CLS: Adjusted EPS Doubled. The Cash Bill Comes Next.

The 60-second read

Five things to know

1. The operating beat is real. Q3 revenue of $5.8297B and adjusted EPS of $3.30 exceeded management’s prior high ends.
2. The cash forecast moved the other way. FY2026 free-cash-flow outlook fell from $600M to $460M as working-capital investment grows.
3. Receivable financing deserves a separate check. The 10-Q reports $700.8M of sold receivables outstanding; gross turnover is not the net cash benefit.
4. New capital changes both sides of the model. The offering brought $3.3913B net cash while period-end shares rose to 126.1M.
5. The 2027 opportunity carries execution risk. A$40B revenue outlook and $1.5–$2.0B capex plan make cash returns and customer commitments central to valuation.

Three scenarios, one probability-weighted number

Analyst 12-month scenarios; all values are USD per share and probabilities are judgments.

Signposts: what would change our mind

Review dates are analytical checkpoints. Except the Oct 27 investor day, future company release dates remain unconfirmed.

The tape and the starting price

Celestica has delivered another large increase in earnings, but the price used here is deliberately frozen before those results reached the market. FMP's quote records $246.40 at 20:00 UTC on October 26, 2026. The SEC accepted the earnings filing at 20:23:49 UTC. The issuer had announced an after-close release; that does not establish an exact publication minute. We therefore make no claim that the recorded daily decline represents the reaction to this print. The conference call and investor day are scheduled for October 27, after this packet's research cutoff.

The historical-price endpoint separately reports a $246.65 daily close. We retain that raw close in the chart and use the explicit $246.40 timestamped quote consistently in the valuation. This small provider disagreement is disclosed rather than silently corrected. All dollar amounts are U.S. dollars despite the Canadian headquarters and dual listing. The float data are a dated provider snapshot, not a promise that every share could trade freely today.

The longer price record is useful because a large retreat can change the valuation without changing the operating business. It cannot establish a bottom. A stock that has fallen still carries exposure to a demanding ramp, customer concentration and a changing funding model. Our HOLD reflects that distinction: the earnings multiple looks reasonable, but the cash evidence is not yet strong enough for the margin of safety we require.

The event-return panel measures the filing-date close to the next available daily close for the eight completed historical windows. It is a descriptive event window, not a causal estimate of the earnings surprise. Other news and broad-market moves occur over the same interval. The new quarter remains pending because its next session has not happened. No zero return is invented for that missing observation.

The print: growth is real

Third-quarter revenue reached $5,829.7 million, an 82.5% increase over the same quarter last year and 24.1% above the preceding quarter. Adjusted diluted earnings per share reached $3.30, versus $1.58 a year earlier. GAAP diluted EPS rose to $3.26 from $2.31. The two growth rates tell different stories because last year's GAAP result contained large gains on total-return swaps. Treating the weaker GAAP growth rate as evidence that the core business stalled would miss that accounting comparison.

Results also exceeded management's prior ranges. Revenue finished above the $5.55 billion high end, while adjusted EPS exceeded the $3.08 high end. FMP's saved current-quarter expectations were $5.394 billion of revenue and $2.93 of adjusted EPS. Those are a dated external benchmark, not the company's own forecast, and older provider historical actuals differ from the filed financial statements. Every historical operating chart in this packet uses the SEC documents rather than replacing them with those provider actuals.

The outlook now points to $21.0 billion of full-year 2026 revenue and $11.55 of adjusted EPS. The previous figures were $20.5 billion and $11.30. Management left the annual adjusted operating-margin outlook at 8.4%. The important offset is that expected free cash flow fell from $600 million to $460 million. Management attributes that reduction primarily to working-capital investment supporting faster growth in 2027.

For 2027, the initial outlook is $40.0 billion of revenue, a 9.0% adjusted operating margin and $21.50 of adjusted EPS. These are forward-looking management estimates, not contracted earnings. The company explicitly says a forward GAAP reconciliation is unavailable without unreasonable effort. We use the outlook as an anchor for scenarios and valuation, while keeping the cost of delivering it visible.

Segments and the capacity challenge

Connectivity & Cloud Solutions produced $4,872.3 million of revenue and $433.2 million of segment income. Advanced Technology Solutions produced $957.4 million and $62.9 million, respectively. Both businesses grew, but the increasing weight of CCS dominates the consolidated trajectory. CCS accounts for roughly eighty-four percent of revenue and carries the largest exposure to hyperscaler and frontier AI infrastructure spending.

The company renamed Communications to Networking and Enterprise to Compute & Storage this quarter. This is a labeling change rather than a new segment boundary. Networking generated $3,263.2 million, while Compute & Storage generated $1,609.1 million. We map the old labels into those new names for historical charts, explicitly preserving the company's unchanged composition. Hardware Platform Solutions, an overlapping product capability rather than a third reportable segment, contributed approximately $2.3 billion of revenue. Adding HPS to CCS and ATS would double count revenue.

The strongest operational case is that Celestica is capturing more complex content within rapidly growing infrastructure programs. Engineering, design and platform capabilities can deepen customer relationships and improve the profit earned per revenue dollar. The limiting factor is that manufacturing scale still requires physical capacity, inventories, skilled employees and supplier coordination. Revenue growth is not equivalent to capital-light software growth.

Customer concentration has increased with the expansion. The top ten customers represented eighty-five percent of this quarter's sales; three customers individually represented thirty-five, seventeen and fourteen percent. Those percentages are disclosed without names, so we do not assign them to a particular hyperscaler. A successful relationship with a large buyer creates efficient scale, but a change in deployment timing can affect revenue, inventory and capacity utilization at the same time. Diversification inside a customer's product portfolio is useful; it does not remove dependence on that customer's spending decisions.

What reaches the shareholder

Adjusted operating earnings were $496.1 million. The path back to GAAP operating earnings is visible: employee stock compensation, swap valuation effects, acquired-intangible amortization and restructuring account for the $36.1 million difference. GAAP operating earnings were $460.0 million. Financing cost, interest income and a small miscellaneous credit then produce $455.8 million of pretax earnings. After $58.6 million of tax, net earnings were $397.2 million.

The prior-year comparison requires care. In Q3 2025, the total-return-swap fair-value adjustment was a $113.3 million gain, which supported GAAP profit and reduced the apparent difference from adjusted earnings. This quarter the adjustment was a $3.9 million loss. The adjusted operating result therefore gives a clearer view of core production performance across those two periods, but it remains a management-defined measure. Excluded compensation and acquisition amortization are not economically imaginary merely because a peer also excludes them.

The tax comparison also matters. The reported GAAP tax rate was thirteen percent and the adjusted rate eighteen percent, below the roughly twenty percent previously anticipated. Management attributed the adjusted benefit primarily to profit mix and a discrete item. Q4 guidance assumes about twenty percent, and the initial 2027 outlook assumes about twenty-one percent. We do not annualize the favorable quarterly rate indefinitely.

The completed share offering changes the denominator. Period-end shares were 126.1 million, compared with 115.0 million a year earlier. Weighted-average diluted shares in the quarter were lower, at 121.8 million, because the new shares were outstanding for only part of the period. For valuation we use the filed period-end count. Combining the full cash proceeds with the old weighted-average share count would overstate value per share. We keep the cash and equity issuance on the same economic footing.

Nine checks on earnings quality

The quality scorecard deliberately mixes favorable evidence and unresolved questions. Adjusted operating margin is expanding, and quarterly GAAP and adjusted EPS are close. Those observations are useful, but they do not cancel the cash-conversion problem. Free cash flow was $127.5 million against $402.1 million of adjusted net earnings. A rapidly growing manufacturer can rationally invest ahead of demand; the investor still has to decide how much cash can ultimately leave the business.

The most important additional disclosure is the receivables-sale program. The company sold $1,297.5 million of receivables during the quarter and $2,176.9 million year to date across the disclosed programs. At September 30, $700.8 million of receivables sold under the A/R program remained outstanding, compared with none at the previous year end. The program limit increased from $450 million to $1 billion in September. Proceeds are classified as operating cash flow under the company's accounting policy.

This is a liquidity tool, not proof of fabricated sales or customer distress. It also means cash collected through those sales can precede the ultimate customer's payment. The gross quarterly sales figure turns over and must not be subtracted mechanically from quarterly operating cash flow. The outstanding balance, program availability, costs and subsequent collections are the more useful monitoring points. Bank and customer programs are uncommitted and can terminate under stated terms.

Reported receivable days improved to fifty-six, while inventory days increased to seventy. Receivable sales help explain why falling reported A/R days need more scrutiny than the headline suggests. Contract assets included in receivables reached $417.2 million. Inventory write-downs were $18.5 million for the quarter. Neither fact is an automatic warning of failure; both belong alongside customer acceptance, product transitions and the scale of purchases. We do not calculate an unsupported forensic score from missing inputs.

The cash bill comes next

The balance sheet now holds $3,961.3 million of cash, against $818.8 million of borrowings and finance-lease obligations. Our simplified net-cash measure is therefore $3,142.5 million, or $24.92 per period-end share. This definition excludes operating-lease liabilities and does not imply that the entire balance is distributable. Manufacturing operations require liquidity in multiple jurisdictions, and expansion plans consume cash before new capacity produces mature returns.

Most of the increase in cash came from financing. The August public offering delivered $3,391.3 million after issuance costs. Operating cash flow was $389.7 million in the quarter, while net capital spending was $262.2 million. Gross capital purchases were $274.1 million, offset by $11.9 million of asset-sale proceeds. That distinction explains the company's $127.5 million free-cash-flow figure and avoids silently mixing gross and net capex definitions.

Management's 2027 capital-expenditure plan of $1.5 billion to $2.0 billion is much larger than historical spending. The company describes demand forecasts for 2028 as meaningfully above the 2027 revenue outlook, but it has not converted that qualitative statement into a verified 2028 revenue forecast in this release. We do not manufacture one. Facilities and equipment are an investment in that opportunity, with utilization and return on invested capital still to be demonstrated.

The annual cash-flow table shows how quickly the funding scale has changed. Historical free cash generation improved, while stock-compensation settlements and repurchases also used cash. A proposed new repurchase authorization is not the same as a commitment to retire shares. For the next stage, capital allocation should be judged on whether new programs generate sustainable cash after their initial ramp, rather than on the gross cash balance or the size of a buyback headline.

Three valuation routes and one margin of safety

Our base fair value is $300 per share, with a wide $180–$420 scenario range. The current $246.40 quote is about eighteen percent below that base value. We retain HOLD with conviction three out of five because our required margin of safety for this high-uncertainty ramp is twenty-five percent. A price near $225 would meet that price test only if the operating thesis and balance-sheet evidence remained intact. The threshold is an analytical decision rule, not a guarantee of a profitable entry.

Route one applies fifteen times management's initial 2027 adjusted EPS outlook of $21.50, giving $322.50. The multiple is our assumption. It gives credit for strong growth while recognizing customer concentration and capital requirements. Route two values $40 billion of revenue at the guided nine percent adjusted operating margin, producing $3.6 billion of adjusted EBIT. Eleven times that result, plus $3,142.5 million of net cash, divided by 126.1 million shares produces about $338.96. This route depends on the full revenue and margin outlook being achieved.

Route three tests cash rather than adjusted earnings. We model annual free cash flow of $800 million, $1.3 billion, $1.8 billion, $2.3 billion and $2.7 billion for 2027 through 2031, discount at 11.5%, and use 3.5% terminal growth. Those are our assumptions, not management guidance. Adding current net cash produces about $234.17 per share. The terminal value contributes most of the enterprise valuation, so this is a sensitive model, not a precise answer.

The simple mean of the three routes is approximately $298.54, rounded to $300. Their dispersion is itself a finding: an earnings-based argument looks more attractive than a conservative cash-based one. At the current quote the market pays about eleven and a half times the 2027 adjusted EPS outlook and roughly $27.93 billion for the operating enterprise under our net-cash definition. The price need not assume failure to be reasonable; it can reflect uncertainty about turning the forecast into cash.

Wall Street: dated, and mostly before this print

The saved FMP target consensus is $456.23, with a range from $375 to $520. These marks were retrieved on October 26, but the underlying dated target headlines precede the new earnings release. A fresh retrieval timestamp does not make an analyst opinion post-results. Our $300 base value is below that aggregate, primarily because we put greater weight on cash generation and the possibility that capital spending and working capital absorb more of the earnings increase.

The most recent saved individual target is BMO Capital's $400 on October 23. Bernstein's $520 is dated September 30, FBN Securities' $375 September 15, and UBS's $430 August 24. These are provider-recorded targets, not independent verification of the complete underlying research reports. We report firm, number and date so the viewer can judge staleness, without implying access to proprietary notes or inventing their rationale.

The ratings distribution in the snapshot is nineteen Buy and nine Hold, with no Sell ratings shown. That is a useful picture of published sentiment, but it is not a probability that the shares will rise. Recommendation categories vary by firm, targets can be updated at different times, and an aggregate can mix analyses prepared under different operating assumptions.

FMP's FY2027 consensus sample shows roughly $36.0 billion of revenue and $19.74 of EPS, versus management's new $40.0 billion and $21.50 outlook. The sample sizes differ: thirteen revenue analysts and seven EPS analysts. The gap is a reason to watch subsequent revisions, not proof that a new surprise has already been priced. We do not compare a future management forecast with stale consensus and label the entire difference an immediate earnings beat.

Management: execution and allocation

Management earns credit for delivering revenue and adjusted EPS above the high ends of its prior quarterly ranges. It also raised the annual operating outlook while providing a much larger initial 2027 revenue forecast. The practical question is whether that visibility supports investments with attractive returns after the ramp, rather than merely investments that make the next revenue milestone possible.

The public offering strengthened liquidity and reduced financing pressure, while diluting existing shareholders. Raising capital is neither automatically good nor automatically bad. Its value depends on what the new capital earns and whether management can maintain discipline as demand forecasts rise. The quarter-end cash balance should not be treated as evidence that the underlying business generated the offering proceeds.

The contemplated normal-course issuer bid could permit repurchases of up to 2.5% of public float after TSX acceptance. It is a proposed authorization and remains discretionary. Comparing that percentage directly with the total-share dilution from the offering would mix different denominators. We will assess actual purchases, their prices, their purpose and their effect on shares outstanding rather than assuming an authorization reverses the issuance.

Our management scorecard therefore has two time horizons. Near-term execution is strong: program delivery, adjusted profitability and results versus guidance improved. Longer-term allocation remains open: receivable financing, capacity investment, the durability of customer commitments and cash returns require subsequent evidence. The investor day should clarify the shape and timing of expansion, but we do not prewrite its conclusions. Management's current forecasts are useful inputs whose accuracy must be tested against future filings.

1. What portion of the 2027 capacity plan is tied to firm customer commitments rather than changeable demand forecasts?
2. How much of the $1.5–$2.0B capex is growth capacity versus maintenance and site transfer?
3. How should investors interpret receivable sales and outstanding sold A/R as volumes expand?
4. What cash conversion should a mature 2027 program earn after inventory and customer-deposit needs?
5. Which milestones would cause management to defer expansion or redirect capital?

Ranked risks and the opposing cases

The largest risk is a mismatch between demand timing and committed capacity. A customer can remain optimistic about AI infrastructure while changing the sequence of individual deployments. If Celestica has already added facilities, equipment and inventories, a delay can reduce utilization and consume liquidity even without a permanent loss of the customer. That is why our valuation range is wide rather than a narrow confidence interval around the guidance midpoint.

Concentration makes the scenario more consequential. Three customers account for approximately two-thirds of current revenue, and the largest accounts for more than one-third. The identity of those customers is not disclosed in this filing. We do not fill that gap with speculation. A change in product architecture, sourcing strategy, program allocation or component availability can affect the economics of a relationship before it becomes visible in annual revenue.

The strongest bull case is that the company is entering a period of unusually strong operational leverage. Revenue growth can spread engineering and administrative costs, product complexity can support the value of Celestica's capabilities, and the new equity capital can fund profitable expansion without excessive leverage. Under that case, the 2027 outlook is achievable and the following year's earnings base is larger still. The current quote would then look inexpensive on earnings.

The strongest bear case does not require AI spending to disappear. It requires only lower cash conversion, more capital tied up in the operating cycle, or a slower utilization ramp than the earnings multiple implies. Receivable funding can provide a bridge, but it does not turn working-capital investment into permanent free cash. Our HOLD is a judgment between those cases: the operating evidence is strong, while the funding and cash-return evidence is still developing.

Dated checkpoints, not vague promises

The first near-term event is the October 27 earnings call and investor day, scheduled for 8:00 a.m. Eastern. At this research cutoff they have not occurred. We will look for capacity commitments, the relationship between 2027 capex and customer forecasts, and a clearer explanation of working-capital funding. We will not treat a more enthusiastic description of the demand environment as equivalent to a change in a numerical outlook.

The next quarterly reporting checkpoint is the Q4 2026 release. Its exact publication date has not been established in the saved source set. The monitoring dates below are therefore analytical review dates, not asserted company event dates. Q4 revenue guidance is $6.275–$6.575 billion and adjusted EPS guidance $3.37–$3.57, with an 8.7% adjusted-margin midpoint. Full-year cash generation will show whether the reduced $460 million outlook was sufficient to absorb the year-end investment cycle.

The 2027 investment program needs separate milestones from the revenue target. Investors should monitor whether facilities and equipment become operational on time, whether inventories turn at expected rates, and whether sold receivables remain a bridge or become a growing permanent funding requirement. A lower reported cash cycle is constructive only when its drivers are understood.

Our decision framework is falsifiable. A buy case needs either a substantially larger price cushion or stronger cash evidence while the earnings outlook stays intact. A more defensive call would follow deteriorating customer commitments, an outlook cut, or sustained cash demands that exceed the assumptions in the model. A stock-price decline alone does not prove the business is impaired, and a price rally alone does not verify the forecast. The next packet should grade each numerical signpost against these recorded starting conditions.

Data, definitions and reproducibility

The dataset contains nine quarters of income statements, GAAP-to-adjusted reconciliations, balance sheets and cash-flow statements, plus three audited fiscal years of cash flows and structural information. Current results come from the October 26 earnings exhibit and the same-day September-quarter 10-Q. Annual data and employee counts come from the 2025 10-K. Historical quarter data are read from the actual release tables, with each source URL retained in the accompanying dataset.

The history crosses a reporting-basis change. Income comparisons in our series use U.S. GAAP values from later domestic filings where available. The September 2024 balance sheet is retained from its original IFRS filing and labeled accordingly. In particular, its lease-inclusive debt figure should not be compared mechanically with the later borrowings-and-finance-lease measure. We do not smooth away that limitation or build a same-basis net-debt trend from incompatible observations.

Derived values are computed in the saved preparation scripts: growth, margins, cash conversion, net cash, enterprise value, per-share values, event windows, valuation routes and scenario weighting. Figures are rounded for readability after calculation. A company's own free-cash-flow definition uses capital purchases net of specified asset proceeds; our annual reconstruction shows that arithmetic. Reported operating cash flow includes proceeds from derecognized receivables under the disclosed programs. This packet does not relabel the company's GAAP cash-flow statement as invalid.

Adjusted EPS and adjusted operating margin exclude items defined in the release. GAAP means U.S. generally accepted accounting principles for the current period. CCS is Connectivity & Cloud Solutions; ATS is Advanced Technology Solutions; HPS is Hardware Platform Solutions and overlaps the reportable segments. A total-return swap can create accounting gains or losses linked to share-price movements. Net cash here means cash less borrowings and finance leases; it excludes operating leases. Fair value, scenario probabilities and required margin of safety are our judgments rather than company guidance or a promise about future market prices.

Historical coverage is retained. The prior local Q2 2026 title emphasized the earnings beat, raised guidance and CCS growth with a HOLD call. This quarter's analytical emphasis is the cash cost of funding the much larger capacity plan. A local prior title is not evidence of a separately verified public long-form URL, and no such URL is invented. All sources and model outputs remain dated so later research can evaluate what was known when this call was made.

Market sources: Financial Modeling Prep quote, profile, float, historical daily prices, earnings expectations, analyst estimates, grades and price-target-news endpoints; all retrieved October 26,2026. Saved original provider responses and source manifests support the dated tables. No forecast is represented as an actual result.