Snapshot: Executive Summary
Marvell Technology has become the public market's premier AI custom-silicon and optical-interconnect franchise, 18 active custom programs (12 across the four largest U.S. hyperscalers, 6 for emerging AI customers), a deepened NVIDIA partnership (a $2B preferred investment plus NVLink Fusion integration), and a FY2028 revenue target management raised to ~$16.5B. The business is firing on all cylinders: FY26 revenue $8.195B (+42%), record Q1 FY27 revenue $2.418B (+28% YoY), and addition to the S&P 500 (effective Jun 22, 2026)1.
The setup, and the catch:
- The thesis is real and it already paid out. The stock ran from ~$89 (early Jan) to a ~$316 ATH closing print (Jun 4, 2026), +231% YTD, before easing to ~$267. The AI-infrastructure narrative is intact and durable; the easy money has been made.
- The valuation is now full. At ~$267 MRVL trades at ~66x forward EPS and ~20x NTM EV/Sales, above Broadcom (AVGO, ~12x EV/Sales / ~23x P/E) and NVIDIA (~17x / ~20x), despite a smaller, more concentrated AI-customer book. The Street's consensus price target (~$242) sits below the current price.
- An idiosyncratic overhang sharpens the caution. The recent slide off the June ATH coincided with a CFO transition (outgoing CFO filed to sell shares) and a broad AI-chip wobble, a reminder that at this multiple sentiment cuts both ways.
This is a valuation call, not a thesis call. We rate MRVL Hold, 12-month PT $255 via a forward EV/Sales model on FY2028E revenue (Bull $320 / Bear $200). The probability-weighted blended fair value is ~$254, essentially at the base PT and below the current $266.77, so the risk/reward is balanced. We would turn constructive again on (a) a pullback toward the high-$100s/low-$200s, or (b) a confirmed third-leg catalyst (a named Google program, faster custom-silicon doubling) that lifts the FY28 base materially above $16.5B. Until then: own the franchise, respect the price.
Tactical: MRVL is trading at $266.77, ~5% above our $255 12-month target and ~10% above the ~$242 Street consensus. The AI custom-silicon thesis is intact and durable, but after a +231% YTD run the stock already discounts excellent execution at ~66x forward EPS / ~20x NTM EV/Sales. Pairs analytically with the AVGO custom-silicon view. Rating Hold: revisit on a pullback.
Investment Thesis
Bull Case
- Google two-AI-chip co-development converts to a named program
- AWS Trainium 4/5 + Microsoft Maia/Cobalt accelerate; custom silicon more than doubles in FY28
- FY28E revenue $18.0B (above the $16.5B guide)
- Optical / PAM4 + NVLink Fusion attach compounds
- Premium multiple holds at 15.6x EV/Sales (pure-play AI narrative sustains)
Base Case
- FY28E revenue $16.5B, management's raised target
- 18 custom programs execute; custom silicon doubles into FY28
- Gross margin ~59% (ASIC-mix pressured), operating margin ~35%
- Multiple normalizes toward, but still above, AVGO at 13.6x EV/Sales
- Stock is ~fairly valued; spot $266.77 sits ~5% above the PT
Bear Case
- Hyperscaler capex digestion; a custom-silicon program slips or is dual-sourced to Broadcom
- Optical pricing normalizes
- FY28E revenue $14.5B (below guide)
- Multiple compresses toward AVGO at 12.2x EV/Sales as the index/NVIDIA-halo premium fades
- Durable multi-year AI-infrastructure demand cushions the down case
Rating: Hold. Probability weights: 20% Bull / 55% Base / 25% Bear → blended fair value ~$254, essentially on the base PT $255 and below the current $266.77. The 55% base weight reflects high confidence in the structural AI-ASIC demand thesis; the 25% bear (above the 20% bull) captures the genuine valuation risk after a +231% YTD move to ~20x NTM sales. This is a valuation call, not a thesis call, we pair it with the AVGO view for a complete custom-silicon read (AVGO #1, MRVL #2), where MRVL now trades at a premium to AVGO on EV/Sales despite a smaller, more concentrated book.
Business Overview
Marvell has simplified its disclosure to two end markets: Data Center, and Communications & Other, replacing the old five-segment view (Data Center / Enterprise Networking / Carrier / Consumer / Automotive-Industrial). The shift is itself the story: the business is now run as an AI-data-center franchise with a legacy connectivity tail. In Q1 FY27, Data Center was $1,832.7M (76% of revenue, +27% YoY) and Communications & Other was $585.1M (24%, +29% YoY). Data Center surpassed $6B in FY26 (+46%).
End-market mix
Q1 FY27 revenue by end market
Data Center is now ~76% of revenue and rising. The old five-segment doughnut is obsolete, Marvell reports just two end markets, and the AI-data-center engine drives essentially all of the growth.
What's inside each end market
- Data Center (~76%): Custom XPU ASICs (hyperscaler AI accelerators such as AWS Trainium and Microsoft Maia/Cobalt), electro-optics / optical DSPs (PAM4, 1.6T / 3.2T), switching (Teralynx), and attach silicon (PCIe / CXL retimers). This is the whole growth thesis.
- Communications & Other (~24%): Carrier (5G base-station ASICs, OTN / optical transport), enterprise networking, and the consumer / automotive legacy. Cyclical and slower-growth; it serves as a cash-flow tail rather than a driver, and carries China / export-control exposure in carrier.
Why the two-segment view matters
The collapse from five segments to two reflects how completely AI data center now dominates the trajectory and the investment case. The non-AI tail still produces cash flow, but it is no longer the analytical anchor it was. The flip side is margin mix: custom ASICs carry lower gross margin (~59% non-GAAP) than the historical standard-product business (~64–66%), so the very mix shift that drives the growth also structurally pressures gross margin. Bull case requires the AI ramp to keep compounding; bear case is AI-capex digestion combined with continued margin pressure.
AI Custom Silicon
This is where the thesis lives, and it is a genuinely strong franchise. Marvell now has 18 active custom-silicon programs: 12 across the four largest U.S. hyperscalers plus 6 for emerging AI customers. Custom silicon doubled to ~$1.5B in FY26, is guided +20%+ in FY27, and to more than double in FY28, targeting >$10B by FY29. The trajectory:
Custom-silicon revenue trajectory ($B)
Custom silicon: ~$0.8B (FY25) → ~$1.5B (FY26) → guided +20%+ (FY27E) → more than doubling (FY28E) → >$10B targeted (FY29E). A steep, multi-year ramp, but one increasingly visible in the numbers, which is exactly why the multiple, not the trajectory, is the debate.
Active custom programs by customer group
18 active programs: 12 across the four major U.S. hyperscalers + 6 emerging
Marquee programs sit at AWS (Trainium, lead ASIC partner) and Microsoft (Maia / Cobalt); a third major U.S. hyperscaler and six emerging AI customers fill out the book; Google is in active talks (no program counted yet). The program count is deep, but a small number of large hyperscaler programs still drive most Data Center revenue, which is the concentration risk.
AWS Trainium: the marquee program
Trainium is Amazon's in-house AI accelerator for AWS data centers, designed by Marvell as a custom ASIC (analogous to AVGO's relationship with Google TPU). Marvell is the lead ASIC partner. The customer dynamics:
- Multi-generation visibility. AWS commits to multi-year roadmaps; Trainium 4/5 are on the horizon, with the re:Invent roadmap (Nov/Dec 2026) the key checkpoint.
- NRE economics. ASIC programs carry heavy upfront non-recurring engineering that recovers over the program lifetime, which is part of why gross margin runs below the historical standard-product level.
- Single-supplier exposure. If AWS dual-sourced a future Trainium generation to Broadcom, Marvell would lose substantial future revenue. This is the bear-case headline risk.
Microsoft Maia / Cobalt and the NVIDIA tie-up
Microsoft's custom AI accelerator (Maia) and Arm server CPU (Cobalt) for Azure are the second marquee relationship. Separately, the NVIDIA NVLink Fusion integration (Mar 31, 2026), alongside NVIDIA's $2B preferred investment, opens semi-custom AI-infrastructure attach, tying Marvell into NVIDIA's ecosystem even as hyperscalers build ASICs that displace NVIDIA GPUs (a double-edged dynamic discussed in Risks).
Google: the bull's key swing factor
The Information reported (Apr 2026) that Google is in active talks with Marvell to co-develop two AI chips. A named win would be a major re-rating event and is the central swing factor in our bull case, but talks are not a contract, and Broadcom is the entrenched TPU incumbent. Acquisitions of Celestial AI, XConn (PCIe / CXL / UALink) and Polariton (silicon photonics) further bolster the connectivity / optical stack. The franchise is excellent; the question the Hold poses is what is already in the price.
Competitive Position vs AVGO: Head-to-Head
The defining analytical exercise for an MRVL analyst is the head-to-head with Broadcom (AVGO). Both are custom-silicon ASIC leaders serving hyperscaler AI workloads. The crux, and a core reason for the Hold: MRVL now trades at a premium to AVGO on EV/Sales (~20x vs ~12x) despite far smaller scale, a more concentrated customer book, lower margins, and no software layer. Multiple parity with, rather than a premium to, AVGO would be more defensible.
Customer wins side by side
| Hyperscaler | AVGO program | MRVL program | Industry position |
|---|---|---|---|
| Amazon (AWS) | Some networking ASIC | Trainium (lead ASIC partner) | MRVL is the AWS preferred AI partner |
| Microsoft (Azure) | Some networking ASIC | Maia / Cobalt | MRVL ahead on Microsoft AI silicon |
| Google (GCP) | TPU (multi-generation) | In active talks (Apr 2026) | AVGO incumbent; MRVL working to enter |
| Meta | MTIA (current gen) | None announced | AVGO incumbent; next-gen could re-bid |
| Emerging AI customers | Select | 6 active programs | MRVL building an emerging-customer book |
| NVIDIA ecosystem | — | NVLink Fusion + $2B investment | MRVL semi-custom attach via NVIDIA |
Structural differences
| Dimension | AVGO | MRVL |
|---|---|---|
| Market cap | ~$1.74T | ~$233B |
| Custom-silicon position | #1 (Google TPU, Meta MTIA) | #2 (AWS, Microsoft, emerging) |
| Customer book | Broader / more diversified | More concentrated (few large programs) |
| Non-AI portfolio | VMware software + non-AI semis + wireless | Communications & Other (no software) |
| Balance sheet | ~$45B net cash | ~$1.4B net debt |
| Non-GAAP gross margin | higher (software mix) | ~59% (ASIC-mix pressured) |
| Forward P/E | ~23x | ~66x |
| NTM EV/Sales | ~12x | ~20x |
| Consensus rating / PT | Strong Buy / ~$524 | ~$242 consensus (below spot) |
Why MRVL trades above AVGO: and whether it should
The market awards MRVL a higher EV/Sales multiple for being the purer, faster-growing AI/custom-silicon play with an NVIDIA-ecosystem tie-up and fresh S&P 500 / index-inclusion liquidity. The bull case is that this premium is justified by growth and sustains as custom silicon doubles. The bear case, and our lean, is that a smaller, more concentrated, lower-margin, net-debt business should not trade at a ~70% EV/Sales premium to the diversified, net-cash, higher-margin category leader. As growth becomes increasingly visible in the numbers, that premium is more likely to normalize than to expand, which is why our base case has the multiple drifting toward (still above) AVGO.
Which to own
For a pure-play AI custom-silicon thesis, MRVL is the higher-beta vehicle (smaller base, faster growth, more concentrated customer book, richer multiple). AVGO is the more diversified, lower-multiple, software-plus-semi vehicle. The two-way pair makes analytical sense; our framework is to own MRVL on conviction that the AWS / Microsoft relationships deepen and a new leg (Google) lands, and on a more attractive entry than today's premium-to-AVGO multiple.
Financial Health & Trends
Revenue history ($B)
FY26 +42% ($8.195B) was the AI inflection. FY27E ~$11.5B (~+41%) and the raised FY28 target ~$16.5B (~+43%) show the ramp accelerating, not fading, the fundamentals are excellent. The debate is the multiple, not the trajectory.
Non-GAAP gross margin (%)
GM declined from ~66% (FY22) to ~59% (Q1 FY27, 58.9%) as the ASIC mix grew, custom silicon is structurally lower-margin than the legacy standard-product business. This is a genuine, durable mix headwind, not a transient one.
Non-GAAP operating margin (%)
Operating margin recovered to a strong 35.0% in Q1 FY27 (op income $846.9M) as revenue scaled and operating leverage kicked in, well above the FY25 transition trough. Gross-margin mix is the pressure point; operating leverage is the offset.
Q1 FY27 print highlights (record quarter)
| Metric | Q1 FY27 actual | YoY | vs guide |
|---|---|---|---|
| Revenue | $2,417.8M | +28% | ~$18M above mid |
| Data Center revenue | $1,832.7M (76%) | +27% | beat |
| Non-GAAP GM | 58.9% | (~1pp) | in-line |
| Non-GAAP Op Margin | 35.0% | +~5pp | beat |
| Non-GAAP EPS | $0.80 | +~33% | beat |
| Operating cash flow | $638.8M (record) | strong | — |
| Q2 FY27 guidance | $2.70B mid / EPS $0.93 | +~35% | — |
Record revenue, Data Center +27% YoY at 76% of mix, operating margin to 35%, record operating cash flow, and a Q2 guide of $2.70B / EPS $0.93, alongside a raised FY28 target of ~$16.5B. Operationally this is about as clean as a print gets; the entire Hold rests on what is already priced in, not on the results.
Capital Allocation
Capital priorities are clear and growth-first: fund custom-silicon R&D and NRE for the AI roadmap, pursue opportunistic M&A, and return a modest amount via buyback plus a token dividend ($0.06/quarter, ~0.1% yield). In Q1 FY27 Marvell returned $200M of buybacks and $53.8M of dividends while generating record operating cash flow of $638.8M. Note the balance sheet is now net-debt (~$1.4B), not net-cash, relevant to the valuation, where net debt is subtracted.
The NVIDIA $2B preferred
In March/April 2026 NVIDIA invested $2B via Series A convertible preferred, adding ~7.9M dilutive shares (reflected in the Q1 FY27 weighted-average diluted count of 893.3M). Beyond the capital, it cements the NVLink Fusion ecosystem tie-up, a strategic endorsement, though one that cuts both ways given NVIDIA is also the merchant-GPU vendor whose customers are building ASICs to displace it.
R&D / NRE intensity
Custom-silicon programs require deep, sustained IP investment, and R&D / NRE is the single most important strategic spend, concentrated on AI/ASIC programs in development, optical-interconnect IP (high-speed SerDes / silicon photonics), and next-generation switching. Underinvesting risks losing future program wins to Broadcom. Capital return is deliberately secondary while the AI roadmap is in its heavy-investment phase.
M&A history matters
Marvell's scale reflects transformational deals, Cavium (2018) and Inphi (2021), plus a recent burst of bolt-ons closing the connectivity / optical gaps: Celestial AI, XConn (PCIe / CXL / UALink), and Polariton (silicon photonics). Integration of four acquisitions is itself an execution risk, but the strategic logic (own more of the AI-data-center interconnect stack) is coherent.
Valuation Overview
NTM EV/Sales history: MRVL vs AVGO
MRVL's multiple compressed in 2022, rerated through 2024, and spiked to ~20x NTM EV/Sales in 2026 on the AI-ASIC / NVIDIA / S&P-500 narrative, opening a wide gap to AVGO (~12x). The two used to track; MRVL now sits well above. That gap is the heart of the Hold.
Peer NTM EV/Sales
MRVL at ~20x sits at the rich end of the AI-infrastructure complex, in line with the hyper-growth connectivity names (CRDO ~20x), above NVDA (~17x) and well above AVGO (~12x), below only the ALAB extreme (~33x). Rich for a smaller, more concentrated book.
Why EV/Sales over P/E for MRVL
MRVL is in a steep revenue-ramp / margin-normalization phase, so forward P/E swings on ASIC NRE cadence. EV/Sales smooths these dynamics and lets us compare cleanly against AVGO and NVDA on the same metric. (Cross-checks confirm the same conclusion: ~66x forward earnings is also a rich multiple.)
The crux of the Hold
At ~20x NTM EV/Sales (EV ~$235B on FY27E revenue $11.53B) and ~66x forward earnings, MRVL is priced for sustained, flawless execution. Three facts define the call: (1) it trades at a ~70% EV/Sales premium to AVGO despite smaller scale, more concentration, lower margin and net debt; (2) it trades above its own forward-EV/Sales-justified base PT of $255; and (3) it trades ~10% above the Street consensus PT (~$242). Our base case anchors on FY28E revenue $16.5B at a still-premium 13.6x EV/Sales, a multiple that normalizes toward, but stays above, AVGO. The thesis is intact and durable; the price has simply caught up to (and slightly through) it.
EV/Sales Scenario Model
FY2028E revenue × EV/Sales multiple = enterprise value; minus net debt ($1.4B) = equity value; divided by 0.875B diluted shares = the 12-month price target. We use FY2028E (the fiscal year ending ~Jan 2028) as the base because the PT is a 12-month-forward target, so no extra discount is applied; FY28E is already the next-year-ahead revenue stack. Net debt is subtracted because Marvell is net-debt, not net-cash.
| Step | Value |
|---|---|
| FY28E revenue × multiple | $16.5B × 13.6x |
| Enterprise value (EV) | $224.4B |
| Less: net debt | −$1.4B |
| Equity value | $223.0B |
| ÷ Diluted shares | 875M |
| 12-month price target | $255 |
| vs current market | — |
Sensitivity grid: FY28E revenue × EV/Sales multiple (PT in $)
Quick PT calculator
Risk / Reward calculator
Ask the Thesis AI-assisted checking…
Describe an AI-ASIC scenario in natural language; the assistant returns a structured impact analysis against this dashboard's forward EV/Sales model, the Hold framing, and the AVGO competitive dynamic. Powered by Claude via a Cloudflare Worker proxy.
Upcoming Catalysts
| Catalyst | Window | Why it matters |
|---|---|---|
| Q2 FY27 print | late Aug 2026 | Validates the $2.70B / EPS $0.93 guide and the FY27 ~$11.5B / FY28 ~$16.5B trajectory. The first checkpoint after the run-up. |
| Google two-AI-chip decision | any quarter | The bull case's key swing factor. A named program would be a major re-rating event; absent it, the Hold holds. |
| AWS re:Invent 2026 | Nov/Dec 2026 | Trainium 4/5 roadmap and whether MRVL remains sole/lead ASIC partner. Any dual-sourcing to Broadcom is the bear case. |
| NVLink Fusion ramp | FY27 | First semi-custom NVLink-Fusion design wins off the Mar-2026 NVIDIA $2B investment + ecosystem tie-up. |
| Custom-silicon doubling proof points | quarterly | Guided to more than double in FY28 and >$10B by FY29, each quarter either confirms or questions the trajectory. |
| Optical / silicon-photonics ramp | FY27 | Teralynx T100 (102.4T), 1.6T/3.2T optics, Celestial AI + Polariton integration. The interconnect leg of the story. |
| S&P 500 inclusion liquidity | effective Jun 22, 2026 | Index-fund forced buying / improved liquidity, but also an inclusion premium that could unwind. |
| Marvell Analyst Day / target refresh | FY27 | Long-range frame update; could lift the FY28+ revenue base above $16.5B (a bull trigger) or simply confirm it. |
Risk Factors
- Valuation, the #1 risk and the core of the Hold. After +231% YTD, MRVL trades ~66x forward earnings and ~20x NTM EV/Sales, with the Street consensus PT (~$242) below the current price. The stock is priced for sustained, flawless execution; multiple compression is the dominant near-term risk.
- Hyperscaler customer / program concentration. A small number of large custom-silicon programs drive Data Center. Any dual-sourcing to Broadcom, in-housing, or program slip (especially AWS Trainium) would be material.
- AI-capex digestion / cyclicality. A hyperscaler capex pause would de-rate the entire AI-semi complex, and MRVL's premium multiple is the most exposed to that.
- Broadcom (AVGO) competition. Deeper SerDes IP and incumbency at Google TPU and Meta MTIA. AVGO is the obvious second-source threat at MRVL's accounts.
- NVIDIA relationship is double-edged. NVIDIA is both a $2B investor / NVLink-Fusion partner and the dominant merchant-GPU vendor whose customers are building ASICs to displace it, strategic ambiguity that can cut either way.
- Margin mix. Custom ASICs carry lower gross margin than standard products; non-GAAP gross margin ~59% vs ~64–66% historically, a structural, not transient, headwind.
- Index-inclusion / halo premium could unwind. Part of the multiple reflects S&P 500 inclusion and the NVIDIA halo; both could fade.
- Execution & geopolitical. Integration of four acquisitions (Celestial AI, XConn, Polariton, plus earlier deals) and China / export-control exposure in carrier.
Bull vs Bear Debate
| Issue | Bull view | Bear view |
|---|---|---|
| Valuation (~20x NTM sales / ~66x EPS) | A premier, fast-growing AI/custom-silicon pure-play with an NVIDIA tie-up and index inclusion deserves a premium multiple; growth grows into it. | Priced for perfection, above AVGO despite a smaller, more concentrated, lower-margin, net-debt book; consensus PT sits below spot. |
| FY28 revenue ($16.5B target) | The raised $16.5B guide is conservative; Google + custom-silicon doubling + optical inflection push toward $18B. | Hyperscaler capex digestion or a program slip/dual-source could leave FY28 nearer $14.5B, below guide. |
| Google two-chip win | Active talks (The Information, Apr 2026) are real; a named win in 12–18 months is a major re-rating event. | Talks are not a contract. Broadcom's TPU incumbency at Google is structurally hard to displace. |
| vs AVGO premium | NVIDIA endorsement (Jensen: "next trillion-dollar company") and pure-play AI exposure justify trading above AVGO. | A ~70% EV/Sales premium to the diversified, net-cash, higher-margin category leader should normalize, not expand. |
| Margin profile | Operating margin recovered to 35% in Q1 FY27; operating leverage offsets the ASIC gross-margin mix. | Gross margin (~59%) is structurally below the legacy ~64–66% and stays mix-pressured as custom silicon grows. |
Resolution = Hold: own the thesis, respect the price. The franchise is excellent and the AI-infrastructure demand is durable, but at ~$267 the base case already discounts that, the blended fair value (~$254) sits below spot, and the next leg up requires a bull-case catalyst (a named Google program, a higher FY28 base) to convert. We turn constructive on a pullback or a confirmed third leg.
Technical Analysis
MRVL trailing-12-month monthly closes
RSI (multi-timeframe)
48 / 60 / 68, the daily has cooled off the ATH while weekly/monthly stay elevated. Consistent with a sharp pullback within a still-strong longer-term uptrend.
Relative strength (2026 YTD)
MRVL +231% YTD vs SOXX ~+34% vs SPY ~+11%, enormous outperformance on the AI-ASIC re-rating. That very outperformance is what leaves little margin of safety in the multiple.
Trader's view
- Price $266.77; 52-week range $61.44–$329.88; all-time-high closing print $316.43 (Jun 4, 2026). +231% YTD 2026, ~+341% trailing twelve months.
- Pulled back ~16% from the June ATH zone into the high-$260s after S&P 500 inclusion (Jun 22), a CFO-transition share-sale overhang, and a broad AI-trade wobble; consolidating below the ATH with extremely high realized volatility.
- Key support: the prior $200–215 breakout shelf. Key resistance: the $316–330 ATH zone.
- R/R from ~$267 to the $320 bull target with a $200 stop = ~0.8:1, weak, and consistent with the Hold: our $255 base PT sits below spot, so a long here has negative base-case reward. A pullback toward the high-$100s/low-$200s improves the setup materially.
- Volatility note: realized vol is very high, typical of AI-thematic semis at a stretched multiple.
Sources & Citations
Inline citations
Superscripted numbers in the body link here. Click any N in the report to jump back to the source.
- Marvell Technology, Inc., First Quarter of Fiscal Year 2027 Financial Results (period ended May 2, 2026; reported May 27, 2026) and Form 10-Q; FY2026 (Q4) results (Mar 5, 2026). Record Q1 FY27 revenue $2,417.8M (+28% YoY); Data Center $1,832.7M (76%); non-GAAP EPS $0.80; non-GAAP gross margin 58.9%; operating margin 35.0%; record operating cash flow $638.8M; Q2 FY27 guide $2.70B / non-GAAP EPS $0.93; FY28 revenue target raised to ~$16.5B; 18 active custom-silicon programs. Marvell IR ↗ · 10-Q ↗ ↩
Sources & further reading
- Marvell Q1 FY2027 results press release and Form 10-Q (period ended May 2, 2026), revenue, segments, margins, guidance, FY28 target.
- Marvell Q1 FY27 8-K exhibit, guidance detail and reconciliations.
- NVIDIA–Marvell NVLink Fusion / $2B investment announcement (Mar 31, 2026), and the Marvell IR version.
- S&P Global: Marvell to join the S&P 500 (effective Jun 22, 2026); CNBC coverage.
- Acquisitions: XConn (PCIe/CXL/UALink) and Polariton (silicon photonics); Celestial AI.
- stockanalysis.com, MRVL statistics and analyst forecast / consensus PT (~$242); AVGO statistics for the comp.
- Analyst notes (post-Q1 FY27): KeyBanc $385 (Street high), Stifel $230, Citi $215, BofA $200, Morgan Stanley / Wells Fargo $195. Benzinga summary · TIKR: FY28 target raised to $16.5B.
See the Important Disclaimers in the footer for the full not-investment-advice notice.