Is the AI-chip future secure? How putting all the eggs in the Nvidia basket unravelled into market panic after Jensen Huang’s exit

In an ideal world, a company as large as Nvidia might be called a "monopoly"—a $4 trillion "monopoly". But in the world of AI and chipmaking, seldom is anything ideal. Jensen Huang’s resignation as Nvidia’s chief executive does not change the fact that AI computing remains the fastest-growing part of the chip industry. And it has also revealed how dependent a single industry has become on one company, one manufacturing partner and a small group of suppliers.

Nvidia’s board announced on November 12 that Huang had resigned as president, chief executive and director with immediate effect. The chipmaker’s stock is now down 27 per cent over the past month. Chief financial officer Colette Kress has been brought in to serve as interim CEO while the board searches for a permanent successor to the once-confident, always-flamboyant leather-jacket-donning Huang.

On the onset, this looks like a disaster. It is a PR nightmare, to be sure. But in the case of chips and big tech, it is always the numbers that speak the truth.

Nvidia is set to post its third-quarter earnings the coming week. Its last reported second-quarter figures saw Nvidia posting revenue of $96.2 billion, up 106 per cent year-on-year. Its data-centre revenue was $89 billion, up 117 per cent from a year ago. GAAP diluted earnings per share (how the Americans measure actual net profit) were $2.46, up 128 per cent year-on-year.

In contrast, Nvidia’s once-head-to-head competitor, AMD, posted a Q2 revenue of $11.5 billion, up 50 per cent. Its data-centre revenue rose 107 per cent to $6.7 billion. GAAP diluted EPS was $1.38, up 156 per cent from the previous year.

The other famous chipmaker, Intel, saw its Q2 revenue increase by 25 per cent to $16.1 billion. Data-centre and AI (DCAI) revenue was $6.3 billion, up 59 per cent. But Intel reported a GAAP loss per share of $2.16.

All of these were once famous companies with their primary product a chip you could touch and feel if you opened up your PC or laptop. Now, they are all data-centre firms masquerading as chipmakers, if we have to go by the financials.

Why Nvidia matters

All three lead a segment that we can now call "AI-chip companies". They do not sell one thing alone; they sell a system. For instance, Nvidia sells GPUs, networking, software and the CUDA development platform. That combination makes it difficult for customers to switch suppliers, even when rival hardware is available.

For many of us, AI simply means ChatGPT, Claude or Gemini—but they are just generative AI models that answer questions, write text or generate images. Those models need massive computing power to train and run. Here is where chips from companies like Nvidia matter. They provide much of that power, more like engines inside the data centres behind your popular AI services.

According to TrendForce, Nvidia leads the AI-server market, while other application-specific chips (ASICs) are gaining share. It expects AI-server shipments to approach 2.8 million units in 2026, up about 30.5 per cent year on year. IDC estimates that worldwide AI infrastructure spending reached $121.4 billion in Q2 2026, up 53 per cent year on year.

Even as the industry grows, Nvidia’s sheer scale affects its rivals. When cloud companies buy Nvidia systems, they often commit to its software, networking and power design. That can make it harder for AMD and Intel to win large AI deals, even if their chips are—at times—competitive on price or performance.

What it means for AMD and Intel

AMD has always been the clearest beneficiary of any shift away from Nvidia. Its data-centre revenue grew two-fold year on year in Q2, driven by Instinct GPUs and server CPUs. But AMD still needs to prove that customers will build long-term AI systems around its platform. Even before the Nvidia debacle, TrendForce expected AMD’s share of the AI-server market to reach about 10 per cent going forward.

Now, Intel’s story is a bit different. Despite Q2 revenue rising 25 per cent, it posted a loss (as per GAAP reporting standards). And the once-market leader will have to depend on improving margins to survive, all the while doubling down on its manufacturing roadmap and converting data-centre demand into profits that stay.

And it is already showing in its numbers. Intel improved its revenue in Q3 by 21 per cent YoY to $16.5 billion. GAAP EPS was $0.60. And despite the recent slowdown, it forecasted Q4 revenue between $15 billion and $16 billion. What needs to be noted is that its data-centre and AI (DCAI) revenue jumped 78 per cent to $7.3 billion. In fact, it posted its strongest data-centre growth in 15 years!

So, AI-chip demand is not going anywhere soon. Yes, the IDC expects the server-market growth to moderate from 2027. Huang’s departure might slow some decisions, and customers might wait for clarity on Nvidia’s leadership, product roadmap and supply commitments. But they are unlikely to abandon AI infrastructure altogether.

The real opportunity is for AIM, which could convert Nvidia’s uncertainty into getting ahead design wins. This could stabilise an industry currently hinging on the promises of a single company, and cut down on its dependence on a few suppliers without slowing the AI growth.

And all of this is if Huang is not in the picture. As of today, multiple reports are trickling in about how Nvidia employees have banded together to demand the "return of the boss". So, as it stands now, relationships may unravel; chips do not.