Custom AI Chips Challenge Nvidia and AMD as Semiconductor Competition Intensifies

The AI chip race is expanding beyond Nvidia and AMD

Nvidia (NVDA) and Advanced Micro Devices (AMD) continue to post rapid growth as demand for AI computing expands. Nvidia’s fiscal first-quarter revenue rose 85% year over year to $81.6 billion, while AMD’s latest quarterly revenue increased 50% to $11.5 billion.

Yet the competitive landscape is becoming more complex. Alphabet (GOOGL) and Amazon (AMZN) are developing chips optimized for their own workloads, Marvell Technology (MRVL) and Broadcom (AVGO) are helping enable custom accelerators, and startups including Etched and Cerebras are targeting specialized parts of the AI-compute market.

Key Points

• Nvidia remains the much larger AI semiconductor company, with fiscal first-quarter revenue up 85% to $81.6 billion and data-center revenue reaching $75.2 billion.
• AMD is expanding its AI challenge through its Helios rack-scale platform, while its data-center revenue reached approximately $6.7 billion in the second quarter.
• Custom chips from major cloud companies and specialized processors from startups are widening the AI semiconductor market beyond traditional GPUs.

Nvidia Faces a Broader Competitive Field

Nvidia enters its Aug. 26 fiscal second-quarter report from a position of considerable scale. Wall Street expects $92 billion in sales and earnings of $2.08 per share, roughly double year-ago levels.

The company’s fiscal first-quarter revenue increased 85% to $81.6 billion, with data-center revenue reaching $75.2 billion. Nvidia’s data-center business grew 92%, illustrating the scale of AI infrastructure demand flowing through the company.

But competitors are attacking different parts of the AI-compute market.

Etched, an AI-chip startup recently valued at approximately $21 billion following a $700 million funding round, is focusing on specialized AI hardware. Michael Burry highlighted the company as potential “serious competition” for Nvidia after reports about its performance and deployment speed. Etched said it took 44 days to get its chips running inference workloads, compared with a process that typically takes six months or longer.

Around 15% of Etched’s workforce consists of former Nvidia employees. Burry also cited information indicating that the startup’s technology demonstrated nearly 10 times the performance at a lower cost per die.

Those claims still represent a competitive proposition rather than evidence that Nvidia’s position has already been displaced. Nvidia’s existing advantages include its CUDA software ecosystem, networking capabilities and years of developer adoption.

Bloomberg Intelligence forecasts Nvidia will retain at least 70% of AI training despite increasing competition.

How Is AMD Challenging Nvidia?

AMD is emerging as the most direct large-scale competitor to Nvidia.

AMD’s second-quarter revenue increased 50% year over year to $11.5 billion, while diluted earnings per share increased 156%. Data-center revenue reached approximately $6.7 billion, and data-center growth of 107% exceeded Nvidia’s 92% growth rate from its most recently reported quarter.

The company is also pushing beyond individual accelerators. AMD’s Helios rack-scale platform combines 72 MI455X accelerators with EPYC CPUs and AMD networking, creating an integrated AI infrastructure alternative to Nvidia’s complete-system approach.

AMD claims Helios can deliver up to 30% more inference tokens per dollar than Nvidia’s Vera Rubin NVL72 on a selected workload. Meta Platforms (META) plans large-scale AMD deployments, while OpenAI expects to bring Helios online in late 2026.

The challenge for AMD remains software. Nvidia’s CUDA ecosystem and entrenched developer adoption make switching platforms costly and complicated.

Valuation also separates the two stocks. Based on the projections included in the available data, AMD trades at 33 times next fiscal year’s earnings compared with 17.6 times for Nvidia. AMD shares have risen about 125% in 2026, compared with roughly 20% for Nvidia, putting greater expectations into AMD’s valuation.

Custom Silicon Signals a New Phase of AI Competition

The AI semiconductor contest is no longer limited to Nvidia versus AMD.

Some of the largest buyers of AI accelerators are developing their own processors. Alphabet and Amazon are designing chips specifically for their workloads, potentially reducing reliance on general-purpose accelerators for selected applications.

Google’s TPU 8i reportedly provides 80% better performance per dollar than its previous generation, while TPU 8t targets large-scale training. Amazon Web Services says Trainium3 systems can provide up to 4.4 times the performance of Trainium2 while lowering training and inference costs.

Marvell is becoming an important participant in this shift. Warrants associated with its Google relationship can fully vest if qualifying Google purchases reach approximately $120 billion through fiscal 2033. That figure is not a guaranteed spending commitment, but it illustrates the potential scale attached to custom silicon.

Broadcom is another enabler of purpose-built accelerators. OpenAI has shown an inference chip developed with Broadcom, while Google is expanding custom-chip production with Marvell.

Cerebras is taking another approach through wafer-scale processors aimed at high-speed inference and has an OpenAI relationship.

Together, these developments show how AI semiconductor competition is spreading across GPUs, rack-scale systems, custom accelerators and specialized inference processors.

What It Means for Investors

The central issue for AI semiconductor investors is increasingly how a rapidly expanding market will be divided among competing architectures.

Nvidia continues to operate at substantially greater scale than AMD and maintains important advantages through CUDA, networking and its broader computing ecosystem. Its upcoming fiscal second-quarter results will provide another measure of AI demand, with Wall Street expecting $92 billion in revenue.

AMD, meanwhile, is growing rapidly in data centers and building a broader AI infrastructure platform around Helios. Its challenge is to translate that growth into greater adoption while competing against Nvidia’s established software ecosystem.

At the same time, hyperscalers are becoming both customers and competitors. Google, Amazon and other large technology companies increasingly have an economic incentive to develop processors tailored to their own workloads, creating opportunities for companies such as Marvell and Broadcom.

Nvidia is also attempting to preserve opportunities in China. The company reportedly plans small-batch shipments of a modified Language Processing Unit to Chinese customers by year-end. The processor complies with U.S. export restrictions and has been adapted through software to work with processors available in China. Several Chinese customers have already submitted orders.

That effort comes as Chinese AI companies face shortages of inference capacity and domestic production capacity has been heavily committed through 2027.

Conclusion

AI semiconductor demand remains a major growth driver for Nvidia and AMD, but the competitive structure surrounding that demand is changing.

Nvidia remains the largest player represented in the available data, with $81.6 billion in fiscal first-quarter revenue and $75.2 billion coming from data centers. AMD is expanding rapidly, with second-quarter revenue reaching $11.5 billion and data-center growth of 107%.

The next layer of competition is coming from custom silicon and specialized processors. Google and Amazon are developing their own chips, Marvell and Broadcom are helping companies build custom accelerators, and startups such as Etched and Cerebras are targeting specific AI workloads.

The result is an AI chip market increasingly defined not by a single Nvidia-versus-AMD contest, but by competition across hardware, software, networking, custom silicon and complete AI computing systems.

FAQs

Why is Nvidia facing more competition in AI chips?

Competition is expanding from AMD to custom chips developed by major cloud companies and specialized processors from startups. Google and Amazon are designing chips for their own workloads, while Etched and Cerebras are targeting specialized AI applications.

How fast is Nvidia’s data-center business growing?

Nvidia’s data-center revenue increased 92% in its fiscal first quarter and reached $75.2 billion. Total company revenue increased 85% year over year to $81.6 billion.

How is AMD competing with Nvidia in AI?

AMD is expanding through its Instinct accelerators and Helios rack-scale platform, which combines 72 MI455X accelerators with EPYC CPUs and AMD networking. AMD’s data-center business grew 107% in its latest reported quarter.

Why do custom AI chips matter for Nvidia and AMD?

Some of Nvidia and AMD’s largest potential customers are developing processors optimized for their own workloads. Google and Amazon are among the companies pursuing custom silicon, while Marvell and Broadcom are helping enable purpose-built accelerators.

What should investors watch next for Nvidia?

Nvidia reports fiscal second-quarter results on Aug. 26. Wall Street expects $92 billion in sales and $2.08 in earnings per share, making the report an important update on AI demand and Nvidia’s growth.

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.