Financial Report Preview | Cloud Giants Launch "AI Cost Revolution," The Era of ASICs Has Arrived! Maweier Technology (MRVL.US) Performance Soaring Soon

CNBC Finance APP has learned that, focusing on large AI data center customized AI chips (i.e., AI ASICs), and as one of the largest partners in Amazon AWS Trainium series AI ASICs, Marvell Technology (MRVL.US) will release its earnings report after the US stock market closes on March 5th Eastern Time. Wall Street analysts unanimously expect that, under the wave of AI inference and the trend of embedding AI large models into enterprise operations through “micro-training,” more cost-effective AI ASICs will pose a strong challenge to Nvidia’s nearly 90% market share in AI chips. Therefore, analysts predict that AI ASIC leaders Marvell and the larger ASIC giant Broadcom (AVGO.US) will both achieve strong performance growth, and management is likely to provide a robust earnings outlook.

In its recent fiscal third quarter 2026 financial report (for the period ending November 1, 2025), Marvell achieved approximately $2.075 billion in net revenue, up about 37% year-over-year and slightly exceeding market expectations. Adjusted earnings per share also surpassed Wall Street forecasts. This strong growth in the third quarter reflects the explosive expansion of demand for customized AI ASICs driven by cloud computing leaders’ new construction and expansion of AI data centers.

According to data compiled by Zacks Investment Research, Wall Street analysts expect Marvell’s fourth fiscal quarter adjusted EPS to be around $0.79, representing a potential 31.7% increase from the same period last year; revenue is expected to be about $2.21 billion, a significant year-over-year increase of 21% on a strong prior base. For the full fiscal year, analysts generally expect EPS of $2.84, an 80.9% increase from the previous year; revenue forecasts for this year and next are $8.18 billion and $10 billion, respectively, implying growth of 41.8% and 22.3%.

Additionally, after completing the acquisition of optical interconnect technology company Celestial AI, Marvell will further enhance its technical capabilities in high-bandwidth, low-latency AI data center infrastructure. This acquisition is expected to gradually contribute to revenue growth over the next few years and help expand the company’s share in the AI ecosystem. In its previous earnings report, besides strong Q3 results and optimistic quarterly guidance, the company also announced a $3.25 billion acquisition of the startup Celestial AI, which focuses on optical interconnect I/O chips, to strengthen its network product portfolio.

Marvell CEO Matt Murphy stated during the earnings call that Celestial’s technology will be integrated into Marvell’s next-generation silicon photonics infrastructure hardware products, which are expected to create a new, potentially $10 billion super blue ocean market.

Murphy and other executives also indicated that they expect to start seeing significant revenue contributions from Celestial AI from the second half of fiscal 2028, with an annualized revenue target of about $500 million by Q4 FY2028, and doubling this to $1 billion by Q4 FY2029.

Market concerns about Nvidia’s prospects are justified

The global surge in generative AI has accelerated the development of AI chips by cloud giants and chip leaders, all racing to design the fastest and most energy-efficient AI infrastructure clusters for advanced large AI data centers. Marvell and its biggest competitor Broadcom mainly focus on leveraging their advantages in high-speed interconnects and chip IP to collaborate with cloud giants like Amazon, Google, and Microsoft to create AI ASIC clusters tailored to their data center needs. This ASIC business has become a vital part of both companies’ portfolios; for example, Google’s TPU AI clusters are a typical example of AI ASIC technology.

Amazon’s new head of AI infrastructure, Peter DeSantis, said in a media interview last Friday: “If we can build models on our own self-developed AI chips, we can do so at a fraction of the cost of pure AI large model providers.”

DeSantis added, “Building ultra-large-scale AI data centers does involve some cost issues. If we ultimately want AI to change everything, the costs must be different.”

The market generally believes that Nvidia (NVDA.US), the “AI chip super-giant,” still holds the majority of the core AI infrastructure market—namely, the AI chip market. The chip giant led by Jensen Huang recently announced quarterly results for FY2026 Q4 and guidance for the next quarter that far exceeded expectations. However, its stock price fell sharply by 5% on Thursday, mainly due to increasing concerns over the recent moves by hyperscalers (large-scale cloud providers) to launch more cost-effective AI ASIC chips based on self-developed models, which signals a potential long-term risk to Nvidia’s dominant position in the global AI infrastructure and AI chip markets.

Undoubtedly, as Anthropic, dubbed an “OpenAI rival,” plans to spend hundreds of billions of dollars to purchase 1 million TPU chips, and Meta (Facebook’s parent company) considers spending billions of dollars in 2026 or 2027 to buy Google’s TPU AI infrastructure for its massive AI data centers, along with Amazon’s announcement to develop large models using Trainium and Inferentia, the market’s concern about Nvidia’s future is justified.

The wave of AI inference is coming, and Nvidia’s “monopoly share” faces fierce challenges

Unquestionably, major constraints related to cost and power consumption are pushing Microsoft, Amazon, Google, and Meta to develop their own AI ASICs for internal cloud systems, aiming for better cost-performance and energy efficiency in AI clusters.

Building ultra-large AI data centers, like the “Stargate,” is extremely expensive. Therefore, tech giants increasingly demand that AI compute systems be more economical and energy-efficient, striving to optimize “cost per token” and “output per watt.” The prosperity of AI ASIC technology is now a reality.

Furthermore, the long-term demand for Nvidia’s advanced AI GPU clusters based on the Blackwell architecture remains high, but costs are rising, and supply chain bottlenecks and delivery schedules are constraining supply. Self-developed AI ASICs can provide “second-curve capacity,” giving companies more leverage in procurement negotiations, pricing, and cloud service margins. Additionally, cloud giants like Google and Microsoft can co-design “chips—interconnects—systems—compilers/runtimes—scheduling—monitoring/reliability,” improving infrastructure utilization and reducing TCO.

Nvidia’s AI GPUs, which almost monopolize AI training, require more versatile AI compute clusters and rapid iteration of the entire system. On the inference side, after the deployment of cutting-edge AI models at scale, the focus shifts to unit token cost, latency, and energy efficiency. For example, Google positions its Ironwood TPU as “born for the AI inference era,” emphasizing performance, energy efficiency, and scalability. Meanwhile, Amazon’s latest moves suggest that AI ASICs have strong potential for training large models.

The AI ASIC compute system will undoubtedly continue to erode Nvidia’s monopoly premium and market share in the medium to long term—not through linear replacement of GPUs, but by reshaping industry profit pools and customer procurement structures. The core reason is that, in the inference era, the competition is no longer just about “peak compute power,” but also about “cost per token,” power consumption, memory bandwidth utilization, interconnect efficiency, and total cost of ownership after hardware-software co-design. In these metrics, ASICs tailored for specific workloads—optimized data flow, compilers, and interconnects—are inherently more cost-effective than general-purpose GPUs.

However, for Nvidia and AMD, this largely means that marginal pressure is real, likely manifesting as reduced bargaining power, market share erosion, and compressed valuation premiums, rather than a complete collapse of demand. AI ASICs will continue to challenge Nvidia’s GPU dominance in AI inference during the super wave, but more as a reshaping of industry profit pools and customer procurement patterns rather than invalidating GPU expansion logic.

AWS explicitly states that Trainium and Inferentia are dedicated accelerators for generative AI training and inference, with Trainium 2 offering about 30–40% better price-performance compared to its AI GPU cloud instances. Google has also publicly announced that Gemini 2.0’s training and inference are 100% run on TPUs. This indicates that “large cloud providers using self-developed ASICs for core model training and inference” is no longer just a concept but is entering a reproducible industrial phase.