Quick View: An AI report card through the lens of Nvidia earnings

Artificial intelligence (AI) has been the undeniable driver of global equity markets over the past few years, and the numbers associated with it  – capital expenditure, revenue growth, market capitalizations – are staggering.

Given the duration of this mega-theme and AI’s dominance over stocks, many investors are wondering whether expectations have moved ahead of reality, and whether this historic level of investment is justified. To address this question, we can use this week’s quarterly earnings report from Nvidia as a report card of sorts, as much of the AI ecosystem either directly or indirectly has its fortunes linked to the semiconductor juggernaut.

Firing on all cylinders

Based on Nvidia’s recent financial and operational performance –  and perhaps more importantly, its expectation for near- to mid-term developments – the AI supercycle is not just progressing as planned but accelerating in many somewhat unexpected ways. As intimated, the unprecedented magnitude of the report’s numbers is news in itself. More germane to investors, though, is how Nvidia’s performance illustrates the pace of AI adoption and how it is evolving.

Revenue for the quarter was roughly $82 billion, a 20% gain over the previous reporting period and up 85% from the same period in 2025. The story just beneath this headline figure was revealing with respect to AI’s evolution: Revenues were split roughly 50/50 between hyperscalers – the tech megacaps behind the historic capital expenditure (CapEx) buildout – and another category that includes the AI cloud plus industrial and enterprise users.

Even more notable is that growth attributable to hyperscalers clocked in at 12% quarter over quarter, while the cloud/industrial/enterprise segment expanded by 31%. Growth at this pace indicates demand for the fundamental building blocks of AI – graphics processing units (GPUs) – is strong.

Agents have arrived

The report also validated our view that the era of agents has arrived. Company management explicitly stated that inference has hit an inflection point. Nvidia’s early ascent was driven on the back of demand for its cutting-edge GPUs used in training AI models. The industry is now transitioning to the inference stage, which essentially means leveraging the trained models to carry out myriad tasks across the global economy. Much of that work will be done by AI agents.

Management echoed our view that AI users will create billions of agents, each tasked with executing the operational aspects of AI. Given the power of networking laws, this implies a nearly incomprehensible number of tokens – fundamental units of data processed – generated.

Also hinting at the increasing role AI will play, Nvidia announced an ambitious rollout of its first standalone central processing unit (CPU), Vera. It’s only recently understood that CPUs will be an essential element in enabling agents to schedule and execute relatively elementary computing tasks. Without CPUs, the best training and GPU-enabled “thinking” would be for naught. Management went so far as to put a $200 billion number on the AI CPU addressable market.

So much more than hyperscalers

In AI’s training stage, hyperscalers had the demand – and the cash – to ramp up this nascent ecosystem. As indicated by the 31% growth in the cloud/industrial/enterprise segment, a handoff is occurring, and these segments will be the end users of AI models, applying them to use cases across industries. Categories that we see as adopters are manufacturing, robotics, healthcare small chemicals, and physical sciences like energy. Not to be ignored is sovereign AI, as countries see it as a strategic imperative to fortify their economic and national security in the age of AI.

We view development as perhaps the most misunderstood concept in AI investing. Many investors hang on every word emanating from the hyperscalers while overlooking that there are hundreds of thousands of companies and other entities racing to integrate AI into their strategies and operations. The adoption of AI within this underappreciated – and massive – segment could exceed that of the adoption of earlier technologies given AI’s potential to improve productivity and eventually establish AI native business models.

Despite a rapidly broadening market of AI applications, frontier models and AI labs still play a major role in future advancement. Nvidia Chief Executive Jensen Huang announced deeper collaboration with model developer Anthropic to help alleviate its shortfall in compute. Furthermore, the universe of frontier models is growing, with all seeking the most advanced GPUs. As incremental – and more complex – compute capacity is deployed, one can argue that AI’s capabilities will grow, expanding its use cases and thus demand.

Another potential secular driver for the AI ecosystem is the growth of data centers whose mission will be to process as many tokens as possible and sell this service to applications providers in need of third-party compute. These AI factories, in our view, will become an increasingly large source of demand for both advanced GPUs and CPUs. These factories will be dispersed globally, with China possibly positioning itself as a dominant player in processing tokens. In this respect, Nvidia’s ability to sell at least older generations of GPUs to China is something investors will want to monitor as the geopolitical climate shifts.

Regardless of where these data centers are located, a dearth of electricity to adequately power them – at least outside of China – will invariably require their operators to seek out the most energy efficient chips to maximize their finite energy resources.

Still early innings

Mr. Huang and his management team, in our view, methodically debunked nearly every tenet of the bear case surrounding the scope of the AI investment cycle. A powerful rejoinder was management’s assessment that annual hyperscaler CapEx could reach $3 trillion to $4 trillion by 2030, compared to an estimated $1 trillion in 2027. This is a gaudy but imaginable number given agentic AI and its ability to be a force multiplier for inference. To be determined are the unknowns, namely what native AI businesses and applications emerge and what will be their incremental demand for compute.

Nvidia’s most recent earnings and outlook can be viewed as a sextant allowing one to peer over the not-too-distant horizon into the unknown world of the global AI economy. In this respect, we saw little in this week’s report to diminish our favorable view of the sector.

Yes, looking at hyperscaler CapEx spend in isolation could create anxiety in many investors. What they fail to recognize, in our view, is that GPUs, data centers, and now CPUs are establishing the foundation of a fundamentally transformed global economy – one in which countless end users across sectors and geographies will allocate resources to gain access, with the aim of improving their economic and societal outcomes.

IMPORTANT INFORMATION

Artificial intelligence (“AI”) focused companies, including those that develop or utilize AI technologies, may face rapid product obsolescence, intense competition, and increased regulatory scrutiny. These companies often rely heavily on intellectual property, invest significantly in research and development, and depend on maintaining and growing consumer demand. Their securities may be more volatile than those of companies offering more established technologies and may be affected by risks tied to the use of AI in business operations, including legal liability or reputational harm. 

Equity securities are subject to risks including market risk. Returns will fluctuate in response to issuer, political and economic developments.

References made to individual securities do not constitute a recommendation to buy, sell or hold any security, investment strategy or market sector, and should not be assumed to be profitable. Janus Henderson Investors, its affiliated advisor, or its employees, may have a position in the securities mentioned.

Technology industries can be significantly affected by obsolescence of existing technology, short product cycles, falling prices and profits, competition from new market entrants, and general economic conditions. A concentrated investment in a single industry could be more volatile than the performance of less concentrated investments and the market as a whole.