Credo Technology filed an 8-K on 1 September 2026 under Item 2.02 to report its fiscal Q1 2027 results (the quarter ended 1 August 2026). Revenue more than doubled year over year to $479.0 million and guidance for the current quarter came in above what analysts expected — but reported GAAP profit fell sharply from the prior quarter, for a reason that is entirely explainable.

The numbers

• Revenue: $479.0m, +114.7% YoY (from $223.1m) and +9.6% QoQ (from $437.0m). Consensus was about $470.4m — a modest beat.
• Non-GAAP EPS: $1.20 vs consensus of roughly $1.16 — a beat. Up 130.8% YoY (from $0.52), but only +3.4% QoQ (from $1.16).
• GAAP EPS: $0.67 — up 97% YoY, but down 24% from $0.88 last quarter. This is the distorted line; see below.
• Non-GAAP gross margin: 68.0%, versus 68.3% last quarter and 67.6% a year ago — essentially flat.
• GAAP gross margin: 64.5%, down from 68.2% last quarter — the entire gap is $11.0m of amortisation of acquired intangibles now running through cost of revenue.
• Non-GAAP operating margin: 48.2%, versus 49.6% last quarter and 43.1% a year ago.
• Operating expenses: GAAP opex $188.4m (from $142.2m); non-GAAP opex $95.2m (from $81.7m) — R&D alone rose to $114.5m from $90.5m.
• Balance sheet: goodwill jumped from $92.8m to $986.4m and intangibles from $29.3m to $378.8m; cash and equivalents fell from $1,165.0m to $466.9m. Cash plus short-term investments ended at $764.3m.
• Guidance: Q2 FY27 revenue of $525–535m (midpoint $530m, about +10.6% QoQ), against analyst consensus of roughly $516.5m — guidance came in above consensus. Non-GAAP gross margin guided to 67.0–69.0%.

What was filed: an 8-K, Item 2.02

A Form 8-K is the event-driven SEC filing — the one a company files when something material happens, rather than on the quarterly calendar. Item 2.02, "Results of Operations and Financial Condition", is the item used to announce results for a completed period. The 8-K cover page is short; the actual numbers live in Exhibit 99.1, the press release attached under Item 9.01.

For a beginner, the practical value of reading the 8-K exhibit rather than a news summary is the tables at the bottom. Credo's release shows three columns — the current quarter, the immediately preceding quarter, and the same quarter a year earlier — plus a full GAAP-to-non-GAAP reconciliation. That reconciliation is where you find out exactly which costs a company is asking you to look past, line by line. Here it is the single most useful part of the document.

Why it matters

The GAAP drop is an acquisition artefact, not a business deterioration. Credo completed its acquisition of DustPhotonics, a silicon-photonics company, in late May 2026, and this is the first full quarter that includes it. Three things follow mechanically: goodwill and intangible assets balloon on the balance sheet, cash falls by the purchase consideration, and a new non-cash amortisation charge starts running through the income statement. That $11.0m intangible amortisation in cost of revenue is the whole of the 3.7-point gap between the 68.2% GAAP gross margin last quarter and 64.5% this quarter. Non-GAAP gross margin, which excludes it, barely moved at 68.0%.

But the reconciliation also shows something real. The gap between GAAP and non-GAAP net income was $109.9m pre-tax, and the largest single component is $88.0m of share-based compensation, not acquisition accounting. Stock compensation is a genuine economic cost — it dilutes existing shareholders — even though it is non-cash. Diluted share count rose from 184.6m a year ago to 194.4m. When a company's non-GAAP profit is nearly double its GAAP profit, that is the number worth understanding rather than skipping past.

Sequential growth is what to track now. The 114.7% year-over-year figure compares against a much smaller company. The more current signal is +9.6% quarter over quarter, with guidance implying about +10.6% next quarter. Non-GAAP EPS grew only 3.4% sequentially, because opex — particularly R&D, up 26% QoQ — is growing faster than revenue right now.

The strategic logic. Credo sells connectivity for AI datacentres: active electrical cables, retimers, optical DSPs and now silicon-photonics components in-house. The DustPhotonics deal moves it further up the optical stack rather than buying those parts from someone else. Management has publicly targeted more than $600m of optical revenue in FY27, weighted to the second half — which is why the second-half guidance ramp is the thing to check against in coming quarters.

Concentration is the structural risk in this business. Connectivity suppliers to AI datacentres typically sell to a small number of very large hyperscale customers. That produces spectacular growth on the way up and considerable volatility if one customer changes its build schedule. Nothing in this filing suggests that has happened; it is simply the shape of the risk in this part of the supply chain.

Beginner takeaway

Credo beat on revenue and on adjusted profit, and guided the current quarter above what analysts expected. The scary-looking fall in reported GAAP earnings per share is almost entirely accounting from an acquisition that closed a quarter ago, not a decline in the underlying business.

The genuinely useful things to carry away: sequential revenue growth is running near 10% a quarter, gross margins are holding around 68% on an adjusted basis, and stock-based compensation is large enough that share count is climbing. A beat plus above-consensus guidance does not always produce a positive market reaction — expectations for high-growth AI suppliers are frequently well ahead of published consensus.

FAQ

Why is Credo's GAAP EPS so much lower than its non-GAAP EPS? Two reasons this quarter: $88.0m of share-based compensation and $11.6m of amortisation on intangibles from the DustPhotonics acquisition, plus $10.4m of deal costs. Non-GAAP excludes all of these. The amortisation is a pure accounting entry; the share compensation is a real cost to existing shareholders even though no cash leaves the company.

What is goodwill, and why did it jump from $93m to $986m? Goodwill is the amount an acquirer pays above the identifiable fair value of what it bought — essentially the premium for the team, technology position and expected synergies. It appears when a deal closes and is not amortised, but it must be tested for impairment; if the acquisition disappoints, goodwill gets written down and that hits reported earnings.

Guidance beat consensus — why might the shares still not react well? Published consensus is an average of analyst models, but the price of a fast-growing stock often reflects expectations well above that average. Beating the formal number while missing the informal "whisper" expectation is a common outcome, which is why a beat and a share-price decline can happen on the same day.

What does Credo actually sell? High-speed connectivity between the parts of an AI datacentre — active electrical cables, optical transceivers, retimers for Ethernet and PCIe, and, since the DustPhotonics deal, silicon-photonics components. As AI clusters grow, moving data between chips and racks quickly and with low power becomes a bottleneck; that bottleneck is Credo's market.

As of 1 September 2026. Source: official SEC filing — read it directly here. We summarise filings for education and may make errors, so always verify against the official document. Educational content only — not investment advice, not a buy/sell recommendation.