Could NVIDIA’s reign in AI chips start to crumble?

As the US tech market reaches a new high, Marc El-Lazidi, chief investment officer at Jesmond Mizzi Financial Advisors, explores the changing dynamics of the AI supply chain, from NVIDIA’s dominance to hyperscalers developing their own chips and Chinese firms waiting in the wings

What exactly is the accelerated computing supply chain, and where does NVIDIA fit in?

AI started in the cloud, but now it’s moving into every part of the real world: enterprise IT, manufacturing, robotics and even autonomous vehicles. That’s created a massive demand for infrastructure.

At the base of this ecosystem are the chipmakers, NVIDIA being the most prominent. Then come the hyperscalers like Amazon AWS, Microsoft Azure and Google Cloud. Next are the system integrators and consultancies, think Accenture or Capgemini.

Finally, at the top, you have the software players selling AI solutions to end users. It’s a complex but critical supply chain that’s rapidly evolving.

NVIDIA has gone from a gaming GPU brand to the engine of the AI revolution. What’s driving that dominance?

Before the rise of large language models, NVIDIA was known mostly to gamers. But they’ve built a full stack: not just chips, but also the programming model and the software layer that sits on top. That’s extremely rare. Enterprises love this level of integration. It’s plug-and-play for AI development. And because of that, NVIDIA is feeding the AI revolution from the bottom up. It has gone from obscure GPU maker to the engine room of generative AI.

How do profits move across the AI supply chain?

The closer you are to the start of the supply chain, the higher the margin and monopolistic position. NVIDIA, of course, has impressive margins, but so does Taiwan Semiconductor.

ASML is the crown jewel of Europe, but beyond that, Europe doesn’t have many players in this sector or in this position within the supply chain.

Hyperscalers also enjoy strong margins, as everything has moved to the cloud. However, further along the chain, in system integration and consulting, profit margins begin to diminish.

Who in the supply chain faces the greatest risk at the moment and why?

Consulting firms are in trouble. Accenture, Capgemini, even Deloitte in some cases, are being challenged. Why? Because the hyperscalers are cutting them out. Generative AI and low-code tools are making it easier to build systems without heavy consulting support. Even financial firms like ours are seeing services directly offered by companies like Microsoft. The traditional middlemen are becoming less relevant.

Can hyperscalers like Amazon and Microsoft truly bypass NVIDIA by designing their own chips?

That’s already happening. Microsoft is building its own AI accelerators like the Maia chip to control the AI stack end-to-end. Amazon is doing the same with Graviton 4. The industry is shifting from CPUs to GPUs, and hyperscalers want to own that technology. Their argument is: why rely on NVIDIA when we can build our own chips? This is a major challenge for NVIDIA, particularly over the next years.

And what about NVIDIA? Is it also moving up the stack to compete directly with hyperscalers and integrators?

Yes, and this is where it gets interesting. NVIDIA is not just a hardware company anymore it’s also a software powerhouse. They’re developing platforms for autonomous driving, enterprise AI, and more. They could go direct to B2B clients, bypassing both hyperscalers and integrators.

Jensen Huang, the CEO, is a visionary. He could well become the next tech figurehead, perhaps even replacing Elon Musk in Trump’s orbit. He’s smart, strategic and politically savvy.

What role does China play in this new ecosystem and should Western investors be concerned?

China is the wildcard. Companies like Huawei and Alibaba Cloud are developing their own alternatives at high speed. There’s intense competition not just domestically but globally. Just like Chinese EVs took Europe by surprise, Chinese AI firms could do the same. And not all of them rely heavily on GPUs; some are finding low-cost ways to deploy LLMs.

The West should not underestimate this. I call it a “Deep Seek” moment when a challenger suddenly becomes dominant. Investors need to look beyond the usual suspects.

Finally, what’s the takeaway for investors trying to navigate this fast-evolving AI landscape?

Understand the stack. Know where the power lies and how it’s shifting. Don’t assume that today’s leaders will still be on top tomorrow. NVIDIA is being challenged from above and below. Hyperscalers are facing their own disintermediation. Consulting firms are being squeezed. And somewhere out there, a Chinese start-up could be building the next killer platform. It’s an arms race and software is not just consuming the world, it’s redefining who gets to consume.

This interview is issued by Jesmond Mizzi Financial Advisors Limited and does not intend to give investment advice and the contents therein should not be construed as such. The company is licensed to conduct investment services by the MFSA, under the Investment Services Act and is a Member of the Malta Stock Exchange. Investors should remember that past performance is no guide to future performance and that the value of investments may go down as well as up. For further information, contact Jesmond Mizzi Financial Advisors Limited of 67, Level 3, South Street, Valletta, on 2122 4410, or e-mail info@jesmondmizzi.com.