Credo Technology: Optical Revenue Set to Exceed $600 Million in Fiscal 2027 as Dust Photonics Acquisition Closes
Q4 Fiscal 2026 Earnings Call, June 1, 2026
Credo Technology delivered another record quarter to close out fiscal 2026, but the more consequential story from Monday's earnings call was the company's aggressive optical revenue forecast for fiscal 2027 and the freshly closed acquisition of Dust Photonics. Together, these developments represent a meaningful strategic pivot for a company that built its reputation on copper connectivity. The question investors are now asking is whether Credo can execute a rapid optical ramp while sustaining the margins and operational discipline that defined its copper business.
Fiscal 2026: Revenue Triples, Earnings Scale Even Faster
Full-year fiscal 2026 revenue came in at $1.3 billion, up 206% year-over-year, while non-GAAP net income increased more than 5x to $662 million, producing earnings per share of $3.46, a 392% improvement over the prior year. The fourth quarter alone generated $437 million in revenue, exceeding Credo's entire fiscal 2025 revenue in a single quarter. Q4 non-GAAP gross margin was 68.3%, above the high end of guidance, and non-GAAP net margin reached 51.9%. Free cash flow for the quarter was $177.5 million, bringing the year-end cash balance to $1.4 billion. These are not incremental improvements. As CFO Dan Fleming noted, "Q4 non-GAAP net income more than tripled year-over-year and was 33% higher than our revenue in the fourth quarter of last year," which speaks to the compounding leverage in the business model.
The Optical Bet: $600 Million Target Across Three Legs
The central new disclosure from this call is Credo's formal fiscal 2027 optical revenue forecast. CEO Bill Brennan stated that the company expects optical DSPs, silicon photonics PICs, and ZeroFlap Optics to each contribute more than $100 million in revenue during fiscal 2027, with the combined optical portfolio exceeding $600 million for the year. That figure would represent roughly a quarter of what the company expects to be an 80%-plus year-over-year total revenue growth year. Fleming clarified the composition of growth: "About half of that growth in absolute dollars is coming from our optical portfolio and about half of that is coming from our existing copper portfolio, predominantly AECs, but also retimers."
Brennan was candid about which optical segment will dominate. ZeroFlap Optics, which carry three-digit ASPs compared to two-digit ASPs for discrete DSPs and silicon photonics PICs, will be "clearly our largest revenue contributor for our optical portfolio." The implication is that while all three segments cross the $100 million threshold, ZeroFlap Optics likely accounts for well over half of the $600 million total. Importantly, the ramp is heavily back-half weighted. The company guided for mid-single-digit sequential revenue growth in Q1 and expects a similar profile in Q2, with the optical inflection arriving in the second half of the fiscal year. First-quarter guidance is $465 million to $475 million.
Dust Photonics Closes: Silicon Photonics Now In-House
The acquisition of Dust Photonics closed last week, and the strategic rationale is significant. Dust brings a silicon photonics PIC portfolio spanning 800G and 1.6T solutions with a roadmap to 3.2 terabits per second. Critically, Dust's architecture reduces laser count in optical designs, which Brennan described as solving both a reliability and a supply chain problem simultaneously: "Laser count reduction can ease the industry supply chain limitations." The acquisition also provides a direct path to co-packaged optics and near-package optics architectures, with initial CPO and NPO revenue now expected in fiscal 2028.
The financial impact is immediate and material. Fleming confirmed that net cash paid for Dust Photonics was approximately $750 million in Q1, drawn against the $1.4 billion cash balance at quarter-end. With cash flow from operations approaching $200 million per quarter, Credo described its post-acquisition cash position as "very comfortable," and management stated there are no current plans to raise additional capital or conduct share buybacks. The balance sheet can absorb further opportunistic acquisitions, though Fleming indicated nothing is imminent.
Supply Chain: 3-Nanometer Is the Constraint Everyone Is Watching
Brennan addressed the industry-wide concern around 3-nanometer wafer availability directly. All of Credo's 200-gigabits-per-lane products across the portfolio are built on 3-nanometer, and the CEO acknowledged broad capacity tightness while expressing confidence in Credo's position: "I feel quite good. Over the last 5 years, we've invested heavily and not just in resources, but also dollars in making sure our supply chain partners understand our commitment." He also noted that for the 1.6T market, "the bulk of the production will happen in 3-nanometer and possibly a shift to 2 over time." Credo's process node diversity, spanning 12-nanometer for high-volume 100G AECs, 7-nanometer for optical DSPs, 5-nanometer for a program in flight, and 3-nanometer for 200G products, provides some insulation against single-node concentration risk. For ZeroFlap Optics, Brennan described owning the full bill of materials and securing capacity commitments through direct financial commitments to supply chain partners, and expressed confidence in "a very aggressive ramp in the second half of our fiscal year and even more than double or triple that the following year."
1.6T Transition: Fiscal 2027 Contribution Will Be Light
One notable piece of candor from Brennan: fiscal 2027 will see "relatively light revenue" from 200-gigabits-per-lane products. The industry has not yet fully committed to the 1.6T transition timeline, and Credo's 80%-plus growth forecast for the year rests predominantly on 100G-per-lane products. Brennan described a hedge that some customers are employing: building systems with two physical ports per GPU, allowing deployment of either 4 lanes of 200G signaling or 8 lanes of 100G, preserving upgrade flexibility. "Not everything has to come at once," he said. When 1.6T does ramp, Credo expects a meaningful ASP uplift across its copper and optical portfolio, but investors should not model that into fiscal 2027 numbers in any meaningful way.
Customer Concentration and Neo Cloud Diversification
The four largest customers each contributed at least 10% of Q4 revenue, with the top customer at 34%, the second at 27%, and the third at 16%. The fourth customer at 10% was new to that threshold, signaling broadening. Brennan disclosed that Credo is now deployed with five of six hyperscalers at high volume, with xAI identified as the one customer where the company considers itself "fully deployed." Across the other four, further penetration opportunities remain. On the Neo cloud front, Brennan put a number on the opportunity: "If we look at that group of customers collectively, I definitely think it could be on the order of 20%" of total revenue over the coming years. He characterized Neo cloud operators as "maybe the perfect customer for Credo" given their emphasis on network performance, reliability, and deployment speed.
OmniConnect and ALC: Fiscal 2028 Ramps With High Per-Unit Economics
Two emerging product categories, the Weaver gearbox under the OmniConnect family and Active LED Cables, are both expected to enter production ramp in fiscal 2028. Brennan provided the clearest articulation yet of Weaver's revenue potential, using customer Positron as a reference point. Positron's inference engine design uses 2 terabytes of LPDDR memory, more than 10x any other announced inference product, requiring many Weaver chips per system. Brennan reiterated prior guidance that "revenue contribution per GPU can be between $2,000 and $3,000," suggesting rapid revenue scaling once inference GPU customers reach volume. For ALC, Brennan drew a direct analogy to the AEC ramp: "The dynamic there could be very much similar to AECs and ZF Optics in the sense that we can see large revenue quickly." ALC extends AEC-class reliability to 30-meter distances using MicroLED technology, addressing a row-scale connectivity gap that current AECs cannot reach.
Margin Outlook and Operating Expense Growth
Management guided for fiscal 2027 non-GAAP gross margins "broadly consistent" with fiscal 2026 levels, meaning around 68%, and for non-GAAP net margins "in the vicinity of 50%." Operating expenses are expected to grow approximately 50% year-over-year, a meaningful absolute increase but well below the 80%-plus revenue growth rate, preserving earnings leverage. The Q1 non-GAAP gross margin guidance range of 67% to 69% is consistent with recent quarters, and the company acknowledged that tariff policy remains a fluid variable in its planning assumptions.
Credo Technology Group Holding Ltd Deep Dive
The Engine of the AI Cluster: Business Model and Economics
In the architecture of modern artificial intelligence infrastructure, compute and memory receive the bulk of market attention, yet the physical connectivity between compute nodes dictates the ultimate efficiency of the entire cluster. Credo Technology Group operates as a pure-play provider of high-speed connectivity solutions, effectively engineering the nervous system for AI data centers. The company monetizes its intellectual property through three primary vectors: Active Electrical Cables, Optical Digital Signal Processors, and the licensing of its proprietary Serializer and Deserializer IP. Credo's flagship product line is the Active Electrical Cable, a copper cable embedded with specialized digital signal processors that regenerates signals over short distances. Unlike competitors who merely design the silicon and rely on third parties for assembly, Credo operates a vertically integrated model where it develops the silicon, oversees the tape-out, and delivers the finalized cable to the customer. This full-stack approach allows the company to capture the entire margin profile of the product. The financial validation of this model is striking. For the fiscal year 2026, which ended in May, the company recorded total revenue of $1.3 billion, representing a year-over-year increase of over 200%. The leverage inherent in Credo's business model is evidenced by its margin structure; in the fourth quarter of fiscal 2026, the company reported a non-GAAP gross margin of 68.3% and a non-GAAP operating margin of 49.6%, yielding a full-year non-GAAP net income of $662 million. These metrics reflect a highly specialized semiconductor firm operating with structural pricing power in a supply-constrained environment.
The Hyperscale Concentration: Key Customers and Suppliers
Credo's revenue base is characterized by extreme customer concentration, a structural reality of supplying the hyperscale cloud market. The demand for next-generation networking components is overwhelmingly driven by a handful of entities deploying capital at unprecedented scales. In the fourth quarter of fiscal 2026, four hyperscale customers accounted for 87% of Credo's total revenue, with individual contributions at 34%, 27%, 16%, and 10%. Industry supply chain dynamics indicate that this cohort is primarily comprised of Microsoft, Amazon Web Services, Google, and Meta. This concentration presents a binary dynamic for the business profile. On one hand, securing design wins with these tier-one operators enables immediate and massive volume ramps, bypassing the fragmented enterprise market. On the other hand, it exposes the company to severe volatility. If a single customer delays a cluster deployment, alters its capital expenditure forecast, or successfully develops an in-house connectivity solution to bypass merchant silicon, the impact on Credo's top line would be immediate and material. On the supplier side, Credo operates a fabless manufacturing model, relying heavily on foundry partners such as Taiwan Semiconductor Manufacturing Company for advanced node production. This reliance subjects Credo to the broader geopolitical and capacity risks inherent in the global semiconductor supply chain.
Trench Warfare in the AI Rack: Competitors and Market Share
The high-speed connectivity market is dominated by incumbent giants and aggressive pure-play challengers, creating an intensely competitive landscape. Credo's primary large-cap rivals are Marvell Technology and Broadcom, two highly diversified semiconductor entities that possess vast intellectual property portfolios and command over 80% of the broader custom silicon and optical digital signal processor markets. Meanwhile, in the adjacent connectivity space, Astera Labs has established itself as the dominant force in PCIe retimers and Compute Express Link technologies, holding an estimated 55% market share in AI accelerator connectivity. Despite the presence of these well-capitalized competitors, Credo has carved out a decisive monopoly in the merchant Active Electrical Cable market, where analysts estimate its share sits between 70% and 85%. While Marvell and Broadcom supply digital signal processors to third-party cable assemblers, Credo's vertically integrated delivery of the complete Active Electrical Cable has resonated with hyperscalers seeking guaranteed signal integrity and single-vendor accountability. However, as the market transitions toward higher bandwidth architectures, the competitive overlap between Credo, Astera Labs, and the legacy incumbents is increasing, setting the stage for aggressive market share defense across all connectivity protocols.
The Silicon Moat: Analyzing Credo's Competitive Advantage
Credo's competitive moat is constructed upon a foundation of deep mixed-signal engineering expertise, specifically within its proprietary Serializer and Deserializer architecture. Designing high-speed interconnects that can seamlessly translate electrical and optical signals requires highly specialized, difficult-to-replicate engineering talent. The primary advantage of Credo's Active Electrical Cables lies in their power efficiency and operational reliability compared to traditional optical transceivers. Within a server rack, where transmission distances are typically under seven meters, Credo's copper-based cables consume 30% to 50% less power than optical alternatives. In data centers where power constraints are the primary bottleneck for AI cluster expansion, this thermal and energy efficiency is a critical differentiator. Furthermore, the company has heavily marketed its ZeroFlap technology, which prevents the intermittent dropping and reconnecting of network links. In AI training environments, a single dropped link can halt an entire training run, costing hyperscalers millions of dollars in idle compute time. The integration of ZeroFlap into its cables provides a tangible, high-value reliability metric that entrenches Credo's products deep within the architecture of major cloud providers.
The Threat of Physics: Industry Dynamics and Disruptive Entrants
The central threat to Credo's core business model is the fundamental limitation of physics. As data center speeds scale from 800 gigabits per second to 1.6 terabits and eventually 3.2 terabits, the signal degradation inherent in copper connections becomes exponentially more difficult to mitigate. The copper versus optics debate defines the structural future of the connectivity market. If copper reaches a physical wall where no amount of digital signal processing can clean the signal without consuming massive amounts of power, the industry will pivot entirely to optics, stranding traditional Active Electrical Cable technologies. This transition is being accelerated by a new class of disruptive entrants focused on silicon photonics and co-packaged optics. Companies such as Celestial AI and Ayar Labs are developing photonic fabrics and optical input/output chiplets that enable high-bandwidth communication between AI accelerators directly via light. These architectures bypass traditional electrical Serializer and Deserializer interfaces altogether, significantly reducing latency and power bottlenecks in large-scale AI training. If hyperscalers adopt co-packaged optics at a fleet scale, the demand for conventional copper interconnects and standalone digital signal processors could face terminal decline.
Pivoting to Optics: New Products as Growth Drivers
Recognizing the existential threat posed by the optical transition, Credo has aggressively expanded its total addressable market beyond copper to secure its future relevance. In late May 2026, the company completed the acquisition of DustPhotonics for $750 million, securing proprietary Silicon Photonics Photonic Integrated Circuit technology. This followed the earlier acquisition of CoMira Solutions, which brought specialized link-layer and error correction intellectual property in-house. Through these transactions, Credo has assembled an end-to-end optical connectivity stack. The company is actively commercializing its new optical digital signal processor families, such as Cardinal and Robin, which are tailored for next-generation 3-nanometer AI compute fabrics. Furthermore, management expects to ramp revenues from ZeroFlap optics and OmniConnect gearboxes designed for Compute Express Link protocols. This strategic pivot transforms Credo from a copper pure-play into a comprehensive connectivity provider. Management has signaled strong confidence in this transition, projecting that the newly integrated optical portfolio will generate over $600 million in revenue during fiscal year 2027, effectively hedging the business against the physical limits of its legacy copper products.
Execution at Hyperscale: Management Track Record
The executive team, led by Chief Executive Officer Bill Brennan, has demonstrated exceptional operational execution and strategic foresight over the past several years. Management successfully identified the Active Electrical Cable bottleneck long before it became a consensus issue, positioning the company as the default supplier during the initial wave of AI infrastructure buildouts. The financial translation of this strategy has been flawless. Scaling a semiconductor firm from niche revenues to $1.3 billion in annual sales while simultaneously expanding gross margins to 68.3% is a rare achievement in the hardware sector. Beyond operational scaling, management has proven its willingness to make decisive capital allocation choices. The $750 million acquisition of DustPhotonics illustrates an executive team that is intellectually honest about the limitations of its core copper technology. Rather than rigidly defending a legacy architecture, management utilized its highly valued equity and robust balance sheet to buy its way into the silicon photonics market. This proactive cannibalization and portfolio expansion reflects a mature, clear-eyed leadership team capable of navigating complex hardware transitions without losing market momentum.
The Scorecard
Credo Technology Group stands as a highly profitable, strategically vital component of the global artificial intelligence infrastructure buildout. The company's dominance in the Active Electrical Cable market, underpinned by proprietary mixed-signal engineering and vertical integration, has yielded a structural moat reflected in 68.3% gross margins and near 50% operating margins. By securing deep integrations with the world's largest hyperscalers, the company has successfully capitalized on the initial waves of the AI super-cycle, delivering a staggering 200% year-over-year revenue growth in fiscal 2026. The financial profile is unassailable, characterized by robust free cash flow generation and a pristine balance sheet that affords strategic flexibility.
However, the risk profile requires clinical evaluation. The extreme customer concentration, with four entities comprising 87% of recent quarterly revenue, leaves the company vulnerable to localized capital expenditure shocks or insourcing decisions. Furthermore, the physical limitations of copper and the rapid advancement of disruptive co-packaged optical technologies by well-capitalized entrants present a persistent terminal threat to the legacy product line. Management's recent aggressive acquisitions in the silicon photonics space successfully pivot the company to address these challenges, setting up an estimated $600 million optical revenue pipeline for fiscal 2027. The ultimate success of the equity will depend on Credo's ability to seamlessly integrate these acquisitions and defend its connectivity dominance against much larger incumbents as the industry crosses the 1.6-terabit threshold.