Is Marvel Cheap After Earnings? The Numbers Say It’s Complicated.

Marvell Technology printed a record quarter, guided aggressively higher, and Wall Street analysts immediately started bumping price targets – Wells Fargo to $240, Jefferies to $235, KeyBanc to $260, Benchmark all the way to $275. And the stock is still down premarket. If that sounds counterintuitive, welcome to AI-era chip investing, where good is never quite good enough and the bar keeps moving.

What Actually Happened

Marvell’s fiscal Q1 FY2027 results (quarter ended May 2, 2026) came in at $2.418 billion in revenue – 28% year-over-year growth, 9% sequential, and a new record. Non-GAAP EPS landed at $0.80, beating the $0.75 analyst consensus by roughly 6.7%. Operating cash flow hit a record $638.8 million. Data center revenue specifically came in at $1.833 billion, growing 27% year-over-year, and management flagged “exceptional AI-related bookings” across interconnect, switching, and custom silicon.

Then the guidance. Q2 FY2027 revenue pegged at $2.7 billion – about 35% year-over-year at the midpoint, beating the analyst consensus of $2.635 billion. Non-GAAP EPS guided to $0.93 at midpoint. Full-year FY2027 revenue raised to nearly $11.5 billion (roughly 40% growth). FY2028 guidance lifted to $16.5 billion – $1.5 billion above what management told the Street just last quarter.

Slight tangent worth noting: management confirmed line of sight to over $10 billion in annual custom silicon revenue by fiscal 2029. That’s not a rounding error. That’s a business that barely existed three years ago becoming the core of the revenue model.

So Why Is It Down?

Here’s the thing. The stock has already done the heavy lifting. Over the past year, MRVL is up roughly 208%. It tripled. The market cap sits near $174 billion. The trailing P/E is around 64x, and the forward multiple – even on aggressive FY2027 estimates – is still demanding. The interconnect business alone is being modeled to grow more than 70% this fiscal year, revised upward from prior guidance of 50% growth. That’s the kind of acceleration that justifies a premium. The question is how much premium is already baked in.

There’s also a structural pattern worth respecting: Marvell has posted negative stock moves on earnings day in four of its last four releases – including drops on beats. The market punishes anything short of blowout guidance, and the sell-side consensus heading into this print was already extremely elevated. Oppenheimer was at $200 before the number dropped. Bank of America’s Vivek Arya had already bumped to $200 calling it a top pick, citing AI networking TAM expansion and NVIDIA design wins. The upside was priced before the report hit.

Add to that: insider selling. Over the past three months, insiders have sold roughly $29.9 million in shares. Not a fire alarm, but not nothing either.

The Business in Plain English

Marvell is a fabless chip designer. It doesn’t fabricate its own silicon – it designs it and outsources production. The core of what it sells today is data infrastructure semiconductors: custom ASICs (application-specific chips built for hyperscaler AI workloads), high-speed optical networking components, PCIe interconnect, Ethernet switches, and coherent DSP modules. The company recently acquired Celestial AI and XConn Technologies, both photonics-focused, to deepen its optical networking capabilities. It also recently acquired Polariton Technologies, which brings silicon photonics expertise. The expanded NVIDIA partnership – embedding Marvell silicon into NVIDIA’s ecosystem – is arguably the most significant development for long-term revenue visibility.

Data center is now roughly 76% of total revenue. The non-data-center business (enterprise, carrier, automotive) is not the story right now.

Is It Cheap?

Honestly, no – not on any traditional metric. The trailing P/E is in the mid-60s. The forward multiple on FY2027 consensus is still elevated. The stock is now trading above the average analyst price target that existed before this earnings print. Post-earnings target hikes are flying in, with the Street rapidly repricing toward $220–$260, but the stock has to keep executing into those numbers to justify the gap.

Where it gets interesting is the growth trajectory. If Marvell actually delivers $16.5 billion in FY2028 revenue and continues expanding non-GAAP operating margins toward 35%+, the earnings power two years out looks very different from what current multiples suggest. The $3 billion quarterly revenue milestone – which management says arrives in Q3, a full quarter ahead of prior expectations – is the near-term confirmation test.

Bull / Base / Bear

• Bull: Marvell hits $3B in Q3 as guided, FY28 $16.5B guidance proves conservative, custom silicon ramp to $10B+ in FY29 becomes the new floor. Stock re-rates toward 30x FY28 earnings – significant upside from current levels.
• Base: Execution is solid but acquisition integration (Celestial AI, XConn) creates some turbulence, gross margins stay compressed around 58–59% non-GAAP, and the stock consolidates in a range while earnings catch up to valuation over 18 months.
• Bear: Hyperscaler AI capex growth slows or gets rerouted, a competitor wins a major custom ASIC program, and the stretched multiple compresses fast. The stock has already shown it can drop 20% in a session – the downside is not small.

Cheap Investor Scorecard

• Q1 Revenue Beat: $2.418B vs. $2.40B guide midpoint – check
• Non-GAAP EPS Beat: $0.80 vs. $0.75 consensus – check
• Q2 Revenue Guide: $2.7B vs. $2.635B consensus – beat
• FY28 Revenue Guidance: Raised to $16.5B (+$1.5B vs. prior) – significant
• Operating Cash Flow: Record $638.8M – check
• Non-GAAP Gross Margin: 58.9% – slight compression, watch closely
• Net Debt/EBITDA: 0.32x – balance sheet not a concern
• Insider Activity: $29.9M in sales last 3 months – yellow flag
• Custom Silicon FY29 Target: $10B+ confirmed – long-term anchor
• Valuation: Trailing 64x P/E, forward still elevated – not cheap on paper

The bottom line: Marvell is not cheap. It was never going to be cheap after tripling. What it is, is a business that just raised its long-term revenue outlook by $1.5 billion in a single quarter, confirmed a $10 billion custom silicon runway into FY2029, and landed a deeper partnership with NVIDIA. The premarket dip is the market asking whether all of that is already in the price. It probably is – partially. Whether the next leg is up or sideways depends on whether Q3’s $3 billion quarter actually lands on time.

That’s the only number that matters right now. Watch it.