Will CRDO's top 3 customer concentration fall below 80% by Q2 FY2027?

The hyperscaler AEC market inherently has limited buyers — 6 US hyperscalers plus Asia and Neoclouds. While diversification improved from 1 to 5 contributing hyperscalers, the revenue distribution remains extremely top-heavy (39%, ~25%, ~24%). Getting from 88% to below 80% requires the 4th and 5th hyperscalers to reach meaningful scale AND the top 3 to not increase their share disproportionately as total revenue grows toward $2B. Given order lumpiness and the structural limited-buyer market, below 80% within 3 quarters is ambitious.

Analyzing the math: if FY2027 Q2 quarterly revenue is ~$500M (consistent with $2B annual), getting top 3 below 80% means top 3 at <$400M and others at >$100M. Currently #4 and #5 are contributing 'initial' and smaller amounts. Growing from perhaps $20-30M combined to >$100M in 3 quarters is aggressive. ZF Optics Neocloud customers (Tensor Wave) add diversification but are small relative to hyperscaler volumes. The largest customer's share has been decreasing (61% → 39%), but the top 3 aggregate has remained stubbornly high.

The diversification trajectory IS positive: 1 → 5 hyperscaler customers in 4 quarters is rapid. If the 5th hyperscaler ramps similarly to the 3rd and 4th (which went from initial to meaningful in 2-3 quarters), and ZF Optics adds Neocloud revenue, breaking below 80% is within the range of plausibility. However, the Q2 FY2027 timeline (3 more quarters) is tight. More likely this takes 5-6 quarters to achieve. Probability is below 40% but not negligible given the rapid expansion pace.

The 88% concentration has proven sticky. The largest customer dropped from 61% to 39%, but the concentration rebalanced among top 3 rather than truly diversifying. This pattern suggests the top 3 hyperscalers have the largest AI infrastructure builds and will continue dominating Credo's revenue even as new customers add. Getting below 80% in 3 quarters requires a structural shift in ordering patterns that is unlikely without a new product category reaching major scale.

Customer concentration below 80% would require the 4th and 5th hyperscalers plus Neoclouds to contribute >20% of total revenue — roughly >$100M on a $500M quarterly base. The 5th hyperscaler just began 'initial' revenue in Q3 FY2026. Even with rapid ramp, achieving $50M+ from the 5th customer alone within 3 quarters is optimistic given qualification cycles. ZF Optics Neocloud customers help but are small-scale. Probability below 35%.

There is genuine uncertainty here because Credo's customer addition pace has been remarkable — 5 hyperscalers in 4 quarters. If the 5th customer ramps as fast as the 3rd did (from initial to ~15% of revenue in 2 quarters), and if ZF Optics opens up the Neocloud segment, breaking 80% is possible but would require favorable ordering timing. The order lumpiness means any single quarter's concentration % can vary significantly. LOW confidence because the outcome depends heavily on customer-specific ordering timing.

88% to below 80% in 3 quarters is a significant improvement. The 5th hyperscaler is at initial revenue stage. While diversification trajectory is positive, the top 3 continue to dominate AEC orders due to their massive AI infrastructure builds. More likely this takes 5+ quarters.

The structural limited-buyer market is the key constraint. Even with 5 hyperscalers, the top 3 have the largest data center builds and will naturally consume the most AECs. ZF Optics Neoclouds are too small to materially shift concentration. Below 80% is unlikely within the timeline.

Diversification is improving but 80% is a high bar to clear in 3 quarters. The trajectory from 1 to 5 customers is positive, and 4th/5th customers are ramping. However, the math requires substantial non-top-3 revenue growth. Probability is near 30%.