By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel consensus is that while Marvell's partnership with Google presents opportunities, the execution risk is substantial, with key concerns being the timing and scale of Google's capex ramp, potential margin compression, and the risk of Google using Marvell as negotiating leverage against Broadcom.
Risk: The single biggest risk flagged is the potential margin compression before revenue scales, which could pressure Marvell's own margins if Google succeeds in forcing Broadcom to lower prices.
Opportunity: The single biggest opportunity flagged is the potential for Marvell to gain a significant revenue stream and equity upside through its partnership with Google, if it can successfully execute on its custom silicon ambition.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Marvell Technology, Inc. (NASDAQ:MRVL) has expanded its partnership with Google to develop custom chips for AI infrastructure, strengthening its position in the growing custom-silicon market. The agreement covers several areas linked to Google's Tensor Processing Unit ecosystem, including AI inference accelerators, storage and networking products, memory interface controllers, and near-memory computing.
Alongside the commercial agreement, Marvell issued Google a warrant to buy up to 58.97 million Marvell shares at $206.58 each. If fully exercised, the warrant would be worth about $12.18 billion. The headline figure is important, but it needs some context because Google is not committing $12.18 billion upfront. Most of the warrants will vest only as Marvell generates revenue from Google over the coming years.
Investors reacted strongly to the news, with Marvell Technology, Inc. (NASDAQ:MRVL) shares gaining more than 11% in premarket trading. Broadcom, Google's existing custom-chip partner, fell more than 2%.
Photo from Marvell website
Bull Case: A Long-Term Growth Opportunity
The most important part of the deal may be the way the warrant is structured. Most of the shares are linked to actual revenue from Google rather than simply vesting over time. That gives investors a fairly clear way to track how the relationship develops.
Every $500 million in qualifying revenue allows another tranche to vest. If Marvell continues winning business from Google, more of the warrant becomes valuable. If the revenue does not materialize, much of the warrant may never vest. In that sense, the agreement creates a direct link between Google's spending on Marvell's products and the potential value of the warrant.
The deal also gives Marvell a larger role in a market that is becoming increasingly important to major cloud companies. Google, Amazon, and other hyperscalers are investing heavily in custom chips as they look for ways to build AI infrastructure that is better suited to their own workloads and less dependent on general-purpose GPUs.
Google's continued investment in its TPU ecosystem fits directly into this trend. The fact that Marvell will work on areas such as memory controllers and near-memory computing suggests the relationship goes beyond a single chip program and could develop into a broader role within Google's AI infrastructure.
There is also a useful precedent. In December 2024, Marvell reached a similar warrant agreement with Amazon Web Services covering 4.18 million shares at an exercise price of $87.78. That warrant was also linked to revenue generated from AWS and remains tied to performance through January 2030.
The Google agreement is much larger. The potential share count is more than 14 times the size of the AWS warrant, while the exercise price is also substantially higher. That difference suggests the potential scale of the Google relationship is considerably greater.
Marvell Technology, Inc. (NASDAQ:MRVL) was already positioned for strong growth in AI infrastructure before this announcement. The company has previously targeted more than $10 billion in custom-chip revenue by fiscal 2029 and expects data-center revenue to grow by roughly 50% this year. If those targets are achieved, the Google agreement could strengthen an existing growth story rather than create an entirely new one.
The deal could also improve Marvell's position against Broadcom. Broadcom already has a long-term relationship with Google covering custom TPU development and networking through 2031. Google's decision to expand its work with Marvell alongside Broadcom suggests that the company is willing to use multiple suppliers as its AI infrastructure requirements grow. For Marvell, gaining a larger share of Google's spending could become a meaningful source of revenue over several years.
The $12.18 Billion Figure Can Be Misleading
The biggest risk is that investors may focus too heavily on the $12.18 billion headline figure. That amount represents the maximum potential value of the warrant, not a guaranteed investment by Google.A
Most of the shares will only vest when Marvell Technology, Inc. (NASDAQ:MRVL) generates enough qualifying revenue. If Google's purchases remain below the required thresholds, a significant portion of the warrant could expire without being exercised.
Competition is another concern. Broadcom already has a significant relationship with Google and is involved in Google's custom TPU development and AI networking. Marvell still has to prove that it can win additional business on the basis of technology, pricing, manufacturing capacity, and execution.
Winning a custom-chip contract is also not the same as generating strong profits from it. Marvell Technology, Inc. (NASDAQ:MRVL) will need to move from development and design wins to large-scale production while maintaining attractive margins. That process can require significant investment, and revenue growth does not automatically translate into equivalent growth in earnings.
There is also a risk that the stock has already priced in much of the good news. An 11% premarket jump shows that investors are expecting the Google relationship to become a meaningful growth driver. That creates a higher bar for Marvell in the quarters ahead.
Conclusion
The Google agreement is a meaningful development for Marvell Technology, Inc. (NASDAQ:MRVL) because it gives the company a broader role in Google's AI infrastructure while potentially creating a substantial source of custom-chip revenue through 2033.
The $12.18 billion warrant should not be treated as a guaranteed investment. Its value depends largely on how much business Google ultimately gives Marvell. That is also what makes the structure interesting for investors: the revenue-linked vesting provides a clear way to measure whether the partnership is actually delivering.
The biggest positives are Google's growing AI infrastructure needs, Marvell's expanding custom-silicon business, and the potential for the company to take a larger share of hyperscaler spending. The main risks are competition from Broadcom, execution challenges, shareholder dilution, and the possibility that the market has already priced in strong growth.
Overall, the deal strengthens the long-term bull case for Marvell, but investors should focus less on the $12.18 billion headline and more on the actual revenue Marvell generates from Google. The progress toward those $500 million revenue milestones will be the clearest indication of whether this becomes a major long-term growth driver for the company.
While we acknowledge the potential of MRVL as an investment, we believe certain AI stocks offer greater upside potential and carry less downside risk. If you're looking for an extremely undervalued AI stock that also stands to benefit significantly from Trump-era tariffs and the onshoring trend, see our free report on thebest short-term AI stock.
Four leading AI models discuss this article
"The real upside hinges on sustained multi-year Google spend converting into material, scalable volume for Marvell; without that, the warrants offer little conventional upside."
Marvell's expansion with Google adds a non-negligible tailwind to its custom-silicon ambition, especially as hyperscalers push bespoke AI accelerators. The revenue-linked warrant creates a quantifiable relationship between Google spend and equity upside, which aligns incentives. Yet there are material caveats the article glosses: the $12.18 billion figure is a potential, not guaranteed; most vesting depends on qualifying revenue, which may take years and may not materialize at scale. The real hurdle is converting design wins into profitable, high-volume manufacturing and margin expansion amid fierce competition from Broadcom and others, plus the risk of Google diversifying or slowing AI infra capex. Execution risk remains substantial.
Against the bullish view: Google’s AI capex could plateau or pivot to Broadcom or in-house TPU solutions, limiting real volume for Marvell. Even with milestones, the warrants are long-dated, and unless MRVL scales quickly, the upside could be largely theoretical and diluted by new shares.
"N/A"
[Unavailable]
"The warrant structure is a performance-based incentive that validates Marvell's growth, but the market is ignoring the risk that aggressive pursuit of Google's volume will lead to significant margin dilution."
The market is rightfully excited about Marvell’s (MRVL) deepening integration into Google’s TPU ecosystem, but the 11% pre-market pop feels like a classic 'sell the news' setup. While the warrant structure aligns incentives, it essentially serves as a performance-based commission for Marvell to become a secondary supplier. The real risk here is margin compression. Custom silicon (ASIC) development at this scale is notoriously capital-intensive and R&D-heavy. If Marvell has to sacrifice gross margins to displace Broadcom or win volume-based tranches, the bottom-line EPS growth may not justify the current forward P/E expansion. Investors are buying the dream of a duopoly, but they are ignoring the execution risk of scaling these complex designs.
If Google is willing to issue such a massive warrant, they are likely signaling a strategic shift to prioritize supply chain redundancy over Broadcom's current dominance, which could lead to higher-than-expected, long-term, high-margin volume.
"The warrant's revenue-contingent structure is honest disclosure, but it also reveals Google is not pre-committing to scale—Marvell must execute flawlessly to hit $500M milestones while competing against Broadcom's entrenched position and hyperscalers' own silicon teams."
The article frames this as a validation of Marvell's AI positioning, but conflates two separate things: a design win and a revenue commitment. The warrant structure is revenue-contingent, which is smart disclosure—but it also means Google has optionality, not obligation. Marvell needs $500M tranches to unlock vesting; that's a high bar in a market where hyperscalers are actively diversifying suppliers AND building more in-house. The 11% pop already prices in execution. What's missing: Marvell's gross margins on custom silicon (typically lower than merchant chips), manufacturing capacity constraints, and whether Google will actually commit capex at scale or use this as negotiating leverage against Broadcom.
If Google's custom-chip roadmap is already locked through 2031 with Broadcom, adding Marvell could be defensive positioning rather than additive growth—meaning Marvell cannibalizes Broadcom's share rather than expanding the total pie, and the revenue thresholds may never materialize.
"The warrants’ upside is highly contingent on a sustained, scalable Google-driven ramp; without that, the compensation is a tail risk rather than a material upside."
Claude, you’re right about margins and capacity, but the bigger, underappreciated risk is the timing and scale of Google’s capex ramp and MRVL’s unit economics. Revenue milestones depend on sustained, high-volume adoption; a capex slowdown or shift to in-house or Broadcom could prevent vesting long before margins improve. The upside from the warrants is a probabilistic tail, not a base-case lift.
[Unavailable]
"Marvell is being used as a tactical lever by Google to compress Broadcom's pricing power, which risks turning this 'win' into a low-margin commodity play."
Claude, you hit the nail on the head regarding the 'negotiating leverage' angle. Most are ignoring the geopolitical and supply-chain risk: Google is likely using Marvell to break Broadcom’s pricing power, not necessarily to shift volume. If this is a tactical move to force Broadcom to lower margins, Marvell is essentially a pawn in a price war. The warrant structure is not just an incentive; it is a hedge for Google against potential execution failure at Marvell.
"Marvell's real downside isn't losing the deal; it's winning volume at compressed margins while Broadcom's pricing power erodes the entire sector."
Gemini's 'pawn in a price war' framing is sharp, but it conflates two outcomes: Marvell as negotiating leverage doesn't preclude real volume. Google issuing $12.18B in warrants to a secondary supplier signals they're willing to pay for redundancy, not just theater. The risk isn't that Marvell gets used tactically—it's that Google *succeeds* in forcing Broadcom lower, which then pressures MRVL's own margins before revenue scales. That's the margin compression nobody fully connected.
The panel consensus is that while Marvell's partnership with Google presents opportunities, the execution risk is substantial, with key concerns being the timing and scale of Google's capex ramp, potential margin compression, and the risk of Google using Marvell as negotiating leverage against Broadcom.
The single biggest opportunity flagged is the potential for Marvell to gain a significant revenue stream and equity upside through its partnership with Google, if it can successfully execute on its custom silicon ambition.
The single biggest risk flagged is the potential margin compression before revenue scales, which could pressure Marvell's own margins if Google succeeds in forcing Broadcom to lower prices.