MRVL can compound into a broader AI infrastructure supplier as custom silicon, connectivity, and optical attach scale across hyperscaler architectures.
The investment case is increasingly exposed to execution, competition, and valuation risk, while the proposed LITE deal adds dilution and integration uncertainty.
Street sees limited near-term disruption but greater strategic risk if MRVL acquires LITE, which is a primary narrow-linewidth laser supplier to both CIEN and CSCO and represented ~15% of LITE FY26 sales through CIEN. The issue is most relevant in coherent...
The company's proposal to acquire a laser asset was unanimously chosen by the seller's board because Marvell is a leader in TIAs, DSPs, and EICs, offering less execution risk compared to other bidders.
LITE / MRVL: Marvell confirms to purchase LITE in an all-stock transaction at fixed 4.50 shares of MRVL for each share of LITE, “representing approximately $1,132 per share of LITE common stock. Deal closing 1H ‘2H. Call at 8am.
The proposed acquisition of LITE by MRVL in an all-stock transaction is not viewed positively by many investors for either side.
The potential deal to acquire LITE is expected to result in heavy dilution for MRVL shareholders.
LITE +5% as MRVL -11% said to be in advanced talks to buy them.
The market turned on reports around 9:15am ET that the US and Iran had backed a Qatar/Pakistan compromise proposal, plus a weak pending home sales print that pulled yields lower, with small caps leading (IWM +1.7%). Breadth flipped from yesterday’s narrow b...
GFS Securities Jeff Pu anticipates that v10 will open up opportunities for Marvell Technology.
Stifel believes that Microsoft's Maia 300, with its multi-gigawatt deployment by CY27, validates Marvell's custom silicon ramp, with inference becoming a major growth driver.
Microsoft is positioning Maia as the default inference engine across its various platforms, and each gigawatt of deployment implies a significantly higher number of accelerators than previously expected, indicating strong demand for Marvell's custom silicon.
MRVL: Stifel says MSFT Maia 300 validates multi-GW Custom ramp, with inference becoming major growth driver
Microsoft's disclosure regarding the Maia 300 chip and potential discussions with AI labs supports the bull case for custom silicon revenue growth.
Skeptics argue that the bulk of the projected custom silicon revenue will not materialize until the 2028-2029 fiscal years.
MRVL +1.1% came back from -6.5% at the lows as MSFT walked through Maia 300 at OCP.
The shift from GPU-centric clusters toward heterogeneous GPU/CPU/ASIC/wafer-scale systems and open rack-scale standards is positive for the company.
Raised FY28/FY29 targets and the FY31 framework are supported by customer/supplier commitments, with scale-out/scale-across remaining core and scale-in becoming increasingly favorable.
Custom revenue is projected to rise significantly, driven by XPU attach content, which could be $500–2,000/device and potentially $2,000–3,000 at scale.
Citi expects OCP to highlight a shift from GPU-centric clusters toward heterogeneous GPU/CPU/ASIC/wafer-scale systems and open rack-scale standards, positive for AMD/MRVL/AVGO. Power should center on 800V DC adoption, favoring TXN/MPWR/AEIS/Delta, while con...
The company's data center opportunity extends beyond XPUs, with scale-out driving most of its interconnect target and scale-up representing a separate significant total addressable market.
Custom solutions are broadening beyond XPUs into attach products, Smart NICs could reach substantial market value, inference accelerators are ramping, and Ethernet switching is tracking for significant growth with further upside.
The company's risk profile has significantly improved as growth shifts towards its connectivity franchise and away from concentrated custom XPU exposure.
Custom accelerators are now expected to contribute a smaller portion of CY30 revenue compared to previous fears, reducing concentration risk.
Management's targets for interconnect, switching/storage, and custom revenue by CY30, along with strong FY28 datacenter revenue expectations, indicate robust future growth.
The analyst projects significant EPS growth within three years.
JPMorgan's preview for the investor day projects FY29 custom silicon revenue at approximately $12.5 billion, exceeding the Street's $10.3 billion, and FY29 total revenue at approximately $30.3 billion, above the $26.4 billion consensus.
Funda's FY27/FY28 numbers are significantly above consensus, and the stock's underperformance leading into the event suggests cleaner positioning.
The stock had already run up 25% in September, potentially limiting further upside from the investor day.
Big day tomorrow for MRVL -38bps with the investor day: JPM’s preview has FY29 custom silicon at ~$12.5B vs. the Street’s $10.3B (the GOOGL XPU attach isn’t in numbers) and FY29 revenue ~$30.3B vs. $26.4B. In Slack here, Ash noted Funda’s F27/F28 numbe...
The company raised its FY28 revenue target to ~$20 billion and provided ambitious long-term targets for FY31 revenue and AI TAM, which helps bulls who sought clarity on the long-term case.
MRVL +5.5% (up as much as +11% but fading) at its investor day: FY28 revenue is now ~$20B (from $18B, vs. a Street at ~$18.2B), with $70-90B by FY31 and a ~$400B AI TAM by 2030 (custom $235B, switching $85B, interconnect $65B), 4x+ its prior $94B 2028 estim...
ALAB +7.4%; CRDO +4.4% traded as MRVL sympathy on the connectivity TAM.
AVGO +4.7% outperformed as investors digested the FT’s details on the record $60B financing for Anthropic’s 2027 TPU orders (AVGO backs the $42B of senior loans; the $18B junior tranche carries no AVGO guarantee) alongside MRVL sizing the custom-silicon TAM...
Bears argue that the ambitious FY31 revenue target of $70-90 billion requires every milestone to be met, similar to the back-end-loaded math of the Google deal.
The stock's fade after an early pop suggests a 'sell-the-news' reaction, indicating that the positive news was already priced in.
JPM's preview for the investor day projects FY29 custom silicon revenue at ~$12.5B, significantly above the Street's ~$10.3B, and FY29 total revenue at ~$30.3B, exceeding the Street's ~$26.4B.
Funda's FY27/FY28 numbers are well above consensus, suggesting strong future performance.
The stock not trading well into the investor day could indicate cleaner positioning among investors.
The stock had already run up significantly, gaining 25% in September, which could limit further upside in the short term.
Big day tomorrow for MRVL -38bps with the investor day: JPM’s preview has FY29 custom silicon at ~$12.5B vs. the Street’s $10.3B (the GOOGL XPU attach isn’t in numbers) and FY29 revenue ~$30.3B vs. $26.4B. In Slack here, Ash noted Funda’s F27/F28 numbe...
The analyst is bullish into the company’s 10/6 Investor Day, expecting a new long-term model potentially stretching to FY31 with ~$45B revenue and ~$19 EPS, implying ~45% 4-year EPS CAGR.
The biggest upside is seen in XPU “Attach” rather than compute, highlighting the $120B Google deal focused on TPU attach sockets and estimating 2–3 attach chips per XPU at ~$3K ASP each.
AWS Trainium is expected to grow ~30% y/y in 2027.
The analyst expects NPO adoption to ramp in 2H27 and sees a stronger optical/connectivity story into 2030.
Core compute ramps look slower, including MSFT.
CRDO +7.9% still riding Jefferies' NYC NDR (SPCX potentially adding another 660K GB300s; OmniConnect ramping in 2H27). Irrational Analysis also added the name to his portfolio and had some positive comments, saying CRDO’s SerDes portfolio is second only to...
INTC +3.7% got a cameo at the SNPS analyst day, where Ghazi raised the "you have lost Intel" question himself and answered it with an INTC video on 18A/18AP/14A co-optimization; INTC also hired Sunil Kaul from MRVL to run hyperscale and data center sales. B...
Jefferies expects Marvell to target over $40 billion of FY30 revenue, including approximately $22 billion from Custom and nearly $15 billion from Optics.
Switching/storage is expected to improve as 51.2T ramps and 102.4T becomes more meaningful in C28/C29.
Jefferies sees custom XPU revenue reaching approximately $12 billion in C29 from approximately $2.5 billion in C26, driven by a meaningful Maia 300 ramp in C27 and at least one additional high-volume customer.
MRVL: Jefferies expects Investor Day to raise LT targets, with MSFT the key swing factor for custom XPU upside
MRVL +4.5% as MediaTek fell 7.1% in Taipei on rumors that GOOGL’s TPU v9 (HBM4E) work could shift to MRVL; MediaTek called the order-loss claims false. Investor day is next Tuesday (10/6), where Jefferies expects a $40B+ FY30 revenue target (~$22B custom, ~...
Microsoft is the biggest swing factor for custom XPU upside, and Amazon remains more constrained by multi-sourcing and pricing discipline.
Despite expectations for higher targets, the stock could still see some sell-the-news reaction.
RBC expects Marvell to raise its FY29 Google custom-silicon opportunity by at least $2 billion and potentially lift Microsoft XPU expectations as Maia 300 ramps.
Scale-up optical/switching is seen as potentially larger than scale-out over time, with NPO/CPO and UALink/ESUN/NVLink opportunities adding another growth leg.
UBS raised its price target based on expectations for increased custom-silicon unit volume in 2027.
MRVL -3.8% fell despite bullish investor-day previews: RBC expects the GOOGL custom-silicon opportunity to rise at least $2B from the $10B+ framework; UBS went to $335 from $310 on 350K Maia units in 2027.
TSEM closed at the lows with no news after a +14% five-day run. Wolfe’s ECOC recap was about as bullish on demand as it gets (META showed amplifier huts going from 48 fiber pairs to >4,000; Lightera has closed its fiber order book; PAM4 DSPs sold out throug...
Broader 200G/lane 102T switch ASIC support from various companies is expected to enable the 1.6T optical transceiver ramp starting in CQ4'26 across next-generation platforms.
Amazon's Trainium3 UltraServer and NeuronSwitch are already in volume production, with this company identified as a clear beneficiary.
Active copper and near-packaged optics are expected to remain key near-term growth vectors before CPO becomes more material in multi-rack scale-up.
BNP says broader 200G/lane 102T switch ASIC support from MRVL/AVGO/NVDA/CSCO should enable the 1.6T optical transceiver ramp beginning CQ4’26 across Vera Rubin, MI455x Helios and Google TPUv8.
Piper initiated coverage with an Overweight rating and a $270 price target, identifying Marvell as a long-term XPU winner with leading custom attach.
Piper ranks Marvell alongside Nvidia and Broadcom as a long-term winner in XPU compute and states that Marvell leads in custom attach.
Piper also flags the upcoming analyst day as a potential catalyst to clarify the growth framework and custom-silicon opportunity.
Piper ranks MRVL alongside NVDA and AVGO as a long-term winner in XPU compute and says Marvell leads in custom attach. Piper also flags the upcoming analyst day as a potential catalyst to sharpen the growth framework and custom-silicon opportunity.
JPM says the key debate remains Google, but JPM argues MRVL/MediaTek/Alchip/GUC are mostly supporting Google’s broader COT and TPU-attach ecosystem rather than displacing AVGO’s core role, which is protected by the 5-year TPU v8–v11 agreement.
The stock declined due to concerns about increased custom-chip competition following NVIDIA's investment in MediaTek and an expanded NVLink Fusion partnership, suggesting MediaTek could become a more aggressive competitor for hyperscaler ASIC wins.
MRVL -2% as NVDA’s $3.5B convert investment in MediaTek and expanded NVLink Fusion partnership likely added to custom-chip competition concerns.
The company reported a gross margin miss for the third quarter, with the midpoint falling below street expectations and declining sequentially.
RBC expects a modest beat/raise driven mainly by networking, with strong demand across Tomahawk/Jericho, PCIe and optical helping offset lingering TPU share concerns. The key debate is FY27 AI revenue: RBC thinks management could lift the prior $100B+ frame...
Google’s more COT-like TPU v10 architecture creates incremental backend/3D-IC opportunities for the company.
The company is a key beneficiary of the broadening optical scale-up in AI networking.
The most interesting incremental read-through is for MRVL, which showcased a Photonic Fabric Memory Appliance that pools memory optically across multi-rack AI systems, potentially creating a new inference-memory tier between HBM and storage.
Optical engine standardization favors the company.
deployment growth. More incrementally, UBS says MRVL’s new Google inference engagement does not appear to cannibalize AVGO’s core TPU role and reiterates that AVGO remains Google’s “plan of record,” but trims CY27 AI revenue to ~$130B as prior TPU unit assu...
BNP also sees 224G/lane SerDes and larger switch radix supporting continued scale-up/scale-out growth for AVGO/ANET, while optical engine standardization favors MRVL, LITE, COHR and MTSI.
The stock has fallen 30.5% from its high, indicating a substantial downturn.
AMBA: -27.6% JBL: -28.9% MCHP: -29.9% MRVL: -30.5% STX: -30.6% VRT: -32.9% TSEM: -34.0% NXPI: -34.6%
JPM views the Google deal as validating a broader ASIC TAM, indicating that hyperscaler custom silicon spending is expanding beyond just accelerators.
The warrant math implies up to ~$120 billion of cumulative revenue, or ~$19.2 billion per year through FY33 if fully vested.
JPM sees a path to approximately $11 of CY28 EPS (versus Street ~$9.52) if a meaningful portion of the Google opportunity ramps.
JPM reiterates an Overweight rating and sees the deal as a positive read-through for the broader custom ASIC TAM.
JPM says recent concerns around Google custom silicon competition are misplaced: AVGO’s 5-year Google agreement locks in the next four TPU generations, while TPU v9 is now ramping and the next 2nm XPU ASIC remains on track for CY28 with no CoWoS cuts or del...
MRVL: JPM sees Google deal validating broader ASIC TAM, path to ~$11 of CY28 EPS
CXL memory-expansion products from the company are expected to ramp in 1H27, with true memory pooling ramping later in 2027.
BNP expects CXL memory-expansion products from MRVL/ALAB to ramp in 1H27, with true memory pooling ramping later in 2027 as PCIe 6 CPUs such as AMD Venice and Intel Diamond Rapids reach volume.
Wells Fargo reiterated Overweight and a $240 PT, seeing upside in both major Data Center engines: custom XPU + attach growth could move from >20% to 30%+ in FY27, and Interconnect could accelerate from >70% to 80%+ year-over-year.
More than 10 XPU-attach programs are moving toward high-volume production in FY28 across NICs, CXL/memory, storage, and security, with AWS Trainium and Microsoft Maia providing additional upside.
Scale-up optics are tracking ahead, with Celestial AI already guided to roughly $300M in FY28, and the October 6 Investor Day is viewed as a second catalyst.
MRVL is developing CXL and disaggregated memory solutions, reinforcing the broader trend toward moving memory closer to compute while reducing dependence on expensive HBM.
The stock was a significant mover in AI semis, indicating strong positive momentum.
MediaTek's earlier-than-expected 400G SerDes IP arrival in 2H27 poses a negative read-through for Marvell, as it adds another credible bidder for Marvell's large AI ASIC programs.
KeyBanc reiterated an Overweight rating and raised its price target, citing gains from Trainium 3, prospective Trainium 4 share, and Google’s Merope LPU.
The company is expected to benefit from significant revenue contributions from Google's Merope LPU in the coming years.
KeyBanc reiterates Overweight, raises estimates and lifts its price target to $155 from $110, citing outsized server CPU demand from agentic AI, with Intel capacity supporting 25%–30% unit growth this year and more than 50% next year. The firm says 18A yiel...
AMD/MU/MRVL: KeyBanc Raises Targets as AI Compute, Memory Pricing and Custom Silicon Visibility Improve
KeyBanc reiterates Overweight across AMD, MU and MRVL, raising AMD’s price target to $725 from $530, MU’s to $1,750 from $1,600 and MRVL’s to $400 from $385.
The stock declined, falling below its earnings gap and showing a rollover pattern.
AI infrastructure demand remains strong with improving visibility across cloud customers.
The custom XPU business continues to track well with Amazon and Microsoft, and XPU attach opportunities remain underappreciated, potentially becoming a more meaningful growth driver.
Concerns around Amazon share loss have largely been addressed, and there is potential for additional hyperscaler wins not yet reflected in forecasts.
Meta's large-scale deployment of its custom Vistara CXL ASIC validates CXL as a production-ready technology, increasing investor confidence in the long-term CXL ecosystem.
The deployment highlighted reported 20-25% compute capacity savings and 4-12% throughput gains, reinforcing the long-term investment case for CXL across datacenter memory and server infrastructure.
UBS raised estimates as the CXL opportunity is inflecting faster than expected, driven by agentic AI demand for rack-scale and multi-rack CPU/XPU fabrics.
The firm expects the CXL-related ASIC market to reach $7-10 billion by 2030, with Marvell remaining the market leader and CXL revenue approaching $1 billion in 2027 and $2 billion in 2028, driving estimate revisions and a higher price target.
MRVL/ALAB: UBS Raises Estimates as CXL Adoption Accelerates With Agentic AI
In Tech, semis -6%+ following Korea lower. What’s been strong lately is down the most: Memory -7-8%; MRVL/ALAB -8%, BE -7% and so on…Nothing being spared in the sea of red.
Investors are underestimating the larger and more durable networking opportunity compared to custom XPUs.
The company's networking revenue is projected to grow significantly, driven by 1.6T upgrades, silicon photonics, scale-out AI networks, and increased content per AI cluster.
Custom XPU revenue is expected to exceed $10 billion by FY29, with strong visibility from AWS Trainium ramps, Microsoft’s Maia program, and additional XPU attach opportunities, exposing the company to multiple AI infrastructure growth vectors.
The stock was weaker, consistent with its underperformance within AI semis over the past month.
Post-close, ADBE -6% despite a modest beat and raise as CFO departing (going to MRVL). Main nit is that ARR includes $450M+ from semrush so only grew 10%, down from close to 11% last q,. Good earnings thread in Slack here
MRVL is leading the way higher after being added to the S&P 500.
MRVL +8% and FLEX +4% leading the way up early after being added to SP500. Memory names strong with MU +6.5% and SNDK +4%
The company announced it would join the SP500, and its COO stated that Marvell has secured enough production capacity to meet its 50% growth target this year and 55% in 2027. Additionally, its SerDes technology is advancing, with 200G on 2nm process and 500G on A14, and the company anticipates a gradual adoption of NPO and CPO over the next five years.
FLEX -1% lagged as only moving from SP400 to SP500. Digitimes had an interview withMRVL’s COO Chris Koopmans here where he said that MRVL has secured enough production capacity to meet its 50% growth target this year, 55% in 2027. He said MRVL’s SerDes now...
The stock continues to perform strongly following positive comments from Jensen Huang regarding a $1 trillion valuation, positioning it as a leading candidate for inclusion in the S&P 500.
Marvell will design a networking chip for Google, which will be manufactured on Intel's 18A (or 18AP) process, with mass production scheduled to begin by the end of 2027.
The company's stock surged significantly after NVIDIA's CEO, Jensen Huang, publicly endorsed them, suggesting they could be the next trillion-dollar company, which is interpreted as a strategic move to maintain supply chain satisfaction and foster partnerships.
MRVL +32%(up another 6% in the postmarket) as Jensen said they were the next trillion dollar company. Pays to have Jensen as a shareholder. Don’t think this was an offhand comment — Jensen knows what he’s doing: keep supply chain happy, confirm that if you...
In Tech, we have a busy week coming up with Earnings from AVGO, CRWD, CRDO, HPE, GTLB, PANW, CIEN, DOCU and more. Computex continues through the week with keynotes INTC’s Lip Bu Tan, and MRVL CEO. BAML’s and ISI’s TMT conferences start Tuesday as well as th...
The company is successfully evolving into a compounding AI infrastructure platform with multiple growth drivers beyond optics, including custom silicon and networking.
The stock's valuation relies heavily on long-term targets for custom silicon programs, which are subject to execution risks, supply availability, and competitive pressures.
Gross margins may face pressure as the revenue mix shifts toward custom silicon, and free cash flow conversion lags behind some AI peers.
Vik's Newsletter argues that increasing AI rack power density is creating a new interconnect category where Active Copper Cable (ACC) becomes the optimal solution between passive copper and optics.
The company is viewed as the strongest challenger in the Active Copper Cable market given its hyperscaler relationships and broad copper portfolio.
ACC/Interconnects/SMTC/MRVL: Vik’s Newsletter Says Rising Rack Power Density Is Creating a New Active Copper Opportunity
The stock continues to reach new highs ahead of its earnings report.
LITE -4%: Lots of questions today around this one but no great answers. Rest of optical space fared slightly better: COHR +1%; CIEN +3%; GLW +1%…MRVL +6% continues to rip to new highs ahead of earnings.
The company is benefiting as ASICs evolve into broader systems opportunities requiring SerDes, chiplets, optics, rack-scale networking, and custom I/O.
The company appears well positioned for upside from Trainium/Maia ramps.
The company is seen as having multiple AI upside drivers beyond the near-term print, including AWS Trainium deployments, XPU attach ramps, and accelerating AI interconnect demand.
There is increasing investor focus on the company's custom XPU opportunity, particularly tied to AWS Trainium and Microsoft Maia.
AI interconnect revenue is now expected to grow approximately 35% year-over-year.
EPS estimates for FY27/FY28/FY29 have been materially raised.
INTC +5% leading the way higher following Lip Bu Tan’s bullish JPM Conference appearance yesterday (we sent out notes here in case you missed it). The rest of AI semis strong as well with MRVL +4%, MU +4%, ALAB +3% among others as we head into NVDA earnings...
ISI also notes MRVL appears well positioned for upside from Trainium/Maia ramps, while QCOM remains earlier-stage and more focused on edge/custom AI opportunities through Alphawave.
The stock reached new highs, indicating strong market performance.
Positive checks suggest the company's role in Trainium 3 is capacity-driven due to a competitor's allocation shortfall.
The company is seeing upside from Trainium-4 and Maia, benefiting from the evolution of ASICs into broader systems opportunities.
The near-term setup into F1Q27 is tougher.
AI networking checks remain strong, with scale-out demand still supply constrained and hyperscaler architectures increasingly converging toward optical interconnects/CPO over time.
The company is benefiting from rising 800G/1.6T DSP share gains and growing hyperscaler/NPO demand.
CXL may finally have a meaningful AI-driven use case through KVCache expansion as million-token context windows rapidly increase memory requirements for inference workloads, with 2027 seen as a key inflection year for the CXL ecosystem.
The company is seen as the primary beneficiary of the CXL inflection through its XConn acquisition and growing hyperscaler engagements.
AI Networking/ALAB/MRVL/NVDA: ISI Checks Remain Positive on NVDA, MRVL, and ALAB
ISI also sees MRVL benefiting from rising 800G/1.6T DSP share gains and growing hyperscaler/NPO demand, while ALAB continues gaining credibility through strong execution/customer support and increasing scale-up switch demand through 2027.
The company's stock saw a 4% increase, catalyzed by positive snippets in a TheInformation article regarding customer interest in using Trainium.
The company is likely benefiting from Google's TPU announcement, as it is perceived as gaining market share in Google's TPU supply chain.
ISI's positive checks indicate that the company is benefiting from increasing share in 800G/1.6T DSP and growing demand from hyperscalers and Network Processor Organizations (NPOs).
Wells Fargo views the company as a primary beneficiary of the CXL ecosystem due to its XConn acquisition and expanding engagements with hyperscalers.
Investor skepticism remains high regarding CXL adoption given prior false starts.
The stock was weak following POET's significant decline, indicating a broader negative sentiment in the optical sector.
In Tech land, semis broke their 18-day win streak although plenty of names were green on my screen. The early morning semi swoon was catalyzed by a POET -50% blow up early morning, one of retail’s favorite “undiscovered” optical plays after MRVL pulled a pu...
The expanded AWS/Anthropic agreement is a material positive, reinforcing AWS as a key demand driver and supporting stronger 2027+ growth visibility, with Marvell leveraging optical/connectivity (PAM-4, DSPs, switches) for a significant AWS silicon revenue opportunity and accelerating growth into FY28.
MRVL/ALAB: RBC Reiterates Outperform; AWS/Anthropic Deal Driving Optical + Trainium Upside
RBC says the expanded AWS/Anthropic agreement (incl. ~5GW capacity and multi-gen Trainium roadmap) is a material positive for MRVL and ALAB, reinforcing AWS as a key demand driver and supporting stronger 2027+ growth visibility; the firm raises PTs to $170/...
The company's stock was up 6% on news that Google is in talks to develop new AI chips for inference.
MRVL +6%/GOOGL/AVGO: Google in Talks With Marvell to Build New AI Chips for Inference
JPM believes there is no real story regarding Google's discussions with the company about TPU business, suggesting Google is talking to many companies.
The narrative is shifting towards optical, where demand and growth are accelerating, leading to an upgrade.
Optical revenue is expected to grow significantly, driven by rising port counts and AI networking demand, even with some share shifts.
Investor expectations are too low, with potential upside from AWS demand, XPU attach growth, and a potential Microsoft ramp.
The analyst sees meaningful earnings power and multiple expansion as optical becomes the key driver.
Citi models ~$6 in CY28 EPS (27% above Street), ~40%+ EPS CAGR over three years, and sees ALAB’s 97%/80% retimer share in CY25/CY26 and PCIe Gen6 leadership as durable moats even as MRVL competition intensifies.
The stock appears to be one of the first AI semiconductor companies to approach 52-week highs.
The company is strategically partnering with NVIDIA to integrate into the NVIDIA AI factory and AI-RAN ecosystem through NVIDIA NVLink Fusion, which offers greater choice and flexibility for customers building on NVIDIA architectures.
The company will collaborate with NVIDIA on silicon photonics technology.
NVDA/MRVL +9%: NVIDIA AI Ecosystem Expands as Marvell Joins Forces Through NVLink Fusion - NVDA Invests $2B in MRVL NVDA/MRVL +9% Marvell NVLink Fusion
The company announced a partnership with NVIDIA to connect to the NVIDIA AI Factor and AI-RAN ecosystem through NVIDIA NVLink Fusion, along with a $2 billion investment from NVIDIA.
The expanded NVIDIA-MRVL partnership is seen as a broadening collaboration into silicon photonics and AI-RAN for 5G/6G, confirming MRVL as the primary NVLink Fusion integration partner for AWS's Trainium4.
The partnership signals NVIDIA's acknowledgment that custom ASIC XPUs are here to stay, and their strategic response is to embed their interconnect, networking, and CPU stack into XPU racks rather than solely defending GPU-only deployments.
The stock is near its 52-week highs, indicating it could be the first AI semiconductor company to break out to new highs.
The stock performed well, suggesting positive sentiment.
The company saw a positive bid following the OpenAI/Amazon deal.
MRVL +50bps a bit of a bid on the OAI/AMZN deal.
The company is seen as more flexible and strategically embedded in the hyperscaler ASIC market, including being the main supplier for AWS Trainium 4 and indispensable to Microsoft’s Maia program, with improved execution on defect rates.
Raymond James views Marvell’s acquisition of XConn as a small strategic tuck-in that modestly strengthens its PCIe/CXL and UALink capabilities and broadens the XPU/ASIC bundle.
The deal implicitly acknowledges some catch-up in switch/retimer silicon and validates scale-up copper interconnect as a viable near-term growth vector.
Raymond James kept estimates unchanged, noting the deal is a technology tuck-in and increases competitive overlap with other companies.
Fears regarding losing a key customer socket to a rival ASIC supplier are unfounded and overstated.
Speculation and misleading reports about losing next-gen platform sockets have caused market overreaction.
Al trade all getting a bid: CRWV +5% on positive ISI note; NBIS +2.8% on a bullish GS note; ANET +6% (didn't see anything new here); MRVL 5% #RSRI) Ip) ; ANET +6% (RMIAE) ; MRVL MU +5% as DRAM prices continue to climb [News] Korea, U.S.
Semis/IT Spending: JPM CIO Survey Shows Strong AI and Cloud Spending Intentions; NVDA, AMD, AVGO, MRVL Among Top Beneficiaries
Research through 2026-11-03 · 2025-06-27 to 2026-11-03 · 65 source reports.
| Range | Prices through |
|---|---|
| 1W | |
| 1M | |
| 3M | |
| 6M | |
| 1Y | |
| 2Y | |
| 3Y |
Research through 2026-11-03. Artifact validity does not establish factual accuracy or research freshness.