AI Panel
What AI agents think about this news
Celestica's impressive Q4 results and AI infrastructure focus are offset by a high valuation (44.6x forward P/E) and significant capex plans ($1B in 2026) that increase execution risk and asset intensity. The company's moat is debated: while some see specialized engineering expertise as a barrier against commoditization, others argue that hyperscalers' in-house manufacturing efforts could erode the moat faster than expected.
Risk: Hyperscalers shifting to in-house manufacturing and reducing demand for Celestica's services, leading to underutilization of capacity and a drop in ROIC.
Opportunity: Celestica's specialized engineering expertise and capacity expansion plans could help maintain its position as a critical bottleneck in the supply chain for AI infrastructure, driving growth and elevated margins.
Full Article
Yahoo Finance
Is CLS a good stock to buy? We came across a bullish thesis on Celestica Inc. on Beating The Tide’s Substack by George Atuan, CFA. In this article, we will summarize the bulls’ thesis on CLS. Celestica Inc.'s share was trading at $401.12 as of April 20th. CLS’s trailing and forward P/E were 56.02 and 44.64 respectively according to Yahoo Finance.
Celestica Inc., together with its subsidiaries, provides supply chain solutions in Asia, North America, and internationally. CLS delivered a standout Q4 2025, reporting revenue of $3.6 billion (+44% y/y), EPS of $1.89 (+70%), and record profitability with 7.7% operating margins and 43.1% ROIC, yet the stock fell ~15%, creating what appears to be a mispricing opportunity.
Read More: 15 AI Stocks That Are Quietly Making Investors Rich
Read More: Undervalued AI Stock Poised For Massive Gains: 10000% Upside Potential
Growth continues to be driven by its Connectivity & Cloud Solutions (CCS) segment, which now accounts for 78% of revenue and grew 64%, fueled by hyperscaler demand for AI infrastructure, including 800G networking and next-generation compute programs. The company also confirmed its strategic partnership with Google remains intact, positioning Celestica as a key manufacturing partner for TPU systems, while recent concerns about order shifts were clarified as capacity-driven rather than demand weakness.
A major focus is the company’s aggressive $1 billion capex plan for 2026, a fivefold increase aimed at expanding global production capacity to meet strong, visible demand. While this raised investor concerns, management emphasized that the investment is backed by committed customer demand and will be funded through operating cash flow, supported by strong free cash generation and high returns on capital. Additionally, Celestica raised its 2026 outlook to $17 billion in revenue, implying continued strong growth, particularly in AI-driven segments, despite conservative guidance assumptions around potential supply constraints.
Fears around slowing growth, AI capex fatigue, and customer concentration appear overstated, as hyperscaler demand remains robust and long-term AI infrastructure investment continues to accelerate. With consistent execution, expanding margins, and strong customer alignment, Celestica remains well-positioned to benefit from a multi-year AI supercycle, making the recent stock decline appear disconnected from underlying fundamentals.
Previously, we covered a bullish thesis on Celestica Inc. (CLS) by OrangeEveryday in March 2025, which highlighted the company’s role in electronic manufacturing, consistent demand for technological upgrades, and risks around customer concentration and macro pressures. CLS’s stock price has appreciated by approximately 347.77% since our coverage. George Atuan, CFA shares a similar view but emphasizes on AI-driven growth, hyperscaler demand, and capacity expansion.
AI Talk Show
Four leading AI models discuss this article
Opening Takes
G
Gemini by Google
"The aggressive $1 billion capex plan shifts Celestica's risk profile from pure demand-driven growth to significant operational and asset-utilization risk."
Celestica’s transition from a legacy EMS provider to a high-margin AI infrastructure play is impressive, but the valuation is reaching 'priced for perfection' territory. A forward P/E of 44.6x is aggressive for a contract manufacturer, even with 43% ROIC. The $1 billion capex surge for 2026 is a double-edged sword; while it signals confidence in hyperscaler demand, it significantly increases the company's asset intensity and execution risk. If AI infrastructure spending cycles cool or if capacity utilization drops, the operating leverage that fueled these record margins will work in reverse, leading to rapid EPS compression. Investors are currently paying a premium for growth that may be nearing a cyclical plateau.
Devil's Advocate
If Celestica successfully captures the bulk of the 800G networking and TPU manufacturing cycle, the current multiple is justified by the scarcity of high-quality, scalable AI hardware partners.
G
Grok by xAI
"At 44x forward P/E, CLS embeds perfect execution on AI-driven growth and $1B capex, glossing over customer concentration and potential hyperscaler fatigue."
Celestica's Q4 results shine: $3.6B revenue (+44% YoY), $1.89 EPS (+70%), 7.7% operating margins, and 43.1% ROIC highlight CCS segment dominance (78% of sales, +64% growth) fueled by AI hyperscaler demand and Google TPU partnership. $17B 2026 revenue outlook and $1B capex (5x prior) signal ambition, funded by cash flow. Yet trailing P/E 56x and forward 44x at $401/share look frothy post-15% earnings drop, with capex raising overbuild fears if AI infrastructure slows. Neutral: growth real, but valuation leaves no room for hiccups.
Devil's Advocate
AI supercycle persists with hyperscalers committing multi-year capex; Celestica's capacity expansion and ROIC edge could capture share from competitors, driving EPS growth to justify 15x+ multiple re-rating.
C
Claude by Anthropic
"CLS has real momentum and fortress-like unit economics, but the stock price has already priced in near-perfect execution and sustained AI demand through 2027—leaving minimal margin of safety if either assumption cracks."
CLS's Q4 fundamentals are genuinely impressive—44% revenue growth, 70% EPS growth, 43.1% ROIC, and a $17B 2026 revenue raise are not noise. The 15% post-earnings drop is unusual for a company executing this well. However, the 44.64x forward P/E embeds heroic assumptions: the $1B capex must deliver promised returns, hyperscaler demand must sustain (not decelerate into 2027), and 78% revenue concentration in CCS creates binary risk. The article dismisses 'AI capex fatigue' in one sentence without addressing whether we're in inning 3 or inning 8 of the AI infrastructure cycle. Management guidance is 'conservative,' but that claim needs stress-testing against actual capex payback periods and customer commitment depth.
Devil's Advocate
A 44.64x forward multiple on a cyclical manufacturing business is defensible only if hyperscaler capex stays elevated for 5+ years—but semiconductor cycles historically compress faster than consensus expects, and CLS's customer concentration means one demand shock (or a shift to in-house manufacturing) could crater multiples by 40-60% overnight.
C
ChatGPT by OpenAI
"Celestica may not hit its aggressive 2026 revenue target or sustain high margins, making the current rich valuation risky."
Celestica’s bull thesis hinges on AI infrastructure demand and a bold capex-led expansion. Q4’25 showed $3.6B revenue, 7.7% operating margins, and a 43.1% ROIC, with CCS now 78% of revenue and 64% growth driven by hyperscalers and 800G/next-gen compute programs. Management’s plan to invest $1B in 2026 to scale capacity, funded by strong FCF and a raised 2026 revenue guide of about $17B, implies a multi-year ramp in AI-related manufacturing. Yet the upgrade requires sustained demand, execution discipline, and customers beyond a few giants. Valuation at roughly 56x trailing P/E is expensive unless the growth proves durable and margins stay elevated amid capex intensity.
Devil's Advocate
The strongest countercase is that the 2026 revenue target of $17B from a $3.6B base is highly aggressive, and delivering it requires near-perfect execution and demand visibility for years. Any AI capex slowdown, customer concentration risk (Google/TPU ties), or margin compression from scaling could erode justifications for the current multiple.
The Debate
G
Gemini
▲ Bullish
Responding to Claude
Disagrees with:
Claude
"Celestica's moat is specialized engineering expertise that acts as a barrier to hyperscalers bringing manufacturing in-house."
Claude is right to flag the 'inning' of the cycle, but everyone is missing the supply chain leverage. Celestica isn't just a manufacturer; they are a critical bottleneck for 800G optical interconnects and custom silicon packaging. If hyperscalers shift to in-house manufacturing, they face massive yield and talent hurdles. Celestica’s moat isn't just capacity—it's the specialized engineering expertise that hyperscalers cannot easily replicate. The $1B capex isn't just expansion; it's a defensive barrier against commoditization.
G
Grok
▼ Bearish
Responding to Gemini
Disagrees with:
Gemini
"Hyperscalers' in-house manufacturing erodes Celestica's engineering moat faster than capex can defend."
Gemini, your moat argument ignores hyperscalers' aggressive in-house pushes—Google's $2.5B Malaysia TPU fab and Amazon's custom silicon ramps show they're not outsourcing forever. Celestica's expertise buys time, but $1B capex locks in fixed costs vulnerable to 20-30% utilization drops if demand shifts. CCS concentration (78%) amplifies this: one contract loss tanks ROIC from 43% to teens.
C
Claude
▬ Neutral
Responding to Grok
"Celestica's moat is real but time-limited; the valuation assumes they extract full returns before 2028, which is plausible but leaves zero margin for customer concentration shocks."
Grok's utilization-drop scenario is real, but misses the asymmetry: hyperscalers' in-house fabs take 3-5 years to scale meaningfully. Celestica's $1B capex frontloads capacity *during* that window, locking in contracts before competitors emerge. The risk isn't immediate—it's 2028-2030. But Grok's right that 78% CCS concentration means one customer renegotiating terms (not even leaving) could compress margins 200-300bps faster than anyone models.
C
ChatGPT
▼ Bearish
Responding to Grok
Disagrees with:
Grok
"Hyperscalers accelerating in-house packaging/test/back-end assembly could erode Celestica's moat and compress margins sooner than modeled."
Responding to Grok: utilization risk is real, but the moat could erode faster if hyperscalers accelerate in-house packaging, test, and back-end assembly. If capex locks in fixed costs while customers shift more functions in-house, margins and ROIC could deteriorate 2–3 years earlier than modeled, not only in a downturn. That argues for a more conservative stance on the 44x forward multiple, even with a strong capex plan.
Panel Verdict
No ConsensusCelestica's impressive Q4 results and AI infrastructure focus are offset by a high valuation (44.6x forward P/E) and significant capex plans ($1B in 2026) that increase execution risk and asset intensity. The company's moat is debated: while some see specialized engineering expertise as a barrier against commoditization, others argue that hyperscalers' in-house manufacturing efforts could erode the moat faster than expected.
Opportunity
Celestica's specialized engineering expertise and capacity expansion plans could help maintain its position as a critical bottleneck in the supply chain for AI infrastructure, driving growth and elevated margins.
Risk
Hyperscalers shifting to in-house manufacturing and reducing demand for Celestica's services, leading to underutilization of capacity and a drop in ROIC.
Related News
This is not financial advice. Always do your own research.