Credo’s beat is not the story; the underpriced risk is hyperscale concentration funding a faster TAM expansion

The first read on this print should be that consensus was too low on the slope of the hyperscale ramp, not merely on the quarter. What was priced in was a company already expected to grow quickly, with the Street at $235.0 million of revenue and $0.49 of EPS, after a Q1 FY2026 base of $223.1 million and +273.6% revenue YoY. What actually surprised was that the company delivered $268.0 million of revenue, a +14.1% surprise, and $0.67 of EPS, a +36.2% surprise, while the quarterly history shows Q2 FY2026 revenue growth of +20.2% QoQ and +272.1% YoY. The variant perception is that investors focused on the obvious beat may still underweight the persistence embedded in the next guide: Daniel Fleming committed to Q3 fiscal ’26 revenue “between $335 million and $345 million, up 27% sequentially at the midpoint.” That language matters because the company did not frame Q2 as a pull-forward; it framed it as a step in a ramp that continues into the next quarter.

The financial trajectory supports that view because revenue acceleration is now arriving with gross margin stability rather than margin sacrifice. Credo’s quarterly revenue moved from $72.0 million in Q2 FY2025 to $135.0 million in Q3 FY2025, $170.0 million in Q4 FY2025, $223.1 million in Q1 FY2026, and $268.0 million in Q2 FY2026, while gross margin moved from 63.2% to 63.6%, 67.2%, 67.4%, and 67.5% across those same periods. That sequence matters more than the single-quarter beat because the market usually discounts fast ramps in semis when margin degrades, inventory builds without visibility, or operating expenses outrun revenue. Here, the printed history shows gross margin at 67.5% in Q2 FY2026, only modestly above 67.4% in Q1 FY2026, but materially above 63.2% in Q2 FY2025. The reported EPS history also moved from -$0.03 in Q2 FY2025 to $0.16 in Q3 FY2025, $0.20 in Q4 FY2025, $0.34 in Q1 FY2026, and $0.44 in Q2 FY2026, giving the beat a quality component rather than a pure revenue surprise.

The margin guide is the only real tension in the bull case, but it does not break the thesis because the revenue guide is large enough to absorb it. Management guided Q3 non-GAAP gross margin to 64% to 66%, below the Q2 FY2026 gross margin of 67.5% in the quarterly history and below the company’s own Q2 non-GAAP gross margin language of 67.7%. If that guide were paired with a flat top line, it would suggest pricing pressure or mix deterioration. Instead, the same guide pairs with $335 million to $345 million of Q3 revenue, compared with $268.0 million in Q2 on the Street-comparison basis and $268 million in the company’s own call language. The conflict is therefore explicit: gross margin guidance steps down to 64% to 66%, while revenue guidance steps up to $335 million to $345 million. That combination points to mix and ramp costs rather than demand weakness, especially because Fleming also guided Q3 non-GAAP operating expenses to $68 million to $72 million after Q2 non-GAAP operating expenses of $57.3 million.

The operating leverage in Q2 is why the EPS surprise should not be dismissed as accounting noise. Fleming said product revenue was $261.3 million in Q2, up 20% sequentially and up 278% year-over-year, while total non-GAAP operating expenses were $57.3 million, up 5% sequentially. That spread between 20% product revenue growth and 5% OpEx growth is the mechanism behind non-GAAP operating income of $124.1 million in Q2 versus $96.2 million in Q1, and non-GAAP operating margin of 46.3% versus 43.1% in the prior quarter. The market may have expected operating leverage directionally, but the magnitude of EPS delivery, $0.67 actual versus $0.49 estimate, suggests models were not carrying enough drop-through from the hyperscale ramp. The caution is that Q3 OpEx guidance of $68 million to $72 million is a step up from $57.3 million, so the next test is not whether Credo can grow revenue, but whether the incremental spending is tied to customer ramps that sustain the path beyond the current large customer.

That customer concentration is the central debate, and management did not try to hide it. Fleming disclosed, “The largest was 42% of revenue, and that was the customer that we've, in the past, said we expect to be the largest customer this fiscal year.” The quote earns attention because it does two things at once: it confirms the bear case that concentration is still high, and it pins that concentration to an already-flagged customer rather than a new surprise. In a $268.0 million quarter, 42% concentration is not a rounding issue; it is the core revenue engine. But the more important forward statement was Fleming’s expectation that “3 to 4 customers will be greater than 10% of revenue in the coming quarters and fiscal year as hyperscale customers continue to ramp to more significant volumes and as we expect to begin to ramp an additional hyperscale customer in the coming quarters.” That is the diversification bridge the stock needs: not a broad customer list today, but a path from one dominant account to 3 to 4 customers over 10% of revenue.

The supply-chain read-through is therefore concentrated but useful: Nanya Technology, identified as a Credo customer for Active Electrical Cables for AI clusters, sits on the positive side of this ramp because Credo’s product business generated $261.3 million of Q2 revenue, up 20% sequentially and up 278% year-over-year. That does not prove Nanya drove the quarter, and the data pack does not quantify Nanya’s share, so the defensible implication is narrower: demand for AECs in AI clusters is strong enough inside Credo’s reported product business to support $335 million to $345 million of Q3 revenue guidance and an expectation of 3 to 4 customers above 10% of revenue. The supplier side gives no named read-through because the supply-chain section lists no suppliers to Credo. For competitors, the implication is that Credo is taking share or expanding its served content in high-speed connectivity faster than large-cap semiconductor and platform peers are growing their reported top lines: NVDA’s latest reported revenue YoY was +85.2%, META’s was +33.1%, MSFT’s was +18.3%, GOOGL’s was +21.8%, and Credo’s Q2 FY2026 revenue YoY was +272.1%. Those are not like-for-like businesses, but the comparison frames how unusual Credo’s connectivity ramp is within the broader AI hardware complex.

The peer comparison also highlights why the stock should be valued on durability of design wins rather than on a generic fabless multiple. Credo’s 67.5% gross margin in Q2 FY2026 is almost level with MSFT’s 67.6% gross margin and above GOOGL’s 62.4%, AAPL’s 49.3%, AMZN’s 51.8%, and 2454.TW’s 46.3%, while still below NVDA’s 74.9% and META’s 81.9%. The relevant point is not that Credo is comparable to hyperscalers, because it is not; it is that Credo is scaling a semiconductor product business without showing the gross-margin compression typical of urgent volume ramps. That gives management more room to invest in the product roadmap. Brennan’s comment that ALC products will sample to lead customers during fiscal ’27 with initial revenue ramping in fiscal ’28 puts a date on the next leg, and his claim that the combined total market opportunity will exceed $10 billion in the coming years, more than triple where the company stood 18 months ago, tells investors what management is asking them to underwrite beyond the immediate AEC cycle.

That forward story would be less credible if the balance sheet and working-capital signals pointed to stress, but the cash and inventory data look more like ramp preparation than a demand crack. Cash flow from operations was $61.7 million in Q2, up $7.5 million sequentially, while free cash flow was $38.5 million, down from $51.3 million in Q1 because CapEx was $23.2 million, driven largely by purchases of production MACsec. Inventory ended Q2 at $150.2 million, up $33.5 million sequentially. The inventory build is the number to watch because it can either validate the $335 million to $345 million Q3 revenue guide or become the first sign of over-ordering if revenue does not materialize. Cash and equivalents of $813.6 million, up $333.9 million from the first quarter largely from proceeds of the ATM offering, also change the risk profile: Credo now has more funding to support production and roadmap investments, but the ATM proceeds mean investors should separate operating cash generation from financing-driven balance-sheet expansion.

The call delivery was constructive but not euphoric, which is exactly what a PM should want after a quarter this far ahead of Street numbers. The tone history shows Q2 FY2026 sentiment at 0.29 versus 0.28 in Q1 FY2026, guidance_tone at 0.53 versus 0.62, prepared_sentiment at 0.58 versus 0.56, qa_sentiment at 0.18 versus 0.19, uncertainty at 54.2 versus 55.6, and qa_evasiveness at -10.8 versus -10.2. In other words, management’s language did not become promotional as the numbers inflected; guidance_tone actually moved down from 0.62 to 0.53, while uncertainty eased from 55.6 to 54.2 and qa_evasiveness remained negative at -10.8. That matters because the print already supplies the upside. The call’s job was to avoid adding a narrative premium that could later unwind. The tone data says management kept the story tied to customer ramps, margin guidance, OpEx guidance, and dated roadmap milestones rather than stretching beyond the numbers.

The remaining concern is that the data pack also includes subsequent quarterly history showing Q3 FY2026 revenue of $407.0 million, gross margin of 68.5%, and diluted EPS of $0.82, followed by Q4 FY2026 revenue of $437.0 million, gross margin of 68.2%, and diluted EPS of $0.88. Because this essay is anchored to the 2025-12-01 earnings event for Q2 FY2026, those later rows should be treated as context for trajectory, not as part of the Q2 beat versus Street. They reinforce the variant perception that the Q2 guide was not aggressive enough, since the Q3 FY2026 history shows $407.0 million against the call guide of $335 million to $345 million, and gross margin of 68.5% against the Q3 non-GAAP gross margin guide of 64% to 66%. The basis is not identical across reported gross margin and non-GAAP guide, so the clean conclusion is about direction and scale, not exact margin beat math: the subsequent history does not support the idea that Q2 was a one-off peak.

What to watch next is precise. The thesis is confirmed if Q3 revenue tracks at or above management’s $335 million to $345 million range, if gross margin does not fall below the 64% to 66% non-GAAP guide, and if non-GAAP operating expenses remain within $68 million to $72 million while revenue growth carries the leverage. Customer concentration needs to move from risk toward validation: the largest customer was 42% of revenue in Q2, so investors should look for management to reaffirm 3 to 4 customers above 10% of revenue in the coming quarters and fiscal year, plus evidence that the additional hyperscale customer is beginning to ramp. Inventory is the early warning signal: Q2 ending inventory was $150.2 million, up $33.5 million sequentially, and that build needs to convert into Q3 revenue rather than linger. The roadmap watch is fiscal ’27 sampling of the first ALC products to lead customers and fiscal ’28 initial revenue ramping; a slip in those dates would weaken the argument that Credo is moving beyond a narrow AEC cycle. The bear case starts to work if Q3 revenue misses $335 million, gross margin breaks below 64%, OpEx exceeds $72 million without a corresponding revenue upside, or the 42% customer remains the only visible growth engine rather than becoming one of 3 to 4 customers over 10%.