Credo Technology's Growth and Margin Expansion Thesis: A Senior Analyst's Multi-Year Earnings Review

Overview of Emerging Themes and Tone Shifts

Across the 2022–2026 period, Credo’s transcripts show a clear evolution from “promising niche connectivity vendor” to “high‑growth, operating‑leverage AI infrastructure beneficiary.” Several durable, valuation‑relevant themes recur and strengthen over time:

1. AEC as an early growth engine, followed by a broader systems portfolio (optical DSP, PCIe/scale‑up, ZF optics, ALC).
2. Progressive improvement in revenue quality (more product, less NRE/IP mix volatility) and margin structure.
3. Customer and product diversification away from a single hyperscaler.
4. Increasing visibility and confidence in multi‑year AI infrastructure demand.
5. Demonstrable operating leverage and capital discipline.

These patterns, taken together, are consistent with profiles that have historically preceded substantial multiple expansion and stock price appreciation in comparable connectivity and semiconductor names.

---

1. Period‑by‑Period Trend Narrative

2022: Product‑Led Inflection, AEC Proof‑Point, Early Leverage

Business mix and positioning

• Credo frames itself as a pure‑play high‑speed connectivity company anchored in SerDes IP, AECs, optical DSPs, and line‑card PHYs, targeting hyperscale data centers and adjacent markets.
• Early emphasis is on AEC leadership at a first hyperscaler and a rapidly expanding Ethernet TAM. Management repeatedly points out they are first to market with system‑level AEC solutions and are taking full responsibility for reliability and qualification.

Revenue quality and growth profile

• FY22 revenue +90% YoY; product revenue up 104% YoY.
• AEC ramps quickly; the largest customer is ~30–40% of revenue and "much greater than 10%" of the total, signaling customer concentration but also a real, non‑speculative volume engine.
• IP is ~16–30% of revenue, well above the long‑term 10–15% target, but management consistently characterizes IP as strategic, not primary, and stresses that long‑term mix will be more product‑heavy.

Margins and operating leverage

• Product gross margin climbs into the low‑50s (≈52–53%), rising ~800–900 bps over several quarters, explicitly tied to scale rather than one‑off factors.
• Total gross margin moves into the 60%+ range when IP is elevated.
• OpEx grows ~50% YoY versus ~90%+ revenue growth, and management calls out an explicit objective that the top line grow at least 2× OpEx “for the foreseeable future.”
• This yields positive operating income and double‑digit operating margins already in FY23 guidance, unusual for a newly public connectivity name.

Tone and risk framing

• Management is clearly bullish but still somewhat “evangelizing” AEC: explaining why AEC wins versus optics, why Ethernet will outgrow InfiniBand, and why 400G/800G transitions create structural tailwinds.
• They acknowledge supply‑chain risk (China COVID lockdowns) but emphasize strong relationships and buffer inventory.
• Guidance is already multi‑quarter and multi‑year (at least $200m in FY23; gross margin model 59–61% then 63–65% over time), signaling confidence and planning rigour.

Investor‑relevant takeaway:
By late 2022, Credo has a validated AEC product, tangible hyperscaler volume, expanding product gross margin, and visible operating leverage. That is a classic setup for the market to begin capitalizing the company as a structural beneficiary of higher‑speed Ethernet rather than a one‑off IP licensor.

---

2023: Volatility Shock, Then Rebuild Around AI and Diversification

Microsoft reset and digestion

• FY23 ends with a sharp reset: Microsoft (the lead AEC customer) cuts forecasts, driving a steep revenue step‑down and guiding FY24 revenue flat YoY.
• Management is explicit that the issue is inventory digestion and program push‑outs, not share loss or product problems. They reiterate Credo remains sole‑sourced there.

Shift in strategic emphasis

• While acknowledging near‑term pain, management doubles down on:
  • Adding a second AEC hyperscaler in FY24 (and starting to engage third/fourth).
  • Building out optical DSP as a second major vector, with design wins at US and non‑US hyperscalers and the LRO architecture positioned for 800G/1.6T.
  • Early chiplet traction (Tesla Dojo) and future PCIe/USB opportunities.

Execution through the downturn

• Despite flat FY24 revenue guidance, they commit to:
  • Stable gross margin (roughly flat YoY) by offsetting lost AEC scale with a richer IP mix and higher‑margin products.
  • OpEx growth well below prior revenue expectations, preserving balance sheet flexibility and operating discipline.
• They outline a path to sequential growth through FY24, exiting the year at a “very high rate of revenue” with more diversified customers and products.

Tone evolution

• The tone shifts from growth evangelism to damage‑containment and risk management but remains confident on the medium term.
• Management openly discusses customer concentration as a risk and lays out a concrete plan to reduce it (adding a second hyperscaler ramp, broadening product lines, targeting 3–4 10%+ customers over time).

Investor‑relevant takeaway:
2023 introduces real volatility, but the transcripts show a management team that: (1) doesn’t flinch on long‑term TAM and technology positioning; (2) uses the reset to pivot the narrative toward diversification and AI data‑center builds; and (3) protects the gross‑margin model. Historically, names that manage through a single‑customer air pocket without margin capitulation often re‑rate once the growth engine restarts.

---

2024: AI Demand Inflection, Mix Upgrade, and Reacceleration

Return to growth, now AI‑centric

• FY24 revenue returns to growth (+5% YoY), but the mix changes dramatically: by Q4‑24, ~75% of revenue is AI‑related, vs much lower a year earlier.
• Product revenue re‑accelerates; IP remains lumpy but within the 10–15% target range on a full‑year basis.

Customer and product diversification

• Multiple data points of reduced single‑customer risk:
  • Three 10% customers by early FY25, then four 10% customers later, spanning AEC, optical DSP, line‑card PHY, and chiplets.
  • Seven customers at >5% of revenue by FY25‑Q2, suggesting a broader demand base.
• AEC is still the largest growth driver, but optical DSP is now clearly a real business with a target of ≥10% of revenue in FY25 and millions of modules already deployed.

Margin and leverage inflect again

• Gross margin pushes into mid‑60s (61–66%), often above the high end of the long‑term 63–65% target.
• Product gross margin (ex‑NRE) steadily increases with scale; management repeatedly attributes improvement to volume leverage and mix, not pricing or one‑time items.
• Operating margin returns to double digits by Q4‑24 and is guided to expand as sequential revenue growth continues and OpEx grows only modestly.

Guidance quality and visibility

• Management begins to talk about fiscal‑year frameworks, not just quarters:
  • FY25 revenue expected to grow 100%+ YoY.
  • AI revenue from Q4‑24 to Q4‑25 expected to double.
  • Gross margin modeled at 61–63% for FY25, with OpEx growing at half the rate of revenue.
• There is clear confidence in mid‑term AI infrastructure spending and the durability of Credo’s attach.

Tone and messaging

• The tone evolves from cautious (post‑Microsoft reset) to measured confidence. Management references “many irons in the fire,” but pairs this with detailed comments on design‑win activity, customer qualification status, and unit volumes.
• They emphasize execution consistency: delivering on previously communicated inflection timing (2H FY25) and margin roadmaps.

Investor‑relevant takeaway:
By late 2024, Credo looks far less like a single‑product/single‑customer story and more like a diversified AI connectivity platform with high and rising gross margins. That combination—post‑volatility stabilization + AI demand tailwind + clean operating leverage—is typically where valuation multiples begin to expand for well‑positioned semiconductor names.

---

2025: Hyper‑Growth, Operating Leverage, and Pipeline Breadth

Revenue and profit acceleration

• FY25 revenue reaches ~$437m (+126% YoY), with product revenue up 157% YoY.
• Quarterly revenue nearly triples from Q1‑25 to Q4‑25 (≈$59m → $170m) as AI deployments ramp across hyperscalers.
• Non‑GAAP gross margin for FY25 is 65%; Q4‑25 gross margin is 67.4%, driven mainly by scale.
• Operating margin improves by 2,500 bps YoY; Q4‑25 operating margin is 36.8%, with net margin 38.4%—well above the long‑term 28–33% model.

Customer scaling and diversification

• AEC remains the primary revenue engine, but with multiple large customers:
  • By Q4‑25, three customers each >10% of revenue (largest ~61%, then 12% and 11%).
  • Management expects two additional hyperscalers to begin ramping in 2H FY26, targeting 3–4 10%+ customers ongoing.
• There is explicit commentary that each hyperscaler deploys multiple SKUs and programs (front‑end, back‑end/scale‑out, disaggregated chassis), increasing depth per account.

Emergence of new vectors: PCIe/scale‑up and ZF/optical systems

• Management introduces PCIe‑based solutions for AI scale‑up networks, targeting design wins in calendar 2025 and production revenue in calendar 2026.
• They frame PCIe/UCIe as a “very large market” across server retimers, scale‑up fabrics, and chiplets, and position Credo as one of few vendors with SerDes capability at 200G/lane.
• On optical, they highlight LRO and full‑DSP solutions, now shipping millions of units, with FY25 optical DSP revenue targeted at ≥10% of company revenue and strong momentum into FY26 and beyond.

Operating model and capital allocation

• Management explicitly guides that from FY24→FY25 and FY25→FY26, OpEx grows at less than half the rate of revenue, resulting in:
  • FY26 net margin approaching 40% (from 65% gross, high‑30s operating).
  • Operating margin expansion tracking gross margin improvement with modest CapEx (even if tape‑outs double).
• They maintain a large cash buffer ($400m+ exiting FY25), positioning the company to invest through cycles without stressing the balance sheet.

Tone: from inflection to multi‑wave growth thesis

• The language shifts to “multiple waves of growth opportunities” tied to:
  • Scale‑out AEC for AI training and inference.
  • Scale‑up PCIe for GPU‑GPU and GPU‑memory connectivity.
  • Optical system‑level solutions (ZF optics, LRO) for 800G/1.6T.
• Management repeatedly emphasizes connectivity bottlenecks everywhere and Credo’s aim to address them at the system level, not just as a component vendor.

Investor‑relevant takeaway:
By FY25, Credo is displaying the textbook pattern of a “new‑category leader” scaling into a structural demand wave: triple‑digit top‑line growth, high‑60s gross margin, 30‑40% operating margin, and a credible multi‑year innovation roadmap that expands TAM beyond the initial AEC niche.

---

2026 YTD: Consolidation at Higher Scale and Structural Leverage

Sustained high growth from a larger base

• Q1‑26 revenue: $223m (+274% YoY, +31% QoQ), Q2‑26 revenue: $268m (+272% YoY, +20% QoQ).
• Management guides FY26 revenue to exceed $800m (>85% YoY) and then lifts the full‑year view to >170% YoY as execution outpaces earlier mid‑single‑digit sequential growth expectations.
• Both Q1 and Q2 revenue repeatedly come in “well above the high end” of guidance; guidance for subsequent quarters remains deliberately conservative.

Margins at structurally higher levels

• Q1‑26 gross margin: 67.6%; Q2‑26: 67.7%, above the high end of the 63–65% long‑term model.
• Product gross margin in mid‑high 60s; management explicitly says they do not intend to raise the long‑term model but expect to operate “at or above” the high end.
• Operating margin in mid‑40s (43.1% in Q1, 46.3% in Q2); net margin ~44–48%.
• Guidance for FY26: non‑GAAP net margin ≈45%, with net income more than quadrupling YoY.

Revenue quality and diversification

• AEC continues to grow “healthy double digits sequentially,” hitting record levels with four domestic hyperscalers, each >10% of revenue in Q2‑26.
• Management expects 3–4 hyperscalers above 10% of revenue in coming quarters and a fifth large customer ramping, though this fifth will not hit 10% in FY26.
• They highlight growing contributions from:
  • ALC (active long‑reach copper), whose TAM they argue is 2× AEC, with similar economics but longer reach and thinner gauges for row‑scale networks.
  • System‑level ZF optics starting at 800G, with a roadmap to 1.6T and differentiated by full‑stack integration (DSP + software + module).

Strategic positioning: protocol and compute agnosticism

• Credo emphasizes its solutions are agnostic to underlying compute: merchant GPUs, custom ASICs, and CPUs are all supported, across Ethernet and PCIe.
• They point to upcoming PCIe design wins (calendar 2025) as an additional attach vector in AI scale‑up networks.

Tone: confident but methodical

• The language is now unmistakably that of a company with a durable franchise, not just a cyclical beneficiary:
  • Repeated emphasis on execution consistency: “We set mid‑single‑digit sequential growth expectations several quarters ago and have outperformed them.”
  • Explicit statements that the long‑term gross‑margin model holds even as the product mix shifts toward system‑level optical and copper solutions.
  • Discipline on OpEx: FY26 OpEx up ~50% YoY vs revenue more than doubling.

Investor‑relevant takeaway:
By mid‑FY26, Credo exhibits sustained scale economics—high‑60s gross margin, mid‑40s operating margin, four 10%+ customers, multiple product pillars, and a strong balance sheet. The tone and numbers together support a thesis of enduring pricing power and structural demand, conditions under which valuation multiples often stay elevated even as growth inevitably moderates.

---

2. Cross‑Cycle Themes that Matter for Valuation

A. Improving Revenue Quality and Diversification

Evidence and trajectory

• Early years: meaningful dependence on IP/NRE and a single AEC hyperscaler.
• Mid‑period shock: Microsoft reset exposes single‑customer risk but does not undermine product or tech.
• Current state: 3–4 hyperscalers at >10% of revenue, 7+ customers >5%, and IP a stable 10–15% of revenue on an annualized basis.

Valuation implication
This pattern—transition from customer‑concentrated, IP‑heavy revenue to a broad base of hyperscale product shipments—is typically where the market moves from valuing a company on peak “deal” earnings to valuing it on structural earnings power. That tends to support higher and more stable multiples.

B. Operating Leverage and Margin Expansion

Observed path

• 2022: low‑double‑digit operating margins emerging on ~60% gross margin.
• 2023 reset: margins pressured but gross margin framework maintained; management consciously protects the model.
• 2024–2026:
  • Gross margin climbs into mid‑high 60s, often above guidance ceilings.
  • Operating margin expands from negative/low‑teens through 30%+ to mid‑40s.
  • Net margin approaches or exceeds 40% with strong FCF conversion.

Drivers

• Scale in AEC and later in ALC/optical system solutions.
• OpEx growth consistently held at ~½ of revenue growth through the hyper‑growth years.
• No visible structural gross‑margin compression as competition emerges; management explicitly maintains margin targets despite bringing new, more complex system offerings to market.

Valuation implication
Sustained margin outperformance vs long‑term models, coupled with explicit OpEx discipline, typically leads investors to re‑rate the quality of earnings—i.e., investors may begin to assume high‑20s/30s operating margins are sustainable, not cyclical peaks. That tends to both lift the earnings base and justify higher P/E or EV/EBIT multiples.

C. Structural AI Demand and Pipeline Visibility

Demand backdrop

• Management repeatedly notes AI clusters growing from tens of thousands of GPUs to hundreds of thousands and “soon millions.”
• They frame a clear linkage between GPU proliferation and interconnect intensity—more GPUs → more AEC/ALC/optical links per GPU, longer reaches (row‑scale), and higher rates (100G→200G per lane).

Visibility signals

• Across 2024–26, Credo consistently guides for:
  • Sequential mid‑single‑digit revenue growth on a much higher base, and then beats that bar.
  • AI revenue doubling from Q4‑24 to Q4‑25, and FY26 revenue more than doubling again.
  • Multi‑year tape‑out and design‑win plans (PCIe Gen6/200G‑lane, ZF optics roadmap, ALC TAM 2× AEC).

Valuation implication
Names that can credibly tie their growth algorithm to hyperscale AI capex, while showing multi‑year design‑win visibility, often earn “secular growth” multiples. The transcripts show Credo steadily building and communicating that visibility rather than relying on one‑off upside surprises.

D. Expansion from Component to System‑Level Solutions

Strategic shift

• Initial positioning: SerDes IP and discrete components (retimers, PHYs).
• Current positioning: system‑level AEC/ALC/optical where Credo owns the full stack (DSP, firmware/software, mechanical design, and reliability spec), and increasingly PCIe‑based scale‑up fabrics.

Implications

• Higher value‑add and differentiation—customers buy reliability, power, and system TCO, not just a chip.
• Stickier engagements: multiple SKUs per hyperscaler, across front‑end and back‑end networks and within disaggregated chassis.
• Greater pricing power and resilience against commoditization.

Valuation implication
System‑level ownership typically commands better long‑term margins and higher switching costs. Markets tend to pay up for companies that migrate from component suppliers to solution partners for the hyperscalers.

E. Execution Credibility and Guidance Culture

Observed behaviors

• The company consistently lays out explicit long‑term models (gross margin 63–65%, net margin 28–33%, then later net margin ~40%) and then meets or beats them.
• During the Microsoft reset, management does not try to “guide through” the issue; instead, they level‑set the year to flat and focus on rebuilding the growth algorithm.
• In the AI upcycle, they keep mid‑single‑digit QoQ growth guidance even as actual results beat by wide margins, avoiding the trap of chasing a hyper‑aggressive bar.

Valuation implication
Markets tend to assign premium multiples to management teams that under‑promise/over‑deliver and can navigate both positive and negative shocks without large guidance whiplash. Credo’s transcript history increasingly fits that pattern.

---

3. Durable, Thesis‑Level Conclusions

Based solely on the transcript record, several durable trends emerge that are typically supportive of a rising stock trajectory over the medium to long term:

1. From single‑customer to multi‑hyperscaler franchise.
   The story evolves from one Microsoft‑heavy AEC ramp to multiple 10%+ customers across AEC, ALC, and optical, with additional hyperscalers in the queue. That materially reduces idiosyncratic risk.

2. Structural margin uplift driven by scale and mix, not one‑offs.
   Gross margins have migrated from high‑50s to high‑60s while operating margins have climbed toward mid‑40s, with management attributing this almost entirely to scale and product mix. This pattern argues for durable earnings power rather than temporary peak margins.

3. Multi‑wave AI connectivity growth rather than a single node.
   Credo is positioning across scale‑out (AEC/ALC Ethernet), scale‑up (PCIe/UCIe), and optics (ZF, DSP, LRO), suggesting participation in multiple future waves—800G, 1.6T, 200G/lanes, and beyond—rather than a one‑generation spike.

4. Upgrade in revenue quality and visibility.
   The shift from lumpy IP/NRE to high‑volume product shipments, alongside more customers at scale and multi‑year visibility into AI programs, supports more reliable forecasting and, in turn, more stable valuation multiples.

5. Execution track record through both up and down cycles.
   The company has now navigated a major customer digestion event and an AI‑driven hyper‑growth phase while largely adhering to, and then exceeding, its long‑term financial models. That builds investor confidence that future guidance is credible.

Taken together, these elements form the backbone of a forward‑looking thesis: Credo is transitioning from an early‑stage, single‑product story into a structurally profitable, multi‑product, multi‑customer platform at the heart of AI data‑center connectivity bottlenecks. Historically, when semiconductor and connectivity companies demonstrate this combination of sustained high growth, rising margins, expanding TAM, and evidence of execution discipline, equity markets have tended to reward them with both earnings upgrades and higher valuation multiples, supporting a rising stock trajectory over the medium to long term.