Since the beginning of 2023, big tech has driven the S&P 500. Cloud service providers have made heavy investments to convert their datacenters from general computing facilities to accelerated computing hubs. This is what enables generative AI and machine learning applications to run, which has brought a flood of investor activity to this portion of the industry.

Amazon, Google, Microsoft, and Meta dramatically expanded their capital expenditures (CapEx) and spent billions on datacenter infrastructure over the past several years. Based on Visible Alpha consensus, that CapEx total is expected to expand to $274 billion. 

A significant component of this investment is shifting from CPUs to GPUs, which is the core enabling factor of the datacenter shift. NVIDIA’s proprietary CUDA platform has helped datacenters perform tasks faster, increase efficiency, and save on costs, all in the name of developing generative AI applications. This massive CapEx investment set off the AI boom and has driven NVIDIA's stock price. So, what's next? 

Outlook and Factors for Future Growth

Looking out to fiscal year (FY) 2027, expectations for datacenter revenues and EPS have doubled from $112 billion in 2024 to nearly $224 billion in FY 2027, driving this segment’s revenue growth up 4.7x in that period.

This top line growth has dropped to the bottom line and driven significant EPS growth expectations for NVIDIA and its stock price. Is all this growth finally priced in, or are there further upward revisions to come? NVIDIA recently launched its new Blackwell product and there is some debate about the growth trajectory. The most optimistic analysts think the consensus number could increase another $100 billion, while the bear analysts see this possibly falling short by $75 billion.

The Visible Alpha AI Monitor, which aggregates publicly traded US technology companies to provide a comprehensive measure of the current state and projected growth of the core AI industry, projects that this range of estimates is close to $180 billion. This wide range of forecasts means there is a lot of debate in the market in terms of what the performance of Blackwell could be and which quarters it should start to be meaningfully calculated into the consensus number. In this February's earnings call, we should receive more color about the firm's visibility for Blackwell. However, it is clear from this range of estimates that there is significant variance in expectations, which may lead to continued volatility until the range narrows.

Data compiled January 2025. Source: Visible Alpha, January 15, 2025. © S&P Global.

Looking ahead to this year and next, growth expectations remain robust for the 10 largest companies generating AI-exposed revenue, driven by Alphabet, Microsoft, NVIDIA, Amazon, and Meta. In 2024, this group generated 52% upside, outperforming smaller cap stocks. 

Who Will be the Winners in 2025? 

There are three key areas we are watching to drive stocks this year. Regulation is a looming question. Europe has been much stricter than the US and it will be interesting to see how this unfolds. We have seen a lot of innovation also at the chip and the model level, which has benefited NVIDIA and the big tech stocks. There has not been as much innovation at the application level for people like you and I, the end users. 

This is an area where we may see some regulation if there is broader adoption. There are questions about which companies could dominate this space. Will the game-changing end-user AI applications come from small startups or the existing large firms or from a variety of firms? As AI begins to penetrate workflows and our daily lives, regulation may start to be a factor. 

One area of excitement is agents, as they can handle domain and persona-specific workflows. For example, a company may want to deploy unique AI agents for research, security, analytics, sales, or customer service to complement the human work that is being done. We are interested to see how firms leverage them to drive revenues and increase profitability.

The Keys for Broader AI Adoption 
When assessing the impact of artificial intelligence companies on the capital markets, it will be interesting to see if a breakthrough at the application level may happen. From our perspective, this is likely to be the gateway to broader adoption.

There was a view last year that Apple might be going to provide that gateway, but we have not seen that. NVIDIA is one of the most important stocks in the space for this type of information. If their range of estimates starts to narrow, we should get a clearer view of the trajectory and direction for their new Blackwell platform, which is widely anticipated to help drive and support that broader adoption.

AI infrastructure companies with smaller market caps have underperformed over the last two years, but our AI Monitor Index picked up greater momentum in that space after the most recent US election. This could present an opportunity for some of those underperformers to reverse back to the mean and push investors to revisit them in 2025.