AI Chip Race Splits as Custom Silicon Challenges the GPU Model

AI Spending Is Creating More Than One Semiconductor Growth Story

The AI semiconductor market is no longer centered on a single type of processor. Broadcom (AVGO) and Marvell Technology (MRVL) are benefiting from hyperscalers developing custom accelerators and networking infrastructure, while Nvidia (NVDA) and Advanced Micro Devices (AMD) remain central to programmable AI computing.

Recent results show strong growth across all four companies, but their scale, valuations and exposure to future customer deployments differ substantially. Broadcom and Nvidia are already generating AI-related revenue at enormous scale, while Marvell and AMD carry higher earnings multiples alongside expectations for future expansion.

Key Points

• Broadcom and Marvell are benefiting from hyperscaler demand for custom AI accelerators and networking infrastructure.
• Nvidia remains the largest accelerated-computing business in the group, while AMD is expanding as an alternative across GPUs and CPUs.
• The supplied comparisons show AMD and Marvell trading at substantially higher forward earnings multiples than Nvidia and Broadcom.

Broadcom vs. Marvell: Two Different Custom Silicon Stories

Broadcom and Marvell both participate in the shift toward chips designed specifically for hyperscalers' AI workloads. The biggest difference is how much of that opportunity has already translated into financial results.

Broadcom reported fiscal third-quarter revenue of $29.59 billion, including $16.7 billion of AI semiconductor revenue, which increased 221% year over year. Free cash flow reached $13.7 billion, equal to 46% of revenue. Management expects fourth-quarter AI semiconductor revenue of approximately $21.7 billion.

Marvell operates from a considerably smaller base. Fiscal second-quarter revenue reached $2.74 billion, with data-center revenue of $2.17 billion, up 46% and representing 79% of total revenue. Management raised its fiscal 2027 data-center growth outlook to approximately 60% and expects fiscal 2028 data-center revenue to grow more than 60%.

*Based on the valuation comparison provided in the content input.

The valuation gap is significant. The supplied figures place Marvell at more than twice Broadcom's forward earnings multiple even though Broadcom currently generates substantially greater revenue and cash flow.

The companies also face different execution questions. Broadcom's custom-chip growth increasingly depends on a relatively small group of hyperscaler programs, while Marvell's smaller base makes the timing and scale of individual customer deployments particularly important.

Nvidia vs. AMD: Scale Meets the Challenger

Nvidia and AMD provide a different comparison because both compete more directly in accelerated computing.

Nvidia generated $96.2 billion of quarterly revenue, up 106%, including $89 billion from Data Center, which increased 117%. Gross margin reached 75%, and the company projected third-quarter revenue of $108 billion.

AMD reported second-quarter revenue of $11.5 billion, up 50%, while Data Center revenue more than doubled to $6.7 billion. Non-GAAP gross margin reached 56%, and management expects approximately $13 billion of third-quarter revenue.

AMD's opportunity extends across EPYC server CPUs, Instinct accelerators and its Helios rack-level offering. Anthropic has agreed to deploy up to two gigawatts of MI450-series systems, with the first gigawatt expected to begin deployment in the first half of 2027.

Nvidia operates from a much larger base and combines its accelerators with CUDA and a broader full-stack platform. The supplied numbers also show Nvidia's Data Center business growing slightly faster than AMD's despite the substantial difference in scale.

Valuation creates another distinction. On the common S&P Global basis supplied, AMD traded at approximately 55.6 times forward earnings compared with 18.7 times for Nvidia. A separate comparison using 2027 earnings estimates also showed a substantial premium for AMD.

That means the financial profiles have diverged along with the technology competition: AMD is expanding from a smaller base as an alternative supplier, while Nvidia continues to generate substantially greater revenue and higher margins.

What Does the Four-Way Comparison Reveal?

Together, the four companies show how AI infrastructure spending is spreading across different parts of the semiconductor market.

Nvidia represents the largest existing accelerated-computing business among the group. AMD is expanding as an alternative across GPUs, CPUs and rack-level systems. Broadcom has developed a large custom-accelerator and networking business, while Marvell combines custom silicon with optical connectivity, switching and networking.

Broadcom's results show that custom AI silicon is already capable of producing revenue at significant scale. Its AI semiconductor revenue reached $16.7 billion in the latest reported quarter and is expected to increase to $21.7 billion in the fourth quarter.

Marvell provides another view of the same trend. Data centers already account for 79% of its revenue, and its expanded custom program with Google includes warrants tied to revenue milestones.

The valuation pattern provides another important distinction. In the supplied comparisons, Nvidia and Broadcom trade below 20 to roughly 20 times forward earnings, while AMD and Marvell trade at substantially higher multiples.

Those valuations do not determine future stock performance. They do show that the market is assigning very different prices to current earnings and anticipated future growth across the four companies.

What It Means for Investors

The AI chip market is becoming more diversified as hyperscalers combine general-purpose accelerators with processors designed around specific workloads.

That broadens the semiconductor story beyond the traditional Nvidia-versus-AMD competition. Broadcom and Marvell provide exposure to another important part of the buildout: custom accelerators and the networking infrastructure required to connect increasingly large AI systems.

The four companies also enter this expansion from very different financial positions. Nvidia and Broadcom combine greater current scale with lower forward earnings multiples in the supplied comparisons. AMD and Marvell operate from smaller bases but carry higher valuations as their respective AI programs expand.

Execution therefore becomes increasingly important. Customer deployments, revenue growth, margins, cash generation and the timing of large custom-chip programs provide measurable ways to follow how these different AI strategies develop.

Conclusion

AI infrastructure spending is creating several semiconductor growth paths rather than a single chip race.

Nvidia remains the largest accelerated-computing business in this comparison. AMD is expanding its position as an alternative platform. Broadcom is generating substantial revenue from custom AI silicon and networking, while Marvell is building its own custom-silicon and connectivity business around growing hyperscaler demand.

The larger market shift is the growing coexistence of general-purpose accelerators and custom processors. How hyperscalers divide future workloads between those architectures will help determine where the next phase of AI semiconductor growth appears.

FAQs

What is the main difference between Broadcom and Marvell?

Both companies have exposure to custom AI silicon and networking, but Broadcom currently operates at substantially greater revenue and cash-flow scale. Marvell has a smaller revenue base with significant exposure to future hyperscaler program ramps.

What is the main difference between Nvidia and AMD?

Nvidia currently generates substantially more Data Center revenue and has higher gross margins, while AMD is expanding as an alternative across GPUs, CPUs and rack-level AI systems.

Why are custom AI chips becoming more important?

The supplied material describes hyperscalers increasingly using silicon optimized for their specific workloads. Broadcom and Marvell both participate in custom silicon and the networking infrastructure supporting those systems.

How do the valuations of the four companies compare?

The supplied comparisons place Nvidia and Broadcom at below 20 to roughly 20 times forward earnings, while AMD and Marvell trade at substantially higher multiples. The exact figures vary depending on the earnings estimates used.

Are Nvidia, AMD, Broadcom and Marvell competing in exactly the same markets?

No. Nvidia and AMD are more directly exposed to programmable accelerated computing, while Broadcom and Marvell have significant exposure to custom hyperscaler silicon and networking. Their businesses increasingly overlap as AI infrastructure expands.

This article was created with AI assistance and reviewed by an editor. For details, please refer to our Terms of Use.

Go Beyond the Market Brief with Market Edge

Follow SharperTrades’ complete approach to trading and investing, combining active trade opportunities through Block Orders, long-term research through Stock Investor, and structured market education through the Swing Trading Masterclass. Try Market Edge for $19 your first month →

Explore Research with Stock Investor

Stock Investor is SharperTrades’ platform for long-term investing research and portfolio management. Members receive research reports, portfolio updates, conviction tracking, and in-depth analysis designed to support disciplined investment decisions.

Explore Active Trading & Income Strategies

Block Orders tracks institutional activity and highlights active trade setups and price behavior across long and short opportunities.

For options-focused traders, Essential Option Income provides a structured approach to options income strategies, while Pro Option Trader offers a broader range of options strategies and trade opportunities.

Think More Clearly with SteadyCapital

SteadyCapital is SharperTrades’ decision-support system for long-term investors, built around the SteadyCapital Method™. Review investment ideas, challenge assumptions, evaluate valuation and risk, compare companies, and think through important buy, hold, add, trim, or sell decisions before you act.

Risk Disclosure

All content is provided for educational purposes only and does not constitute investment advice. Trading involves risk, and past performance is not indicative of future results. Please review our full Risk Disclosure for additional information.