Marvell Earnings Preview: Can Custom AI Chips and 1.6T Optics Keep MRVL Growing?

 

Marvell Technology reports fiscal second-quarter 2027 results after the U.S. close on August 27, with a conference call scheduled for 1:45 p.m. Pacific time. Coming the night after Nvidia, it is the second semiconductor report drawing concentrated attention this week, and the first one capable of answering how far along hyperscaler custom silicon orders actually are.

 

What makes this print unusual is that the current quarter holds almost no suspense. Guidance called for revenue of $2.700 billion plus or minus 5% and non-GAAP earnings of $0.93 per share plus or minus $0.05. The Zacks consensus sits at $2.71 billion in revenue, up 35.2% year over year, with earnings of $0.93, up 39%. Consensus runs only about 0.3% above the guidance midpoint, among the narrowest gaps in recent years.

 

What will actually move the stock is third-quarter guidance, the gross margin structure of the custom business, and how management quantifies the Google agreement disclosed on August 19. There is also a thread that consistently gets overlooked: the generational shift in optical interconnect from 800G to 1.6T is becoming a revenue engine of comparable importance to custom silicon.

 

 

 

On expectations, consensus sits near $2.7084 billion, roughly 0.3% above the guidance midpoint, leaving essentially no room for a meaningful beat this quarter. The real bar sits one quarter out, where analysis puts third-quarter consensus revenue near $3.03 billion. Guidance that merely matches or falls below that level could be read as decelerating growth even if the current quarter beats.

 

On earnings quality, non-GAAP gross margin guidance runs 58.25% to 59.25% against 58.9% last quarter and market expectations near 59%. That metric says more than revenue about whether the custom ramp is delivering scale economics or buying orders through price concessions.

 

On price, the stock set an all-time high of $329 on June 18, fell to $162.90 in July, and has since rebounded roughly 49%, trading at $235.36 in pre-market on August 25. It is up about 179% year to date and roughly 221% over twelve months, yet remains more than 25% below the record. Options imply a post-earnings move near 10%.

 

On the sell side, TipRanks shows a Strong Buy consensus built on 24 Buy ratings against 5 Hold, with the highest target reaching $400.

 

 

 

Start by pinning down the expectation structure. May guidance called for $2.700 billion plus or minus 5% and non-GAAP earnings of $0.93 plus or minus $0.05. Consensus tracks that almost exactly.

 

For reference, the year-ago quarter produced revenue of $2.006 billion, up 58% year over year, with data center growing 69% and reaching 74% of the total, non-GAAP operating margin expanding 870 basis points to 34.8%, and earnings of $0.67 per share. This quarter's 35% growth is substantial but slower than the 58% it laps, and consensus is built on precisely that decelerating path.

 

Under that structure, exceeding the guidance midpoint by ten or twenty million dollars carries almost no share price meaning. Last quarter revenue of $2.418 billion came in $18 million above the midpoint and earnings of $0.80 edged the $0.79 estimate. Those figures appear in the results release and the market absorbed them long ago.

 

 

The pricing variable lives one quarter ahead. Moving from $2.71 billion to $3.03 billion requires sequential growth near 12%, a steep ramp in semiconductors. The market's willingness to assume it comes from the company having previously raised its outlook, with management citing exceptional AI-related bookings as the reason for significantly increasing revenue expectations for both fiscal 2027 and fiscal 2028.

 

Published expectations put current fiscal year growth near 40%, followed by roughly 45% growth to around $16.5 billion next year, with the Zacks consensus even higher at $16.63 billion. On earnings, the market looks for roughly 42% growth this year and about 53% next year to $6.18 per share.

 

That set of numbers carries the entire valuation. If third-quarter guidance lands below $3.03 billion, what the market has to recalculate is not one quarter but the slope of the whole curve.

 

 

A structural feature of custom ASIC work is that it typically carries lower margins than the corporate average. As that business takes a rising share of revenue, blended margin gets naturally diluted.

 

The most informative combination in this print is therefore whether accelerating revenue growth coincides with gross margin approaching the top of the guided band. Both together would indicate the custom ramp is producing scale economics. Revenue on target with margin at the bottom of the range would suggest growth is being purchased with price, and the quality of that growth would get repriced.

 

 

 

The regulatory filing submitted on August 19 is the most important new variable heading into this report. It disclosed a commercial agreement with Google signed on July 29 for custom semiconductor products, alongside a warrant covering up to 58,970,907 shares at an exercise price of $206.58. Roughly 57.6 million of those shares vest across 240 tranches, one per $500 million of custom product revenue, implying a cumulative purchase ceiling near $120 billion.

 

Timing is the crux. Performance vesting begins in the fiscal third quarter of 2027, meaning after the quarter being reported. Investors will therefore see no revenue from this agreement in the August 27 statements. Whether management supplies a recognition cadence, a delivery timeline and a gross margin band for the business matters far more to valuation models than whether the second quarter hits.

 

 

If most of the warrant vests and is exercised, potential dilution runs to roughly 7%. That cost sits in none of the current consensus earnings estimates, because when it occurs depends entirely on Google's purchasing pace.

 

The result is a timing mismatch: the positive revenue effect and the negative share count effect arrive roughly together, but markets habitually price revenue first and share count second. The earnings call is the first opportunity management has to explain that structure, and how it is described will directly shape how sell-side models get adjusted.

 

 

The custom business currently serves Amazon, Microsoft and Google. That distribution is itself an advantage, since it avoids dependence on a single account. But custom work is inherently lumpy, and management has previously acknowledged pronounced quarter-to-quarter variability in this business.

 

A useful approach when reading this print is therefore not to stop at aggregate data center growth but to note how management describes where individual customer programmes sit in their ramp. Data center revenue was $1.833 billion last quarter, up 27% and representing 76% of the total. Whether that concentration keeps climbing or stabilises shapes how the market assesses volatility.

 

 

 

While custom silicon absorbed all the attention, the generational transition in optical interconnect moved into substance. Per the company's March product announcement, Marvell introduced the first 200G per lane 1.6T DSP in 5nm with Nova in 2023, followed by the 3nm Ara platform in 2024. Ara now ships in mass volume to global customers, enabling hyperscalers to deploy 1.6T pluggable connectivity.

 

The March expansion added Ara X, Ara T, Petra and Aquila M, sampling to customers from the first quarter of 2026. Among them, Ara T is the first 8x200G transmit-retimed optics PAM4 DSP, delivering generational gains in performance and power against fully retimed alternatives. The company is also advancing its Photonic Fabric platform for multi-rack optical scale-up.

 

 

The significance here is easy to underweight. Industry analysis of a prior report argued that the driver is not general cloud expansion but specifically the transition to 1.6T networking and scale-up cluster architectures, noting that optical attach rates rise with every accelerator rack deployed, and that channel checks suggested optics spending growing faster than compute spending in several Tier 1 hyperscaler environments.

 

If that holds, the implication is direct: optical revenue does not depend entirely on the progress of any single customer's custom programme but tracks industry-wide rack deployment. Against the quarterly lumpiness of custom work, that is a smoother curve.

 

The company continues extending the use cases through partnerships, including a demonstration with Lumentum of optical circuit switching for next-generation AI scale-up infrastructure, using both the Aquila 1.6T coherent-lite DSP and the Ara 1.6T PAM4 DSP.

 

 

Three pieces of information matter here: whether 1.6T products are broken out as a share of data center revenue, whether Ara shipments continue ramping, and the timeline from sampling to volume production for the newest platforms. None of these appears in the financial tables. They surface only through management commentary, and their impact on next year's revenue curve may rival that of custom silicon.

 

For anyone tracking the stock through the print, the MEXC US equities page provides a live quote and access point via MEXC.

 

 

 

 

Assessing the risk-reward around this print requires seeing where price sits. The stock set an all-time high of $329 on June 18 and fell to $162.90 in July, a decline of more than half. It has since rebounded roughly 49% from the low on the back of an AI sector recovery and the Google disclosure, trading at $235.36 in pre-market on August 25.

 

The position is uncomfortable in both directions. A year-to-date gain near 179% means a great deal of growth is already priced. Yet the stock remains more than 25% below its record, so the market has not returned to peak optimism either. The report therefore becomes the event that decides direction rather than one that confirms an existing trend.

 

 

Options price a post-earnings move near 10%. Around $235, that is roughly $23 in either direction, framing a band of approximately $212 to $259.

 

That is considerably wider than the roughly 5.4% implied for Nvidia the same week, reflecting two things: a smaller market capitalisation with thinner liquidity, and results more sensitive to the progress of a handful of custom programmes, which widens the distribution of outcomes.

 

For position management, a 10% implied move means any stop set inside that band can be jumped through on the open. The approach that fits the event is adjusting size before the print rather than relying on stops afterward.

 

 

In the constructive case, second-quarter revenue and earnings meet or modestly exceed, gross margin approaches the 59.25% top of the range, third-quarter guidance exceeds $3.03 billion, and management supplies a recognition timeline for the Google programme. That combination validates both the pace and the quality of growth, creating conditions for a retest above $260.

 

In the neutral case, the quarter lands in line, third-quarter guidance sits near consensus, and management stays vague on Google. The report then supplies no new information, price digests within the roughly $212 to $259 implied band, and the stock returns to trading on sector beta. The October 6 Investor Day in New York becomes the next catalyst.

 

In the adverse case, third-quarter guidance comes in below $3.03 billion or gross margin lands at the bottom of the range. The former forces the market to revisit the $16.5 billion assumption for next fiscal year, the latter raises questions about growth quality. After a year-to-date gain approaching 180%, tolerance for negative surprise is limited.

 

Two external variables belong in the frame as well. Nvidia's results and guidance the previous evening will set sector risk appetite first, and Marvell's report will be read against that backdrop. And on competition, whether Broadcom's established position in custom silicon is materially affected remains unproven, which suggests the sector divergence priced on August 19 may have been premature.

 

 

The likeliest misreading of this report is treating it as an event about whether the quarter beats. With roughly 0.3% separating consensus from the guidance midpoint, the information content of the current quarter approaches zero. What actually needs pricing are three things not yet public: the specific third-quarter guidance number, the gross margin structure of the custom business, and the revenue recognition path for the Google programme. The first two appear in the release. The third can only emerge in the question and answer session.

 

The second overlooked issue is a shift in narrative weight. Over two months the market has placed nearly all attention on custom silicon while the optical business quietly completed its transition from 800G to 1.6T. The two lines carry entirely different risk characteristics. Custom depends on the schedules of a few customers and produces visibly lumpy revenue. Optics tracks industry-wide rack deployment and produces a smoother curve. When a company holds both lumpy and following revenue, the market tends to price only the lumpy portion, because it tells a better story. That produces two errors in sequence: overstating the certainty of custom during optimism, and understating the resilience of optics during pessimism.

 

The third point worth isolating is the timing structure of dilution. The warrant corresponds to roughly 7% potential dilution, but vesting depends entirely on Google's actual purchases, with performance tranches only beginning in the fiscal third quarter. The positive revenue effect and the negative share count effect therefore arrive on roughly the same schedule, and current consensus earnings estimates have not incorporated the second. That is not an immediate problem, but any valuation built on today's earnings estimates carries a gap that has not yet been filled.

 

For a watch list after the print, three data series outrank the share price. First, the direction of data center concentration, which stood at 76% last quarter, since rising concentration increases exposure to the AI capital expenditure cycle, an advantage on the way up and a risk on the way down. Second, where non-GAAP gross margin lands within the 58.25% to 59.25% band, the only public indicator of whether custom scale economics have arrived. Third, how warrant vesting progress gets disclosed in subsequent quarters, the only public metric that can quantify Google's actual purchase volume.

 

For cross-asset investors, the value of this report is that it provides an observation window independent of Nvidia. When compute suppliers and connectivity suppliers guide in different directions, it usually means the composition of capital expenditure is changing rather than its total. AI infrastructure investment is the central driver of the Nasdaq, and crypto has correlated closely with technology equities through much of 2026, so structural signals of this kind tend to appear before any directional change at the index level. All of the above is an analytical framework built on public information and is not a judgment on results or price direction.

 

 

 

The company reports fiscal second-quarter 2027 results after the U.S. close on August 27, 2026, with a conference call scheduled for 1:45 p.m. Pacific time. Results typically arrive between 4:20 and 4:30 p.m. Eastern. The first full trading session after the print is August 28. The company will also host an Investor Day in New York on October 6.

 

 

Guidance called for revenue of $2.700 billion plus or minus 5% and non-GAAP earnings of $0.93 per share plus or minus $0.05. Consensus sits at roughly $2.71 billion, up 35.2% year over year, with earnings of $0.93, up 39%. That is only about 0.3% above the guidance midpoint, among the narrowest gaps in recent years, leaving very little room for a meaningful beat. The year-ago quarter produced revenue of $2.006 billion.

 

 

Third-quarter guidance. Consensus sits near $3.03 billion, and guidance that merely matches or falls below that could read as decelerating growth even alongside a current-quarter beat. Non-GAAP gross margin comes second, guided at 58.25% to 59.25% against 58.9% last quarter. Margin near the top implies the custom ramp is delivering scale economics; the bottom of the range would suggest growth bought with price concessions.

 

 

No. Per the August 19 regulatory filing, performance-based vesting begins in the fiscal third quarter of 2027, meaning after the quarter being reported, so no related custom product revenue appears in these statements. The only information available will come from management's verbal commentary on the recognition timeline, delivery cadence and gross margin profile of that business.

 

 

Because it carries a different risk profile from custom silicon. The company introduced the first 200G per lane 1.6T DSP in 2023 and the 3nm Ara platform in 2024, which now ships in mass volume, with Ara X, Ara T, Petra and Aquila M added in March. Custom work depends on a few customers' project schedules and produces lumpy quarters, while optics tracks industry-wide rack deployment and produces a smoother revenue curve.

 

 

The warrant covers up to 58,970,907 shares at $206.58, and full vesting plus exercise corresponds to roughly 7% potential dilution. But vesting depends entirely on Google's actual purchases across 240 tranches, one per $500 million of custom product revenue. Dilution therefore arrives roughly in step with the revenue that creates it, and current consensus earnings estimates have not yet incorporated it.

 

 

Shares set an all-time high of $329 on June 18, fell to $162.90 in July, then rebounded roughly 49% from that low, trading at $235.36 in pre-market on August 25. The stock is up about 179% year to date and roughly 221% over twelve months, yet remains more than 25% below its record. TipRanks shows a Strong Buy consensus from 24 Buy and 5 Hold ratings, with the highest target at $400.

 

 

Options imply roughly a 10% move in either direction, about $23 around $235, framing a band of approximately $212 to $259. That is considerably wider than the roughly 5.4% implied for Nvidia the same week, reflecting a smaller market capitalisation, thinner liquidity, and results more sensitive to the progress of a few custom programmes. Any stop inside that band can be gapped through, so sizing matters more than stop placement.

 

 

This article is provided for information and market analysis purposes only and does not constitute investment advice, financial advice, legal advice, tax advice, or any recommendation to transact. Prices of equities, crypto assets and other related financial instruments can move sharply, and gaps around earnings events may substantially exceed both the options-implied range and the coverage of conventional risk controls. None of the guidance figures, consensus estimates, gross margin bands, technical levels or analyst targets referenced here can guarantee future outcomes, and consensus methodologies differ across data providers, so readers should rely on the company's official results release. The scenarios, revenue curve assumptions and dilution estimates described are forward-looking and may not be realised. Investors should reach independent conclusions based on their own financial circumstances, investment objectives, experience and risk tolerance, consulting a qualified professional where appropriate. The MEXC Crypto Pulse team accepts no liability for any direct or indirect loss arising from the use of, or reliance on, the information contained in this article.

 

 

James Mitchell specializes in technical analysis, market trends, and trading strategies for both Bitcoin and altcoins. Based in London, he has over 10 years of experience in financial markets. Before joining MEXC Learn, James worked as a senior analyst at a leading European investment firm, where he developed expertise in risk management and quantitative trading. His transition to cryptocurrency markets began in 2017, and he has since become recognized for his data-driven approach. He holds a Master's degree in Financial Economics from the London School of Economics. His analytical approach combines traditional technical analysis with on-chain metrics to provide readers with actionable insights.

 

His areas of expertise span technical analysis, market trends and cycles, trading strategies, Bitcoin and altcoin analysis, and risk management.

 

 

 

Want the fastest access to MEXC's latest updates? Join our official Telegram group now!

MEXC Platform: Download App | Zero Fee Trading | Buy USDT Now

Join MEXC Community: X (Twitter) | Telegram | Discord

Get Insights: MEXC Learn | Terms of Service | MEXC Blog | MEXC News

Account Verification: Understand KYC | How to Complete KYC

External Content Platforms: Substack | Medium | Paragraph | LinkedIn | X(News)