View Past Polls
Live Voting

Diamond or Dud?

The weekly showdown where investors decide if controversial stocks are hidden gems or complete trash

⚡ A few minutes a week to sharpen your market knowledge.

Diamond or Dud is for entertainment purposes only. AssetRoom does not provide financial advice. Figures collected prior to poll publication.
💡
Why we're voting on MRVL this week

On Tuesday, Jensen Huang walked onstage at COMPUTEX and called Marvell “the next trillion-dollar company.” The stock surged 32% in a single session - its biggest day ever - and has now more than tripled this year.

Marvell is not a household name, but it is becoming essential plumbing for the AI buildout. The company designs custom AI chips and networking silicon for the same hyperscalers spending $725 billion on infrastructure this year. Amazon taps Marvell to co-design its Trainium processors. Microsoft taps it for Azure Maia accelerators. Google is a customer too. Nvidia liked what it saw enough to invest $2 billion and forge a partnership around silicon photonics.

The Q1 earnings backed it up: record revenue of $2.4 billion (up 28%), a raised full-year outlook to $11.5 billion, and a $16.5 billion target for FY28. But the stock now trades at over 75x forward earnings after a 270% YTD run, and its biggest customers have a history of switching chip suppliers when it suits them. Broadcom just crashed 13% when its AI chip forecast fell short.

MRVL
Marvell Technology
Poll Closes In
d h m s
New poll every Friday
Ytd
~+270%
Q1 Rev
$2.4B (+28% YoY)
Market Cap
~$276B
Forward P/E
~75x

What's your verdict on MRVL?

Vote to reveal what others think

Verify Your Vote

We verify emails to keep votes authentic and prevent spam. One person, one vote.

Already have an account? You'll be redirected to sign in.

Quick email verification - no spam, unsubscribe anytime. Free forever.

💎
Diamond Case
  • Jensen Huang does not hand out trillion-dollar predictions casually. Nvidia invested $2 billion in Marvell and partnered on silicon photonics through NVLink Fusion, integrating Marvell’s networking and custom silicon directly into Nvidia’s AI factory ecosystem. That is not a throwaway endorsement - it is a supply chain commitment.
  • Marvell is the second-largest custom AI chip designer behind Broadcom, and it is gaining ground fast. Data center revenue hit $1.83 billion in Q1 (76% of total), management expects interconnect revenue to grow over 70% this year, and the company raised its FY27 outlook to $11.5 billion with a $16.5 billion FY28 target - implying 40-45% annual growth for the next two years. Record operating cash flow of $639 million in Q1 shows the growth is translating into real cash.
  • The custom silicon market is expanding because hyperscalers want alternatives to Nvidia’s GPUs. Amazon, Google, and Microsoft are all building proprietary chips, and Marvell co-designs them. Every dollar the hyperscalers shift away from off-the-shelf GPUs flows toward companies like Marvell, not away from them.
💩
Dud Case
  • At ~75x forward earnings after a 270% YTD run, the stock is priced for perfection. Marvell’s market cap needs to roughly 3.5x to reach the trillion-dollar mark Huang dangled, and at $276 billion it already trades at a premium to Qualcomm on a fraction of the revenue. One CEO’s conference remark is not a business plan, and Nvidia has every incentive to talk up a company it just invested $2 billion in.
  • Customer concentration is a real vulnerability. A handful of hyperscalers generate the vast majority of Marvell’s data center revenue, and those customers have leverage. Amazon has already brought in competitor Alchip for future Trainium chip generations, and Microsoft is reportedly in talks to shift custom chip work to Broadcom. Losing even one anchor customer would reset the growth story overnight.
  • Broadcom just crashed 13% after its AI chip forecast disappointed, and it holds roughly 60-70% of the custom chip market versus Marvell’s 20-25%. If the dominant player cannot meet expectations, the smaller challenger faces the same risk at a much richer valuation. The AI chip hype cycle has a pattern: stocks run on promises and sell off when the numbers fall short.
💎
Diamond Case
  • Jensen Huang does not hand out trillion-dollar predictions casually. Nvidia invested $2 billion in Marvell and partnered on silicon photonics through NVLink Fusion, integrating Marvell’s networking and custom silicon directly into Nvidia’s AI factory ecosystem. That is not a throwaway endorsement - it is a supply chain commitment.
  • Marvell is the second-largest custom AI chip designer behind Broadcom, and it is gaining ground fast. Data center revenue hit $1.83 billion in Q1 (76% of total), management expects interconnect revenue to grow over 70% this year, and the company raised its FY27 outlook to $11.5 billion with a $16.5 billion FY28 target - implying 40-45% annual growth for the next two years. Record operating cash flow of $639 million in Q1 shows the growth is translating into real cash.
  • The custom silicon market is expanding because hyperscalers want alternatives to Nvidia’s GPUs. Amazon, Google, and Microsoft are all building proprietary chips, and Marvell co-designs them. Every dollar the hyperscalers shift away from off-the-shelf GPUs flows toward companies like Marvell, not away from them.
💩
Dud Case
  • At ~75x forward earnings after a 270% YTD run, the stock is priced for perfection. Marvell’s market cap needs to roughly 3.5x to reach the trillion-dollar mark Huang dangled, and at $276 billion it already trades at a premium to Qualcomm on a fraction of the revenue. One CEO’s conference remark is not a business plan, and Nvidia has every incentive to talk up a company it just invested $2 billion in.
  • Customer concentration is a real vulnerability. A handful of hyperscalers generate the vast majority of Marvell’s data center revenue, and those customers have leverage. Amazon has already brought in competitor Alchip for future Trainium chip generations, and Microsoft is reportedly in talks to shift custom chip work to Broadcom. Losing even one anchor customer would reset the growth story overnight.
  • Broadcom just crashed 13% after its AI chip forecast disappointed, and it holds roughly 60-70% of the custom chip market versus Marvell’s 20-25%. If the dominant player cannot meet expectations, the smaller challenger faces the same risk at a much richer valuation. The AI chip hype cycle has a pattern: stocks run on promises and sell off when the numbers fall short.
Read our AI summaries of MRVL's quarterly and annual reports

Discussion

Be the first to share your opinion!

Sign up or Sign in
💬 Be first to comment

Share your vote. See who agrees.

$MRVL - Diamond or Dud?

💎 ⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜ 0%
💩 ⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜ 0%

https://www.assetroom.net/p/Mzqhzr
Or share on