaneously. Valuation scenarios should reflect that correlation by reducing both earnings and multiples, rather than merely trimming a high revenue growth rate.

The source of any decline in net debt also matters. Repayment from operating cash strengthens the balance sheet for existing shareholders. A debt-for-equity conversion replaces a fixed claim with an equity claim, without necessarily delivering a comparable per-share benefit to those shareholders. It can reduce repayment risk while spreading future earnings across more shares. Once preferred stock has converted into common shares, its converted principal should not still be deducted from enterprise value while the enlarged share count is used. That double deduction understates equity value. The opposite error is to assume post-conversion interest savings while retaining the smaller pre-conversion share count.

Analyst forecasts depend on execution

CMBI’s Saiyi He, Kevin Zhang and Aaron Guo linked Coherent’s profit outlook to its indium-phosphide, module and systems capabilities in their June 12 report. Their FY2028 adjusted diluted EPS estimate of $15.03 and 31-times multiple supported an approximately $465 historical target, with capacity execution, adoption delays and pricing among the risks. June market inputs are not October valuation evidence or current consensus.CMBI, pages 1 and 16

On September 23, Stifel analyst Ruben Roy maintained a positive rating on Lumentum. According to Investing.com, he remained constructive on high-power laser supply and demand after discussions with management. StreetInsider's public summary confirms the analyst and rating action. This analysis did not obtain Stifel's full original report: the supply-demand assessment is treated as a media account, and management's description of demand is not treated as realized revenue. Investing.com report; StreetInsider summary.

These views capture the commercial logic that scarce components can claim a larger share of industry profit. The counterargument in this analysis is that sustained shortages do not translate automatically into durable high returns: expansion costs and customer bargaining power stand in between. If customers are willing to qualify second sources, today's profit pool will attract more supply. Should new capacity arrive together, supply recovery could depress prices before end-market demand slows. An indispensable product does not confer unlimited scope for margin expansion.

Coherent's cash shortfall is a tangible constraint on the optimistic forecast. Successful expansion can convert today's spending into future revenue, but a forecast must also account for inventory normalization, equipment utilization and financing costs. Raising distant revenue and margin estimates without revisiting those items leaves construction-period risks with shareholders while recognizing the benefits upfront in valuation.

Sensitivity of earnings and valuation

A one-point margin improvement can be tested independently of a volume forecast. Hold revenue, expenses and shares constant, assume tax rates of 15% for the Chinese companies and 20% for the US companies, and calculate revenue × 1% × (1 − tax rate). The resulting increases are roughly RMB355 million and RMB178 million for Innolight and Eoptolink’s half-years, and $56.95 million and $24.11 million for Coherent and Lumentum’s full years. Different periods, minority claims and actual tax treatment prevent ranking them or treating the exercise as an earnings forecast.

The same arithmetic works in reverse when gross margins fall. A module supplier facing higher component costs without higher selling prices, or a laser manufacturer incurring greater-than-expected initial depreciation on new lines, could experience that outcome. The thinner the net income base, the larger the effect of a small margin change on EPS. Where profitability is already high, the more important test is how long those margins can persist. Valuation disagreements often concern duration, not just next quarter's revenue.

Next, use CMBI's historical FY2028 forecast as an arithmetic reference and vary both EPS and the price-to-earnings multiple. The table does not use the current share price and does not represent this article's price target. The multiples and upward or downward adjustments are the author's assumptions, not evidence that the shares are undervalued.

Reference: CMBI, pp. 1 and 16. The difference between the middle value and the report's approximately USD 465 target reflects precision and rounding. The other cases are this article's hypothetical scenarios. No full operating forecast has been updated, and no discounted cash flow model has been constructed.

If earnings fall 20% and the multiple declines from 31× to 24×, modeled value falls approximately 38.1%, substantially more than earnings. Conversely, the upside scenario relies on two favorable assumptions when faster growth also earns a higher multiple. Investors should not treat those assumptions as independent sources of a margin of safety. Share count changes have a similar effect: if net income grows 20% and diluted shares increase 10%, EPS grows only about 9.1%, or 1.2 / 1.1 − 1.

Current prices, fully diluted share counts and forecasts for a common calendar year have not been verified on a consistent basis across the four companies. This article therefore offers no ranking of which stock is cheapest. Chinese profit excluding nonrecurring items is not automatically equivalent to US adjusted earnings; stock-based compensation, amortization, tax rates and preferred-equity claims must first be reconciled. Cross-currency analysis need not convert every company's earnings into one currency: comparable returns can be calculated in each company's local currency. Enterprise value and its associated cash flows, however, must never use different currencies.

Multiples can also be tested against cash available to shareholders. Suppose, purely illustratively, a company earns 100 and has an equity value of 3,000: its accounting PE is 30×. If only 40 in cash remains after the net reinvestment needed to stay competitive, equity value is actually 75 times shareholder cash flow. If greater net reinvestment leaves only 20, that ratio rises to 150×. Every number in this example is hypothetical, not a cash flow forecast for any of the four companies. It explains why identical earnings growth can command different valuations: rapid product upgrades require research, equipment and working capital, potentially leaving a substantial portion of profit unavailable for distribution over an extended period. Retained earnings deserve a value above their cash amount only when the incremental investment earns adequate returns.

The evidence needed to reduce the cash discount

The cash-quality assessment should change when the reports change. If Innolight and Eoptolink’s rolling twelve-month operating cash catches up with earnings while inventory growth slows and margins hold, the financing concern diminishes. Persistent extensions of collection periods, impairments and additional funding needs would instead require lower capital-return assumptions.

For Coherent, a decline in cash capital expenditure as a share of revenue after expansion, recovering operating cash flow and per-share earnings gains not materially diluted by new shares would strengthen the financial case for integration. Improving margins alongside persistently negative free cash flow would weaken the justification for a high forward multiple. For Lumentum, weaker component prices and margins, slower purchases by concentrated customers, or continued reliance on equity financing to improve net cash would require a reassessment of the valuation premium attributed to scarcity.

The distinction between Chinese and US optical suppliers cannot be reduced to “China manufactures; the US owns the technology.” Manufacturing yields, component availability, customer qualification and capital structure all influence who captures the returns. The more useful comparison now is how much funding each incremental unit of revenue requires—and how much cash remains per common share after financing costs and dilution.