CELESTICA INC. (CLS): what the price assumes

In the published model solve dated 2026-Q2, anchored at $299.55, CELESTICA INC. (CLS) is priced for today's economics sustained for ~5.6 years. boothcheck publishes no house fair value, target price, or buy/sell rating; individual model outputs and user-controlled scenarios are analytical inputs, not Boothcheck targets. Narrative composed 2026-07-25.

Generated: 2026-08-30 · Exported: 2026-08-31 · Source: https://boothcheck.com/report/CLS

Headline

What The Price Assumes (Inversion)

The assumption today's price embeds, recovered by inverting the valuation.

The operating-margin figure is value-band context at year 6: derived from the framework's value band, a separate calculation — not part of the priced-in solve.

Solve inputs: computed at a 12.6% cost of capital; growth searched up to the 38.4% self-funding ceiling.

How unusual the bet is: n/a

Valuation X-Ray

Asset, earnings-power and peer-multiple models all land far below the price; ONLY the growth-DCF reaches it. The bet is durable compounding the static frames structurally cannot price (a moat/durability premium).

How the valuation models price the stock relative to the market price. Price/FV above 1.0 means the market pays more than that lens defends (expensive); at or below 1.0 the lens can defend the price.

Families that justify the price: Growth Families that call it expensive: Asset, Earnings, Relative

The models below discount at their own flat-beta convention rates (cost of equity 9.3%, WACC 9.1%); the inversion above states its own rate.

Per-Model Detail (n=15)

Economic-Unit Decomposition (Sum Of The Parts)

One or more material disclosed units has unresolved economics. Unknown/general is not evidence of homogeneity: segment SOTP is primary but incomplete, consolidated cash flow may remain only a secondary cross-check when every unit shares an enterprise basis, and one sector multiple or target margin is withheld.

No total common-equity value is stated. One or more material units lack a supported unit value. The displayed values are an indicative subtotal; consolidated debt and cash cannot be applied to a fraction of the company.

Solvency

Bullet Takeaways

• Hardware Platform Solutions, the engineering-led business Celestica sells alongside assembly, reached 5.0 billion dollars in 2025 and accounted for 41% of the year's total, against 29% a year earlier, which is the plainest evidence that the mix has moved off pure contract manufacturing.
• Concentration is the standing exposure: in 2025 three customers, all inside the connectivity and cloud segment, each represented 10% or more of total revenue, specifically 32%, 14% and 12%.
• The Enterprise end market is the near-term tell, since its revenue fell 19% during 2025 through a technology transition on an AI compute program at one hyperscaler customer, and the pace of its recovery lands in each quarterly print.

Bull Case

Contract manufacturing is a mature-stage business almost everywhere it is practised. Assemble what someone else designed, earn three or four cents on the sales dollar, and hope the customer's product cycle lasts. Celestica has been drifting out of that stage for several years, and the 2025 numbers are where the drift becomes visible. The 10-K describes the Hardware Platform Solutions business as an offering that helps to differentiate us from other EMS providers, by supporting customers in the design of advanced hardware and systems-level solutions, primarily for data centers, that can be customized to support their specific requirements, and calls it an engineering-led, IP-based offering. That is a claim about stage, not just about product. The supporting number is that HPS revenue grew 81% in 2025 to 5.0 billion dollars and moved from 29% of total revenue to 41%.

Where the growth came from is unusually specific. Communications end-market revenue rose 3,179.7 million dollars in 2025, an 81% increase the company attributes to data center networking demand and the ramping of its switch programs. Connectivity and cloud segment revenue for the year was 9.19 billion dollars, up 42% from 6.49 billion dollars. That is not a diversified lift across a broad customer base; it is one end market, running hard, on programs the company designed itself. The distinction matters for margins, because design content is what allows a manufacturer to keep more than the assembly fee.

The margin data supports the claim, if not dramatically. Celestica earns roughly 7.2% at the operating line on 13.79 billion dollars of trailing revenue. Set that against the cohort it is usually grouped with: SANM reports a 3.6% operating margin, JBL 4.3%, FLEX 4.9%, PLXS 5.2% and BHE 3.1%. Only TTMI, at 9.2%, runs higher, and TTMI is a printed-circuit specialist with a 21.0% gross margin, a different business shape entirely. The advanced technology segment is improving on its own account too: the 10-K reports ATS segment margin rising from 4.6% in Q4 2024 to 5.3% in Q4 2025, driven by better profitability in aerospace and defense and by the discontinuation of a margin-dilutive program. Management is spending to hold the position, telling shareholders it has increased our R&D investment in our global design services and capabilities to develop differentiated HPS product solutions for our customers and expects to increase it further.

The balance sheet is not the constraint. Net borrowings sit at about 0.4 times operating profit, trailing operating income covers interest roughly eighteen times over, and the company generates cash rather than consuming it. A growth-stage company with an investment-grade-looking capital structure can fund a design build-out from operations instead of from equity issuance, which is the difference between compounding and diluting. The bull case is not that Celestica has escaped the economics of its industry. It is that it has bought itself a genuine seat at the design table in the one end market currently spending the most, and that it can pay for the seat itself.

Bear Case

The moat, such as it is, rests on design content: Celestica is worth more than an assembler because customers hand it hardware problems, not just build instructions. The 2025 results contain a direct test of how much protection that actually buys. Enterprise end-market revenue fell 19% during the year, and the company's own explanation is a technology transition on an AI and machine-learning compute program with one hyperscaler customer. Design involvement did not prevent that customer from re-architecting, and it did not stop nineteen cents of every Enterprise dollar from leaving. The engineering relationship is real; it simply does not convert into switching cost at the moment switching is being considered.

The scale of that exposure is in the filing without any interpretation needed. In 2025 three customers, all in the connectivity and cloud segment, individually represented 10% or more of total revenue, at 32%, 14% and 12%. In 2024 it was two customers at 28% and 11%; in 2023 it was one at 22%. Concentration has been increasing, not diluting, through exactly the period the design story has been getting stronger. The risk factors decline to promise otherwise, stating there can be no assurance that our customers will renew their long-term manufacturing or services contracts with us on acceptable terms or at all. The receivables ledger carries the same shape: two customers each represented 10% or more of total accounts receivable at the end of 2025.

Competitors are moving into the same demand. SANM grew revenue 44.5% over its trailing year and TTMI grew 23.1%, and both describe data center and AI workloads in their own filings as the target. When five or six capable manufacturers chase the same handful of hyperscaler programs, the design premium is precisely the thing that gets negotiated away first, because the customer knows what the alternative bid looks like. Working capital compounds the pressure. The company notes that suppliers may also require us to purchase materials and components in minimum order quantities that may exceed current customer forecast requirements, which is the mechanism by which a cancelled program becomes an inventory writedown rather than merely a lost order.

All of which matters because of what the price is asking for. Today's quote implies operating profit compounding at roughly the fastest rate the business can fund out of its own earnings, near 25% a year, and holding there for something like 28 years. Only about 15% of comparable fast-growers have sustained that pace even a decade. The demand attaches to the advanced technology side of the business, the smaller of the two segments and the one that reported a 5.3% margin in the most recent fourth quarter. If that requirement is not met, the support underneath the price thins quickly: the peer-multiple methods already land well below today's quote, and the methods that capitalize current earnings with no growth credited at all land at a small fraction of it. Nothing about the balance sheet forces a bad outcome here. The valuation simply leaves no room for an ordinary one.

Valuation

Start with what the quote is asking the business to do. At 306.03 dollars a share, the price implies operating profit growing at about the ceiling the company can finance from its own earnings, roughly 25% annually, sustained for something on the order of 28 years, with the demand concentrated on the advanced technology segment. That is a long time. The reference points are not encouraging in either direction the data can be checked: the multiple sits at the very top of its peer distribution, well beyond the upper quartile, and among comparable fast-growers only about 15% held that growth rate for even a decade. Each additional percentage point of growth shortens the implied horizon by roughly three and a half years, which gives a sense of how sensitive the arithmetic is.

The methods used to triangulate the business split cleanly, and the split is the information. Only the forward-growth methods reach today's price. Every book-value-anchored approach lands below it, including the one that credits five full years of the trailing 45.7% return on equity before fading. The peer-multiple methods land roughly 27% under the current quote. The methods that capitalize what the business earns now, with no growth credited at all, reach only a small fraction of it. That pattern describes a durability premium rather than a mispricing: the static frames are structurally incapable of pricing compounding, so a company that genuinely compounds will always look expensive to them. The question is whether this one does.

Even inside the growth methods the agreement is partial. The version that carries today's enterprise-value multiple held flat through a seven-year projection clears the price comfortably; the version that fades growth into a perpetuity does not, landing under it. So the price is not simply defended by discounted cash flow. It is defended by the specific assumption that the current multiple survives the projection period intact, which is an assumption about market psychology as much as about the business.

Two filing-sourced anchors are worth holding next to that. Connectivity and cloud revenue of 9.19 billion dollars in 2025 was 42% above the prior year, so the growth being extrapolated is real and recent. And the segment the price leans on hardest reported a 5.3% margin in Q4 2025, which is respectable for its industry and thin against what a 28-year compounding assumption normally accompanies. Q1 2026 GAAP earnings came in at 1.83 dollars a share, including a 0.37 dollar per-share pre-tax charge, so the trailing earnings base is being built quarter by quarter rather than assumed.

On the downside, solvency is not what constrains this story. Net borrowings run about 0.4 times operating profit, trailing operating income covers interest expense roughly eighteen times, and free cash flow is positive rather than negative. Book value stands at 18.13 dollars a share against trailing earnings of 8.25 dollars, which is what a 45.7% return on equity looks like from the other side: very little capital is tied up producing the profit. The floor under the price is therefore an earnings floor, not an asset floor, and earnings floors move.

Catalysts

The financing question was settled in April. On April 27, 2026 the company amended its senior credit agreement, raising revolving commitments from 750.0 million dollars to 1,750.0 million dollars, refinancing the existing term A facility, which had 228.1 million dollars outstanding at March 31, 2026, into a new 250.0 million dollar term loan drawn in full at closing, and pushing the maturity of both from June 2029 out to April 2031. For a business whose growth consumes inventory and receivables before it produces cash, more than doubling the revolver is the operationally relevant part; the maturity extension mostly removes a 2029 conversation from the calendar.

Leadership of the segment carrying most of the revenue changed in July. Steven Dorwart became President of Connectivity and Cloud Solutions effective July 6, 2026, succeeding Jason Phillips, who retires at year end and stays on in an advisory capacity through the transition. Dorwart is a twenty-one-year veteran of the company and previously ran global accounts for the same segment, so the appointment reads as continuity of customer relationships rather than a change of direction. Given that three customers accounted for 32%, 14% and 12% of 2025 revenue, who owns those relationships is not a minor personnel detail.

The most recent print gives the operating baseline. Q1 2026 GAAP earnings were 1.83 dollars a share and included a 0.37 dollar per-share pre-tax charge, while research and development expense rose 23.6 million dollars year over year to support the platform solutions business. Interest expense under the credit facility, including the effect of interest rate swaps, was 10.7 million dollars in the quarter against 13.1 million dollars a year earlier. The line to watch through the next few reports is Enterprise, where the hyperscaler compute program transition took revenue down 19% across 2025 and where the shape of the rebound will say more about the durability of the design position than any single quarter's total.

Peer Cohorts (Per Segment, With Filing Citations)

CCS (Connectivity & Cloud Solutions) / ATS (Advanced Technology Solutions) (reported)

• SANM (Sanmina Corporation)
• FY2025 10-K: …solutions ideal for a wide range of applications including rack scale data storage and AI and machine learning workloads. With the increasing demand for AI, hybrid cloud and edge solutions for compute workloads in hyperscaler and enterprise data centers, VES is well positioned to take advantage of these trends. SCI.…
• FY2025 10-K: …original equipment manufacturers ("OEMs") in the following industries: industrial, medical, defense and aerospace, automotive, communications networks and cloud infrastructure. Our customer-focused organization with 39,000 employees, including 4,000 temporary employees, supports our customers from 20 countries on…
• JBL (JABIL INC)
• FY2025 10-K: …solutions provider. As we work to achieve our vision, we pursue the following strategies: • Establish and Maintain Long-Term Customer Relationships. An important element of our strategy is to develop and expand long-term relationships with leading companies in expanding industries that benefit from global, highly…
• FY2025 10-K: …wide spectrum of value-add design services to enhance our relationships with current customers and to build new customer relationships. Our teams are strategically staffed to support development projects of all sizes, from turnkey system design and joint development to industrialization and product optimization…
• PLXS (PLEXUS CORP.)
• FY2025 10-K: …oversee and provide leadership to business development and customer relationship management teams, supply chain, engineering, manufacturing and sustaining services subject matter experts and market sector specialists. These teams maintain expertise related to each market sector and execute sector strategies aligned…
• FY2025 10-K: …the world. • Sustaining Services - We are committed to protecting our customers' brand reputation, supporting the success of each product in the market and extending a product's useful life through repair, refurbishment and related services. From influencing a product's design for serviceability, which creates early…
• FLEX (FLEX LTD.)
• FY2025 10-K: ("Flex" or the "Company") is the advanced, end-to-end manufacturing partner of choice that helps a diverse customer base design, build, deliver and manage innovative products that improve the world. Through the collective strength of a global workforce across approximately 30 countries with responsible, sustainable…
• FY2025 10-K: …and Cloud , including data center, edge, and communications infrastructure • Lifestyle , including appliances, floorcare, smart living, HVAC, and power tools • Consumer Devices , including mobile and high velocity consumer devices. The FRS segment is optimized for longer product lifecycles requiring complex ramps…
• BHE (BENCHMARK ELECTRONICS, INC.)
• FY2025 10-K: …market opportunities and instead focus on low-to-medium-volume, high-complexity manufacturing programs, primarily within the A&D, semi-cap, medical, and industrial markets, which are often highly regulated and increasingly outsourced by OEMs. In the AC&C markets, we focus on customers with more complex requirements…
• FY2025 10-K: …growth or operational efficiencies, these cost increases could have a material adverse impact on our operating results and cash flow. Our business or stock price could be negatively affected by the actions of activist shareholders or others. Responding to actions by activist shareholders or others can be costly and…
• TTMI (TTM TECHNOLOGIES INC)
• FY2025 10-K: …components, assemblies, and subsystems. Orders for products generally correspond to the production schedules of our customers and are supported with firm purchase orders. Our customers have continuous control of the work in progress and finished goods throughout the PCB and engineered systems manufacturing process,…
• FY2025 10-K: …one-stop solution for customers. We are one of the largest technology manufacturers in North America and have a global sales and manufacturing presence. We believe we have significant economies of scale, an increasing end-to-end value chain solution for customers, a regionally diversified and resilient manufacturing…

Methodology Note

• Priced-in inversion: the valuation is inverted on the current price to recover the operating-income growth, duration, and steady-state margin the price embeds (ROE for financials, FFO growth for REITs).
• Valuation x-ray: the valuation models, grouped into four families (asset, earnings, relative, growth). Each model is expressed as a price/FV ratio (distance from price), not a point fair-value estimate. The spread across families is the disagreement.
• Economic-unit decomposition (SOTP): each disclosed business unit is assigned its native valuation basis before any multiple is applied. Operating units are valued on enterprise value; funded financial units are valued on their own common equity, because their borrowings fund earning assets rather than levering the parent. A company total is stated only once every material unit carries a supported value and the parent capital bridge reconciles.
• Solvency: net cash/debt, net-debt-to-NOPAT, interest coverage, and share-count CAGR from EDGAR financials (net debt / FFO and fixed-charge coverage for REITs; regulatory-capital framing for financials).
• Peer cohorts: per-segment comparables with deep-linkable SEC filing citations.

Fundamentals sourced from SEC EDGAR filings. Current price from Databento. The priced-in inversion and valuation x-ray are computed by the boothcheck engine; narrative composed by AI from the structured data.