MRVL vs AVGO Stock Comparison: AI Score, Valuation, Performance and Upside
MRVL (Marvell Technology) and AVGO (Broadcom) are both significant semiconductor players in AI data center infrastructure — Broadcom is larger and more established with market-dominant networking switch chips, a leading custom ASIC business (for Google and Meta), and VMware software providing recurring revenue, while Marvell is a focused data infrastructure specialist reorienting to custom AI ASICs and electro-optics for hyperscalers. Broadcom is the diversified AI semiconductor leader; Marvell is the pure-play data infrastructure AI semiconductor growth story.
MRVL vs AVGO is pure-play AI semiconductor growth with hyperscaler custom ASIC exposure (Marvell's repositioned data infrastructure business betting on custom chip programs and optical networking for AI) versus diversified AI semiconductor and software giant (Broadcom's dominant networking chips, established custom ASIC relationships, and VMware software providing multiple revenue streams) — focused AI bet versus diversified semiconductor/software leader.
AVGO holds the edge across 3 of 5 key metrics in this comparison. MRVL has delivered stronger 1-year price return (+315.50% vs +64.96%), though AVGO trades at the lower forward P/E (19.74x vs 50.31x). Analyst consensus implies meaningfully more upside for AVGO (+36.64%) than for MRVL (-23.13%).
- →Want a more concentrated bet on hyperscaler custom AI ASIC growth — Marvell's revenue is increasingly dominated by AI data infrastructure, providing higher sensitivity to AI semiconductor demand
- →Value Marvell's custom ASIC and optical DSP capabilities as positioning it to capture significant value from the growing hyperscaler preference for custom AI silicon over general-purpose GPUs
- →Accept the execution risk and revenue lumpiness of custom ASIC programs in exchange for the potentially higher growth rate of a smaller, more focused AI semiconductor company
- →Want the most diversified AI semiconductor exposure from the largest non-NVIDIA semiconductor company in the AI infrastructure stack — Broadcom's networking chips are essential for every AI cluster
- →Value Broadcom's VMware software as providing high-margin recurring revenue that cushions semiconductor cycle impacts while also providing software infrastructure for enterprise AI deployments
- →Prefer Broadcom's market-leading competitive position in networking switches and established custom ASIC relationships as creating durable competitive moats versus smaller AI chip competitors
| Metric | MRVL | AVGO |
|---|---|---|
| AI score | 66.2 | 74.5 |
| AI rank | #55 | #24 |
| Latest close | $310.58 | $411.35 |
| 1M return | +76.20% | +0.07% |
| 6M return | +280.61% | +26.17% |
| 1Y return | +315.50% | +64.96% |
How much would $10,000 be worth today if invested at the start of each period, with all dividends reinvested?
| Period | MRVL | AVGO |
|---|---|---|
| 1Y ago | $41.66K (+316.6%) started 2025-06-18 | $16.37K (+63.7%) started 2025-06-18 |
| 5Y ago | $59.5K (+495.0%) started 2021-06-18 | $106.02K (+960.2%) started 2021-06-21 |
| 10Y ago | $355.4K (+3454.0%) started 2016-06-20 | $452.71K (+4427.1%) started 2016-06-20 |
Hypothetical — past performance does not guarantee future results.
| Metric | MRVL | AVGO |
|---|---|---|
| Market cap | $271.7B | $1.82T |
| Trailing P/E | 107.10 | 63.68 |
| Forward P/E | 50.31 | 19.74 |
| Price/Sales | 31.17 | N/A |
| EV/Revenue | 28.13 | 24.69 |
| Analyst target | $238.75 | $522.06 |
| Target upside | -23.13% | +36.64% |
| Metric | MRVL | AVGO |
|---|---|---|
| Revenue growth | 27.60% | 47.90% |
| Earnings growth | -80.40% | 85.40% |
| EPS growth | -80.40% | +85.40% |
| FCF margin | +26.04% | +36.06% |
| Operating margin | N/A | 48.99% |
| Profit margin | 28.99% | 38.85% |
| ROIC proxy | 16.03% | 37.28% |
| Return on equity | 16.03% | 37.28% |
| Dividend yield | 0.09% | 0.68% |
| Beta | 2.28 | 1.43 |
| Debt/equity | 28.97 | 74.02 |
| Current ratio | 3.28 | 2.24 |
| Quick ratio | 2.51 | 1.93 |
Lower drawdown and smaller single-period drops generally indicate a smoother ride, though they do not guarantee lower future risk.
| Period | Metric | MRVL | AVGO |
|---|---|---|---|
| 1Y | Growth | +315.50% | +63.71% |
| CAGR | +315.91% | +63.83% | |
| Sharpe ratio | 2.27 | 1.21 | |
| Max drawdown | 26.36% | 28.95% | |
| Max daily drop | 18.59% | 12.59% | |
| Max wkly drop | 18.00% | 22.35% | |
| 5Y | Growth | +483.49% | +863.18% |
| CAGR | +42.31% | +57.42% | |
| Sharpe ratio | 0.80 | 1.16 | |
| Max drawdown | 61.88% | 41.15% | |
| Max daily drop | 19.81% | 17.40% | |
| Max wkly drop | 23.97% | 22.35% | |
| 10Y | Growth | +3194.79% | +3286.54% |
| CAGR | +41.87% | +42.25% | |
| Sharpe ratio | 0.84 | 0.98 | |
| Max drawdown | 61.88% | 48.30% | |
| Max daily drop | 19.81% | 19.91% | |
| Max wkly drop | 23.97% | 31.75% |
| Category | MRVL | AVGO |
|---|---|---|
| Company | Marvell Technology, Inc. | Broadcom Inc. |
| Sector | Technology - Semiconductors | Technology |
| Industry | N/A | N/A |
| Core business | Marvell Technology is a fabless semiconductor company designing chips for data infrastructure — providing custom AI accelerator ASICs (co-designed with Google, Amazon, and Microsoft), optical digital signal processors for data center networking, PCIe switches, and storage controllers. Marvell has repositioned from a general-purpose chip company to a data infrastructure AI semiconductor specialist. | Broadcom is a diversified semiconductor and infrastructure software company — providing networking switch chips (dominant in data center Ethernet switching with Tomahawk and Trident series), custom AI accelerator ASICs (Google's TPUs, Meta's AI chips), broadband semiconductors, storage controllers, wireless chips for Apple iPhones, and after acquiring VMware, enterprise software (VMware cloud infrastructure). Broadcom is one of the largest semiconductor companies globally. |
| Investor focus | Investors track Marvell's AI revenue (custom ASIC + electro-optics for AI data centers), the ramp of custom ASIC programs with hyperscaler customers, 5G semiconductor recovery, gross margin improvement as AI mix improves, and Marvell's TAM opportunity in the hyperscaler custom silicon market. | Investors track Broadcom's AI semiconductor revenue (networking chips + custom ASICs), VMware software integration progress and revenue, gross margins (very high due to IP-rich products and software), capital return through dividends and buybacks, and the breakdown of AI versus non-AI semiconductor revenue. |
- →Custom AI ASIC design capability — Marvell's ability to design full custom AI training and inference chips for hyperscalers (working under NDA with multiple large cloud providers) positions it in the fastest-growing semiconductor market
- →Electro-optics for AI networking — Marvell's optical DSP chips enable high-bandwidth data center networking for AI clusters; as GPU interconnect demands increase, Marvell's networking silicon is a key enabler
- →Repositioned to AI data infrastructure — Marvell has divested non-core businesses and focused R&D on AI data center silicon, aligning the company's growth engine with the most powerful technology trend
- →Dominant networking switching chip market position — Broadcom's Tomahawk and Trident series control the majority of data center Ethernet switching; as AI clusters require massive networking bandwidth, Broadcom's switch chip upgrades drive recurring revenue from hyperscalers
- →Custom ASIC market leadership — Broadcom designs custom AI chips for Google (TPU), Meta, and other major hyperscalers; its custom ASIC design track record and production scale give it a strong competitive moat
- →VMware software added a high-margin recurring revenue stream — Broadcom's VMware acquisition added ~$8B of infrastructure software revenue with very high margins, providing an earnings buffer when semiconductor cycles weaken
- →Customer concentration in custom ASIC — Marvell's custom AI ASIC business is concentrated among a small number of hyperscaler customers; losing a key ASIC program would have material revenue impact
- →Long design cycles for custom ASICs — hyperscaler AI chip designs take 3-5 years from engagement to volume production; Marvell's revenue ramp from custom ASICs is lumpy and difficult to predict quarter-to-quarter
- →Competition from Broadcom in custom AI ASICs — Broadcom is the market leader in custom AI chips (designing for Google TPUs and Meta's MTIA); Marvell competes in the same hyperscaler custom silicon market
- →Customer concentration in networking chips — Broadcom's networking chip revenue is concentrated among a small number of hyperscalers; changes in hyperscaler architectural choices (such as NVIDIA's InfiniBand winning share from Ethernet) affect Broadcom's networking volumes
- →VMware integration execution — Broadcom significantly raised VMware's pricing after acquisition, which has driven some customers to evaluate alternatives; managing VMware customer retention while maximizing per-seat revenue is an ongoing execution challenge
- →NVIDIA competition for AI networking — NVIDIA's acquisition of Mellanox and push into AI Ethernet networking (Spectrum-X) competes with Broadcom's networking switching business for AI cluster fabric revenue
Want deeper AI forecasts?
This comparison page is public and free forever. Subscribers can unlock saved watchlists, full AI rankings, detailed forecasts, and interactive analysis tools.