$255 Billion AI CAPEX = New Zealand’s GDP—Are End Users Catching Up?

All Magnificent 7 stocks have now reported their results for the final quarter of 2024, making it the perfect time to review their performance. Below is a quick summary.

Here are the key observations:

1. Revenue guidance suggests a slowdown in growth for the next quarter. This is particularly crucial for cloud giants, as they must justify the massive CAPEX spent on AI.

2. The three cloud giants saw their stock prices decline after their earnings reports. While the exact reason is uncertain, one possible explanation is their commitment to even higher CAPEX in 2025:

   1. Amazon: $100 billion (+20%)

   2. Microsoft: $80 billion (+44%)

   3. Alphabet: $75 billion (+43%)

   4. Together, $255 billion is as much as New Zealand’s GDP

3. Comparing these CAPEX increases with their cloud revenue growth reveals that capacity expansion is outpacing demand, particularly for Microsoft and Alphabet. Amazon appears more measured in its growth. While cloud giants may be anticipating higher future demand and accelerating investments due to lead times, forecasting demand accurately remains a challenge, and overcapacity is always a risk.

   1. AWS: +19%

   2. Microsoft Azure: +31%

   3. Google Cloud: +30%

4. Most companies posted double-digit revenue growth. Tesla, however, managed only 2% growth yet saw its share price surge 102% over the past year—outpacing even Nvidia. This may be attributed to Trump’s election victory, as Elon Musk's association with him has turned Tesla into a "Trump trade."

5. Meanwhile, Microsoft was a surprising underperformer, gaining just 3% over the past year—significantly lagging behind many stocks, including the S&P 500, which rose more than 20%.

Where are the AI end users?

The bulk of the demand for Nvidia GPUs comes from cloud giants, which, in turn, rely on AI model builders as their key growth driver. However, this doesn’t necessarily mean that AI model builders have a large base of paying customers. In a recent survey I conducted on my socials, 80% of respondents indicated that less than 25% of their work involves AI.

Take Microsoft’s CoPilot, for example. Priced at an additional $30 per user for enterprise customers, its revenue falls under Microsoft’s 365 Commercial Cloud segment, which reported just 16% growth—far from the acceleration in AI adoption that many had anticipated.

The Decoder compiled monthly active user data for various AI platforms, and the results are telling. Microsoft CoPilot's adoption lags behind ChatGPT, which has been a runaway success. In fact, CoPilot’s active users declined from June to September 2024 after an initial growth spurt. Meanwhile, Bing, another AI-powered product from Microsoft, has seen user numbers decline steadily since January 2024.

While CoPilot and Bing may not be taking off as expected, Microsoft still benefits from its investment in OpenAI, which owns ChatGPT. As part of the deal, OpenAI is required to use Microsoft Azure, ensuring that ChatGPT’s explosive growth translates into tangible revenue for Microsoft.

Unlike Microsoft, Google Workspace is bundling Gemini AI into its suite and raising prices by $2 per month. This move suggests that Google's product team may doubt the standalone demand for Gemini AI—if it were truly compelling, it would likely be sold separately rather than bundled into an existing subscription.

The Cloud Giants’ Prisoner’s Dilemma

The cloud giants find themselves locked in a classic prisoner’s dilemma, where each must decide between two strategies:

1. Invest Heavily in AI (Defect)

Pros:

• Gains a first-mover advantage

• Generates buzz around AI capabilities

• Secures top AI talent

• Preempts competitors

Cons:

• Requires massive CAPEX

• Uncertainty about return on investment

• Risk of overextending resources

2. Hold Off or Invest Lightly (Cooperate)

Pros:

• Preserves capital and avoids risky, large-scale R&D bets

• Frees up funds for other uses (e.g., acquisitions, share buybacks, new product lines)

Cons:

• Risk of being overtaken in core markets or innovation races

• Competitors could capture the lion’s share of AI-driven value

Each tech giant considers what its rivals might do. Executives understand that others face the same pressures—no one wants to be left behind.

Even if an industry-wide restraint would keep costs down and maintain stable profits, no company can trust that its competitors will hold off. The fear is simple:

• If a competitor invests heavily while they don’t, they risk falling behind—losing market share, AI dominance, and top talent.

• If all invest, at least they don’t lose relative positioning in the AI race.

Thus, each company independently arrives at the same conclusion:

"If I don’t invest while others do, I lose. If we both invest, at least I don’t fall behind."

This leads to a “Defect-Defect” equilibrium, where all cloud giants aggressively invest, despite the risk of overbuilding capacity.

Even though end-user demand has yet to materialize at scale, cloud giants can’t afford to wait—they are securing AI infrastructure now, fueling insatiable demand for Nvidia chips and other AI-related investments.

The big question: Have they overbuilt capacity?

If so, we could see a correction in cloud and AI-related stock prices in the future. Only time will tell whether this massive investment spree pays off—or if it leads to another cycle of overinvestment and retrenchment.