Portfolio Manager, Rand Swiss

Viv is a senior analyst and investment specialist, focusing on international and local markets. He frequently appears in the media, contributing to channels such as: SABC, CNBC, SAFM, 702 and ETV. He is also a regular guest lecturer at a number of prominent business schools and advanced education programmes, and previously lectured at tertiary institutions such as UKZN and DUT.

You can follow Viv on YouTube at @RandSwiss for regular market updates and AI insights.

Nvidia’s results just confirmed the AI buildout is accelerating

Thursday, 21 May 2026

Nvidia’s latest quarterly results were exceptional by almost any measure.

Revenue for the first quarter of fiscal 2027 surged 85% year-on-year to $81.6bn, comfortably ahead of expectations. Adjusted earnings also beat forecasts, while Data Centre revenue, now the company’s core growth engine, climbed 92% to $75.2bn.

The group generated nearly $49bn in free cash flow during the quarter, increased its dividend and announced an additional $80bn share buyback programme.

Yet despite those extraordinary figures, the market reaction was surprisingly restrained.

That may be because investors are beginning to realise Nvidia itself is no longer the entire story.

The more important development is what Nvidia’s results reveal about the broader AI economy.

AI spending is becoming commercially driven

For much of the past two years, investors have debated whether the AI boom resembles previous speculative technology bubbles.

Nvidia’s results suggest something very different is emerging.

The companies purchasing Nvidia’s hardware are increasingly generating meaningful revenues from their own AI products and services. What initially appeared to be experimental infrastructure spending is rapidly evolving into a large-scale commercial ecosystem.

Anthropic offers perhaps the clearest illustration of how quickly this shift is unfolding.

The company, which developed the Claude AI platform, reportedly generated annual recurring revenue of roughly $100m in early 2024. By April 2026, estimates suggest that figure had risen to approximately $30bn, with some projections implying it could ultimately exit this year generating between $80bn and $100bn in recurring annual revenue.

Even allowing for the uncertainty surrounding private market estimates, the scale of that growth is remarkable.

More importantly, much of this demand is increasingly enterprise-led rather than consumer-driven.

Businesses are now deploying AI systems across software development, legal workflows, customer support, research and internal productivity functions. In many cases, the economics are already compelling enough to justify significant ongoing infrastructure expenditure.

That distinction matters.

Technology booms become far more durable when they move from speculative adoption into measurable productivity and revenue generation.

The constraint is now infrastructure

The most important takeaway from Nvidia’s results may not be Nvidia’s continued dominance in AI chips.

It may be the growing evidence that the entire AI ecosystem is beginning to run into infrastructure constraints.

Anthropic itself has reportedly acknowledged that compute availability is becoming one of the key factors limiting the rollout of its products.

In other words, the challenge is no longer finding customers.

The challenge is securing enough compute capacity to meet demand.

That changes how investors should think about the opportunity set emerging around AI.

The market is gradually recognising that this is not simply a software story. It is also an infrastructure story involving semiconductors, networking, memory, advanced packaging, hyperscale data centres and power systems.

Every new AI model requires substantial processing power.

Every enterprise deployment increases demand for compute infrastructure.

And every AI agent integrated into the economy adds additional pressure to the systems supporting it.

That is why the broader AI supply chain may become increasingly important over the coming decade.

China still faces major structural hurdles

Nvidia’s position also continues to benefit from geopolitical and technological barriers that remain difficult to overcome.

US export restrictions have limited China’s access to the most advanced AI chips, accelerating Beijing’s push to develop domestic alternatives.

Huawei has emerged as China’s primary national AI hardware champion, but manufacturing bottlenecks remain significant.

The issue is not simply chip design.

Advanced semiconductor manufacturing requires leading-edge lithography, advanced memory integration, sophisticated packaging capabilities and fabrication scale that remain extremely difficult to replicate.

Even with substantial state support, China still appears several years behind Nvidia and the broader Western semiconductor ecosystem in many critical areas.

That reinforces Nvidia’s strategic importance within global AI infrastructure markets, even as China continues investing aggressively in domestic capacity.

The next phase may extend beyond Nvidia

• Advanced semiconductor manufacturing
• AI networking infrastructure
• Data centre expansion
• GPU cloud infrastructure
• Packaging and memory supply chains
• Enterprise AI integration