Credo Technology (NASDAQ:CRDO) has emerged as one of the fastest-growing semiconductor infrastructure companies benefiting from the artificial-intelligence investment boom. Rather than designing GPUs, Credo solves an increasingly important problem created by them: how to move enormous quantities of data quickly, reliably and efficiently between servers, switches, accelerators and racks inside AI data centres.
That puts Credo in the networking and connectivity layer of the AI ecosystem, alongside — and sometimes competing with — companies such as Broadcom (NASDAQ:AVGO), Marvell Technology (NASDAQ:MRVL) and Astera Labs (NASDAQ:ALAB).
Credo Technology investor relations
| Credo Technology (NASDAQ:CRDO) | Price: $212.07 (+48% YTD) | Market cap: $39.55bn |
The investment case is compelling: booming AI infrastructure spending, leadership in Active Electrical Cables (AECs), rising networking speeds and opportunities in optical connectivity. The complication for investors is valuation. At roughly $212 a share and a market capitalisation close to $40 billion, expectations are already extremely demanding.
Credo at a glance
| Metric | Latest/approximate |
| Ticker | CRDO |
| Exchange | Nasdaq |
| Share price | ~$212 |
| Market capitalisation | ~$39.55bn |
| FY2026 revenue | ~$1.34bn |
| Trailing PE | ~85x |
| Core growth market | AI/data-centre connectivity |
| Key products | AECs, retimers, DSPs, optical connectivity, SerDes |
| Major customers | Hyperscalers/cloud and AI infrastructure operators |
| Investment profile | High-growth / high-valuation semiconductor infrastructure |
Market data as of 21 July 2026; valuation multiples move with the share price and earnings estimates.
Credo reported approximately $1.34 billion of revenue in fiscal 2026 and >750% EPS growth, representing an extraordinary step-up in scale as AI infrastructure demand accelerated.
What does Credo actually do?
The easiest way to understand Credo is to think about an AI datacentre as a giant nervous system.
Nvidia (NASDAQ:NVDA) and other accelerators provide the computing brain. Credo helps build the high-speed nervous system connecting everything together.
Modern AI models can run across tens of thousands — potentially hundreds of thousands — of GPUs and accelerators. Those processors constantly exchange enormous volumes of data.
If the connections between them are too slow, unreliable or power hungry, expensive GPUs sit idle waiting for data.
Credo develops high-speed connectivity technologies designed to solve this bottleneck.
Its portfolio includes ZeroFlap Active Electrical Cables, optical transceivers, OmniConnect memory connectivity products, retimers and digital signal processors (DSPs) for Ethernet and PCIe, supported by its PILOT diagnostics and analytics platform.
Where Credo sits in the AI ecosystem
| AI infrastructure layer | Examples | Credo’s role |
| AI accelerators | Nvidia, AMD, custom ASICs | GPUs generate huge connectivity requirements |
| Memory | SK Hynix, Micron, Samsung | High-bandwidth data feeds AI processors |
| Networking silicon | Broadcom, Marvell, Arista | Switches and networking architecture |
| High-speed connectivity | Credo, Astera Labs, Marvell | Moves data efficiently between systems |
| Optical networking | Coherent, Lumentum and others | Longer-distance high-bandwidth links |
| Servers/systems | Super Micro, Dell, HPE | Integrate AI infrastructure |
| Hyperscalers | Microsoft, Amazon, Google, Meta | Major end customers |
Credo says its technology supports connections from server-to-switch, switch-to-switch, server-to-server and rack-to-rack, with solutions reaching speeds as high as 1.6 terabits per second.
The company also says its products are used by four leading US hyperscalers.
That makes Credo a classic AI picks-and-shovels investment: it does not need to predict which AI model ultimately wins if overall AI infrastructure continues requiring dramatically more bandwidth.
The secret weapon: Active Electrical Cables
One of the biggest drivers of Credo’s growth has been Active Electrical Cables, or AECs.
Traditional passive copper cables become increasingly difficult to use as networking speeds rise. Signal degradation limits the distance over which high-speed data can travel reliably.
AECs incorporate electronics into the cable assembly to improve signal integrity.
The result can be:
Longer reach + high bandwidth + lower power than some optical alternatives + lower cost in appropriate applications.
Credo claims to have invented the AEC category and cites third-party 650 Group data indicating an AEC market share of roughly 88%.
That leadership matters because AI clusters are becoming vastly more interconnected.
An AI rack containing dozens of accelerators can require numerous high-speed connections. Multiply that across thousands of racks and the connectivity opportunity becomes enormous.
Why AI could drive a structural AEC boom
The transition looks roughly like this:
More AI accelerators → more data traffic → faster networking → more connectivity complexity → greater demand for advanced cables, retimers and optical solutions.
Networking speeds are simultaneously moving through:
400G → 800G → 1.6T
Every transition makes signal integrity more difficult.
That plays directly into Credo’s expertise.
Opportunity 1: AI spending is shifting towards networking
The first wave of generative-AI spending was overwhelmingly focused on GPUs.
But GPUs cannot operate efficiently without networking.
As AI clusters become larger, networking can represent an increasingly significant proportion of total system cost.
This creates an attractive second-order investment theme.
Instead of asking:
‘How many Nvidia GPUs will hyperscalers buy?’
Credo investors can ask:
‘How much connectivity infrastructure will be required to make all those GPUs work together?’
The answer appears to be: a lot more than previous generations of datacentres required.
Credo specifically highlights AI clusters potentially scaling towards one million GPUs, requiring extremely high-bandwidth, low-latency GPU-to-GPU communication.
Opportunity 2: Credo can expand beyond AECs
It would be a mistake to view Credo purely as a cable company.
Its technology portfolio is expanding across several connectivity markets.
| Technology | Opportunity |
| Active Electrical Cables | Short-reach AI/data-centre connectivity |
| Retimers | Improve signal integrity at high data rates |
| Optical DSPs | Optical transceiver connectivity |
| Optical transceivers | Longer-reach AI networking |
| SerDes/chiplets | High-speed chip-to-chip/data transmission |
| OmniConnect | Memory connectivity |
| PILOT software | Diagnostics and network analytics |
Optical connectivity could become particularly important.
Copper remains attractive where cost and power advantages make sense, but as distances and bandwidth requirements increase, optical technology becomes increasingly necessary.
Credo has therefore been expanding its optical portfolio, including through the acquisition of DustPhotonics and development of optical solutions for AI scale-out networks. This year, optics revenue is estimates at ~$600m, about a quarter of total FY2027 revenue.
If successful, this could significantly increase Credo’s addressable market.
Opportunity 3: the hyperscaler custom-silicon boom
AI infrastructure is also becoming less dependent exclusively on Nvidia GPUs.
Google has TPUs. Amazon has Trainium. Microsoft and Meta are developing custom accelerators.
That diversification could actually benefit connectivity suppliers.
Whatever processor architecture wins, increasingly large clusters still need:
high-speed Ethernet, PCIe connectivity, retimers, cables, DSPs and optical links.
Credo therefore offers investors exposure to AI infrastructure without making a pure bet on one accelerator vendor.
Opportunity 4: extraordinary operating leverage
Semiconductor companies can become highly profitable once revenue scales because much of the initial R&D expense has already been incurred.
Credo’s rapid revenue expansion illustrates this potential.
FY2026 revenue reached roughly $1.3 billion, dramatically increasing the company’s scale.
If revenue continues growing rapidly while gross margins remain strong, earnings can grow considerably faster than sales.
That operating leverage is one reason investors are willing to award CRDO such a premium valuation.
But it cuts both ways.
A slowdown in revenue growth could cause earnings expectations — and therefore the valuation multiple — to compress quickly.
The biggest risk: valuation
Credo’s biggest problem may simply be that investors already recognise how attractive the story is.
At around $212, CRDO trades on a trailing PE of roughly 85x earnings, based on Stockopedia data.
Yet traditional valuation metrics like PEs are limited for a company experiencing explosive earnings growth – for example, the 12-month rolling forward PE falls to ~32x on current FY2027 estimates, although even that multiple demonstrates the expectations already embedded in the shares.
Indicative peer valuation comparison
| Company | Primary exposure | Relative valuation profile | Growth expectations |
| Credo | AI connectivity/AECs | High | Very high |
| Astera Labs | AI connectivity | Extremely high | Very high |
| Marvell | AI/custom silicon/networking | High | High |
| Broadcom | AI networking/custom silicon | Lower | High |
| Nvidia | AI accelerators/networking | <20x | High |
Peer comparisons need care because the businesses differ substantially.
Broadcom, for example, is vastly larger and more diversified, while Astera Labs is closer to Credo in being a concentrated high-growth AI connectivity play.
July 2026 market data shows substantial forward sales premiums across the group, including roughly 23.8x forward sales for Broadcom, 19.6x for Marvell and 52.9x for Astera Labs at one snapshot. These multiples can change rapidly as share prices and forecasts move.
The key issue isn’t whether Credo deserves a premium.
It almost certainly does if current growth persists.
The question is how large that premium should be.
What the bull case assumes
A bullish valuation effectively assumes Credo can evolve from a niche semiconductor company into a major strategic supplier to global AI infrastructure.
The argument rests on several assumptions:
- AI infrastructure spending remains elevated for years.
- AEC adoption continues expanding.
- Credo retains technological and market-share leadership.
- Revenue becomes more diversified across hyperscalers.
- Optical products become another meaningful growth engine.
- 800G and 1.6T networking create larger dollar content per AI cluster.
- Earnings scale much faster than operating expenses.
If those assumptions hold, today’s apparently extreme valuation could gradually be justified by earnings growth.
This is similar to what has happened historically with exceptional semiconductor growth companies: the denominator — earnings — catches up with the share price.
But there is very little room for execution mistakes.
Risk 1: customer concentration
This is probably the most important fundamental risk after valuation.
Credo’s top 10 customers accounted for roughly 90% of FY2026 revenue, with two customers individually representing at least 10%.
That creates considerable concentration risk.
Imagine one hyperscaler represents a major portion of revenue and then:
changes architecture, adopts an alternative supplier, develops technology internally or delays an AI data-centre deployment.
Credo’s growth rate could change dramatically.
Customer diversification therefore deserves close attention in every earnings report.
Risk 2: competition from giants
Credo operates in markets populated by extremely capable competitors.
These include:
Broadcom, Marvell, Astera Labs and numerous optical/networking semiconductor specialists.
Broadcom and Marvell have enormous R&D budgets and deep hyperscaler relationships.
Meanwhile, large customers increasingly design custom silicon themselves.
Credo’s advantage comes from specialised connectivity expertise, execution and power-efficient architectures.
But technology leadership is never permanent in semiconductors.
Risk 3: copper versus optics
Credo’s AEC leadership is extremely valuable today.
Longer term, however, AI architectures could increasingly shift towards optical connectivity.
That does not necessarily spell trouble — Credo is deliberately expanding into optical DSPs and transceivers.
Its optical DSPs already target AI clusters, hyperscale data centres and networking infrastructure, supporting connections extending from metres to kilometres.
But the transition creates both an opportunity and a threat.
Credo must ensure it participates in the move towards optics rather than allowing optical technology to cannibalise its strongest franchise.
Risk 4: AI capital expenditure eventually slows
The entire investment thesis ultimately depends on sustained infrastructure spending.
Microsoft, Meta, Amazon and Alphabet are investing extraordinary sums in AI infrastructure.
If AI monetisation disappoints, hyperscalers could moderate capital expenditure.
That would ripple through the supply chain:
Hyperscaler capex ↓ → fewer AI servers ↓ → fewer switches ↓ → fewer high-speed connections ↓ → weaker Credo demand.
A stock trading at a very high multiple can react violently to even modest changes in growth expectations.
What Wall Street thinks
Analyst sentiment remains broadly constructive, reflecting Credo’s unusually strong exposure to AI networking.
One notable recent call came from Evercore ISI, which initiated coverage in June with an Outperform rating and a $325 price target.
Rothschild & Co Redburn initiated coverage earlier with a Buy rating and $206 target. At today’s roughly $212 share price, that particular target no longer implies upside — a useful reminder that analyst recommendations and valuation can quickly diverge when a stock rallies.
Selected analyst target spectrum
| Analyst/broker | Rating | Price target | Read-through |
| Evercore ISI | Outperform | $325 | Strong AI connectivity growth case |
| Rothschild & Co Redburn | Buy | $206 | Positive fundamentals, but valuation has caught up/overtaken target |
| Broader Street view | Generally positive | Wide range | Growth optimism offset by valuation risk |
The range itself is informative.
At $325, the bull case implies Credo can continue compounding at exceptional rates and capture a larger share of AI connectivity spending.
Targets closer to or below the current share price imply that much of this success is already reflected in CRDO’s valuation.
Positive analyst argument: Credo has become strategically important
The bullish Wall Street thesis centres on Credo’s position at the intersection of several powerful trends:
AI compute growth + higher network speeds + increasing connectivity density + power efficiency.
Evercore’s $325 target illustrates how optimistic analysts can become if Credo maintains its growth trajectory.
The strongest argument is that networking requirements may grow faster than the number of accelerators themselves as clusters become increasingly interconnected.
Credo doesn’t simply benefit from more AI servers.
It potentially benefits from more connections per server and higher value per connection.
That is a powerful combination.
The more cautious analyst argument
The bearish case doesn’t necessarily require believing Credo is a bad company.
Quite the opposite.
The central concern is:
Can even an excellent company outperform expectations that are already extremely high?
With the shares valued at more than 100 times trailing earnings, investors are paying today for substantial future growth.
Potential disappointment could come from surprisingly small things:
- Revenue growth slowing from exceptional to merely strong.
- A hyperscaler delaying deployment.
- Gross margins declining.
- Increased competition.
- Customer concentration remaining high.
- Faster migration towards alternative connectivity architectures.
- AI capex expectations being revised down.
The Redburn $206 target compared with a current share price around $212 demonstrates this valuation tension: an analyst can remain fundamentally positive while the stock price itself catches up with or surpasses their valuation framework.
Bull case versus bear case
| Bull case | Bear case |
| AI infrastructure spending remains structurally high | AI capex eventually normalises |
| Dominant position in AECs | Competition erodes market share |
| 800G/1.6T increases connectivity value | Technology transitions faster than expected |
| Optical creates a major second growth engine | Optics cannibalises copper faster than Credo captures it |
| More hyperscalers adopt Credo | Customer concentration remains excessive |
| Huge operating leverage | Margins disappoint as product mix changes |
| Earnings grow into valuation | Multiple compression overwhelms earnings growth |
What UK investors should watch next
There are five metrics that arguably matter more than headline EPS.
Revenue growth and guidance come first. At this valuation, the market needs continued evidence of exceptional expansion.
Second is hyperscaler diversification. Investors should watch whether Credo can reduce dependence on its largest customers while adding new large-scale deployments.
Third is AEC market share and adoption. Credo’s leadership is a major competitive asset, but investors need evidence that AEC demand continues growing as 800G and 1.6T architectures develop.
Fourth is optical revenue. Successful expansion into optical connectivity could materially increase the addressable market and reduce the risk of technological substitution.
Finally, watch gross margins. Strong margins would demonstrate pricing power and differentiation; deterioration could signal competition or an unfavourable product mix.
Investor verdict
Credo may be one of the most interesting second-order AI investments available to UK investors having emerged as one of the fastest-growing semiconductor infrastructure companies.
The attraction is straightforward.
Rather than competing directly in GPUs, Credo supplies technology needed to connect increasingly enormous AI clusters.
Its combination of AEC leadership, high-speed SerDes expertise, retimers, optical DSPs and expanding optical capabilities places it squarely in one of the most important bottlenecks facing AI infrastructure: moving data quickly without consuming excessive power.
The company’s reported AEC leadership and relationships across leading hyperscalers suggest it has built genuine competitive advantages rather than simply benefiting from AI hype.
But price matters.
At around a $40 billion valuation and more than 100 times trailing earnings, CRDO is priced more like an emerging AI infrastructure champion than an unrecognised semiconductor opportunity.
That creates an unusual risk/reward profile:
Business quality: ★★★★★
Structural growth opportunity: ★★★★★
Competitive positioning: ★★★★☆
Customer diversification: ★★☆☆☆
Valuation attractiveness: ★★☆☆☆
Overall risk: High
For UK retail investors who already own Nvidia or other mega-cap AI stocks, Credo could provide differentiated exposure to the AI networking and connectivity bottleneck.
However, investors should expect substantial volatility. The risk isn’t necessarily that Credo’s business fails. It is that a very good business delivers results that aren’t quite good enough to satisfy an exceptionally high valuation.
Indirect exposure
For investors unwilling to take single-stock risk, there is another route worth considering: UK-accessible active funds holding CRDO, including BlackRock Next Generation Technology, AXA Framlington Global Technology and Liontrust Global Innovation, while Polar Capital Technology Trust (LON:PCT) provides a UK-listed investment-trust route to broader AI infrastructure exposure.
AI leadership drives another standout year for popular Polar Capital Technology Trust
Bottom line: Credo offers unusually pure exposure to the infrastructure required to connect the AI revolution. The fundamental story remains powerful, but after its spectacular re-rating, future returns increasingly depend not simply on strong growth, but on Credo repeatedly exceeding already elevated expectations. For growth-oriented UK investors, it is a compelling company to watch — particularly if volatility creates a more attractive entry valuation.
Disclaimer: The author Steven Frazer has a personal interest in Polar Capital Technology Trust.
You might also like:
