Marvell in the AI stack

The business and what the filings show

Marvell designs custom chips for hyperscalers and the digital signal processors inside optical transceivers. Over its latest four quarters it reported $9.5B of revenue, +31%, at a 52% gross margin; its data-center revenue was $7.2B, +33%.

Marvell competes with Broadcom for custom accelerator programs and supplies optical components that grow with the number of high-speed links in AI clusters. Its filings disclose a warrant issued to a customer in fiscal 2025, part of the pattern of suppliers using equity to win large orders.

Like every custom-silicon supplier, Marvell depends on a few large design wins. A program won or lost at one hyperscaler moves its numbers more than the overall growth of AI spending does.

Marvell’s position in the stack is between the chip and the network. Custom accelerators put it in competition with Nvidia’s merchant GPUs as well as with Broadcom, because every workload a hyperscaler moves onto its own chip is one it no longer buys from Nvidia. Its optical and electrical connectivity products, by contrast, grow with the number and speed of links in a cluster whoever makes the accelerators, which gives Marvell a second route into the build-out that does not depend on winning the next custom program. Reading its data-center revenue alongside its gross margin shows which of the two is driving growth: custom work tends to lift revenue faster than margin.

How AI shows up in the numbers

AI shows up at Marvell as custom accelerator programs for hyperscalers and as the optical and electrical connectivity chips that link AI servers. Its data-center end market, which it reports separately, is the part of the business that carries the AI story.

For scale, Marvell’s revenue over the latest four quarters equals 2% of the $566B the five largest buyers spent on capital projects over their latest four quarters. Their capex rose +81% on the year before, while Marvell’s revenue grew +31%: slower than the spending, so it is capturing a smaller share of each new capex dollar, or more of its business sits outside the build-out.

How it talks about AI

In its latest earnings release, furnished to the SEC on August 27, 2026, Marvell mentioned AI 2 times in 1,237 words of narrative text, about 1.6 mentions per 1,000 words. Of its 2 sentences about AI, 0 included a number: a revenue figure, a growth rate, a customer count or another measurable result. Across the 8 releases we read, back to March 5, 2025, AI mentions per 1,000 words went from 0.9 to 1.6.

Specific claims are easier to check against the numbers than general enthusiasm, which is why the share of AI sentences that carry a figure is worth watching alongside how often AI is mentioned. The sentences below are quoted from the release as published, after removing safe-harbor text and company boilerplate.

• “AI-related bookings remain exceptionally robust, and we expect our revenue growth to accelerate further through the remainder of fiscal 2027.”
• “We look forward to showcasing these growth drivers and sharing Marvell’s long-term strategy for enabling the continued expansion of AI infrastructure at our Investor Day on October 6 th , 2026.”

Source: earnings release (Form 8-K, Exhibit 99.1), August 27, 2026.

The trend in the numbers

Since the start of 2023, Marvell’s largest quarter by revenue was Q2 2026 at $2.7B, and its highest gross margin was 53% in Q2 2026. The latest quarter set the revenue record for the period.

Quarter by quarter, Marvell’s revenue grew +37% in Q3 2025, +22% in Q4 2025, +28% in Q1 2026, +37% in Q2 2026 on the same quarter a year earlier. The pace has been steady.

Gross margin in the latest quarter was 53%, against 50% a year earlier, so it is keeping more of each dollar as it grows. Operating margin moved from 14% to 17% over the same span. Over the latest four quarters it turned 23% of its revenue into operating cash flow and spent 5% of revenue on capital projects, against 4% a year earlier.

Segments that matter for AI

The parts of Marvell’s reported results most tied to the AI build-out, from the segment and product disclosures in its 10-K and 10-Q filings.

Where it sits in the stack

Marvell sits in the chips layer of the AI stack. It is paid by the AI clouds and servers and networking layers; its own spending flows to foundry & tools, memory.

Among the 7 companies in this layer on our map, it ranks number 6 by revenue and number 4 of 6 by revenue growth. Its gross margin of 52% compares with 66% for the layer as a whole, so it keeps less of each dollar than the layer average. The full layer analysis is on the chips page.

Latest four quarters for 10-Q filers; latest fiscal year for 20-F filers. Companies under $100 million of revenue are left out.

Against the 6 US-listed companies in its layer, Marvell grew +31% over the latest year versus +52% for the layer combined, and converted 18% of revenue into free cash flow versus 35% for the layer. It is growing more slowly than its layer, so its share of the layer’s revenue is shrinking.

What has to go right, and what could go wrong

What has to go right. Its custom-silicon programs have to reach volume at large hyperscalers, and optical connectivity has to keep growing with the size of AI clusters.

What could go wrong. Custom programs are lumpy: a single program lost to a rival or ended by a customer can move Marvell’s data-center revenue more than overall AI spending does.

These describe the conditions the business depends on, as the filings show them. They are not a view on the stock.

Through the lens of Investing in AI

Rob May’s book offers six mental models for telling AI moats from traps. The one most useful for Marvell’s layer is “AI changes markets and corporate strategy”: When creation gets cheap, choke points move. Like the internet shifting newspaper profits to discovery, AI pushes moats toward coordination and control points. The winners sit at the new choke point, not the best AI story.

The question it asks of any company is: What becomes scarce in this value chain as AI makes the work abundant? Applied to Marvell, the filings on this page are the place to start: the margin trend shows whether it holds scarce value or passes it through, the customer and financing disclosures show who it depends on, and the commitments show what it has promised before the revenue arrives.

Customers and concentration

US companies must disclose any customer that accounts for 10% or more of revenue, and foreign filers usually describe their largest customers. Concentration matters because a company’s growth is only as durable as the handful of buyers behind it, and in the AI stack those buyers are often the same few companies.

• We have not found a customer-concentration statement in Marvell’s latest filings that our reader can quote cleanly; see its filings on SEC EDGAR.

Commitments and backlog

• Backlog. Marvell doesn’t report a comparable total of remaining performance obligations in its XBRL data.

What to watch

Marvell’s next quarterly report will cover the quarter ending around October 31, 2026. Large companies must file a 10-Q within 40 days of a quarter’s end and a 10-K within 60 days of the fiscal year’s end, so the new figures should arrive by about December 30, 2026 at the latest. This page, the layer pages and the next quarterly report all update from that filing.

• Data-center revenue growth, where its AI programs show up.
• Custom-silicon program disclosures in its filings, since single programs move its numbers.
• Gross margin as custom work grows, which tends to carry lower margins than standard products.
• Its place in the next quarterly AI Stack Report, which updates every company on the map from the new filings.

Figures from Marvell’s SEC filings, latest period ended August 1, 2026. Other companies: all company pages.