To The Moon

For much of the past two years, the market has rewarded Big Tech companies for investing aggressively in artificial intelligence. Building data centres, securing advanced chips and expanding computing capacity were seen as necessary steps to capture the next major technology opportunity.

But the market is starting to change its expectations.

Investors are no longer satisfied with hearing how much companies plan to spend on AI or how quickly their cloud businesses are growing. They want evidence that these investments can generate enough additional revenue, profit and free cash flow to justify the enormous costs.

Thursday’s market reaction suggests that this shift is becoming increasingly important. The Nasdaq 100 fell 1.42%, while the S&P 500 declined 0.30%. Technology shares faced renewed pressure as investors assessed whether the returns from AI infrastructure spending would eventually match the scale of investment.

Amazon was particularly affected. The company reported $201.9 billion in revenue and adjusted EPS of $1.89, missing market expectations. AWS revenue increased 39% to $48.1 billion, demonstrating that demand for cloud computing remains strong. However, trailing 12-month free cash flow declined 35% to $42.2 billion. Shares subsequently fell another 7.30% after hours.

The interesting part is that AWS is still growing rapidly. Yet the market is questioning what that growth is costing the company and how much cash is ultimately left after infrastructure investment.

This raises an important question: among Amazon, Microsoft and Meta, which company has the clearest path to turning AI investment into sustainable financial returns?

1. Amazon: Impressive AWS growth, but can cash generation keep up?

Amazon’s biggest advantage is AWS. As businesses increasingly adopt generative AI, they need cloud infrastructure, computing capacity and data storage. AWS is positioned to benefit from this demand.

Its 39% revenue growth is an encouraging sign. However, building data centres and purchasing advanced computing equipment require substantial upfront capital.

The challenge is that infrastructure spending often happens well before the resulting revenue is fully realised. This creates pressure on free cash flow in the short term.

Investors need to distinguish between temporary cash flow weakness caused by investment for future growth and a more concerning situation where spending continues to increase without sufficient returns.

For Amazon, I would watch AWS operating income, capital expenditure growth and free cash flow conversion. If AWS continues expanding while cash generation gradually improves, the current concerns could eventually ease.

2. Microsoft: Turning AI demand into recurring enterprise revenue

Microsoft has a different advantage. It is not relying on one business to monetise AI.

Azure provides cloud infrastructure, while Microsoft 365, Copilot, GitHub and its broader enterprise software ecosystem offer multiple opportunities to generate additional revenue from existing customers.

This is important because selling AI services to established enterprise customers could create recurring revenue rather than relying entirely on new infrastructure demand.

Microsoft also has an existing software distribution network, which gives it an opportunity to integrate AI into products that businesses already use.

However, there is still a financial question. AI infrastructure is expensive to build and operate. Even if customers are adopting AI tools, investors need to see whether the additional revenue translates into higher operating margins and stronger free cash flow.

Microsoft looks particularly interesting because it has several potential monetisation channels. Its challenge is demonstrating that AI revenue can grow faster than the costs associated with delivering those services.

3. Meta: An advertising business funding its AI ambitions

Meta is an interesting case because its AI strategy is closely connected to its existing advertising business.

Unlike cloud providers, Meta does not need to sell all its AI computing capacity directly to customers. It can use AI internally to improve content recommendations, user engagement and advertising targeting.

Better recommendations can increase time spent on its platforms, while more effective advertising systems can improve returns for advertisers. If these improvements translate into higher advertising revenue, Meta could generate financial benefits from AI without having to sell a separate AI subscription to every user.

Its existing advertising business also provides a substantial source of cash to fund investment.

However, Meta is not immune to the same spending concerns. Expanding data centre capacity and developing AI models require significant capital. Investors will want evidence that improved advertising performance and new AI products can justify the additional expenditure.

The key question is whether Meta can improve the profitability of its existing business while expanding its AI capabilities.

What should investors actually measure?

I think the next stage of the AI investment cycle requires looking beyond headline revenue growth.

There are four financial measures worth watching:

1. Free cash flow: How much cash remains after capital expenditure? A company can report strong earnings but still experience pressure on cash generation when infrastructure spending accelerates.

2. Return on invested capital (ROIC): Are companies generating sufficient returns from the billions invested in data centres, chips and AI infrastructure? This helps determine whether spending is creating economic value.

3. Operating margins: As AI services scale, are companies becoming more efficient, or are higher computing and operating costs absorbing the additional revenue?

4. Revenue growth relative to capital expenditure: If spending increases much faster than revenue for an extended period, investors may begin questioning the economics of the investment cycle.

There is also a valuation issue. Even if all three companies successfully monetise AI, their share prices already reflect significant expectations for future growth. A good business does not automatically represent a good investment at every price.

My view: Microsoft has the strongest overall position, but Meta deserves attention.

If I had to choose one, I would currently lean towards Microsoft because of its combination of cloud infrastructure, enterprise distribution and multiple AI monetisation opportunities.

Amazon remains an important competitor, particularly with AWS growing at 39%. However, its declining free cash flow highlights why investors are becoming more selective about the cost of AI expansion.

Meta may have a different advantage. Its advertising business gives it an established source of cash, while AI can potentially improve the economics of that business.

I would not automatically interpret Amazon’s share price decline as evidence that its AI strategy is failing. Equally, I would not assume that strong cloud growth guarantees attractive shareholder returns.

The real test will be what happens over the next few quarters. Can these companies maintain their AI investment programmes, grow revenue, protect margins and eventually produce stronger free cash flow?

The AI opportunity remains significant, but the market is entering a more demanding phase. Companies will need to demonstrate that their spending decisions create measurable financial value.

The winners of the AI race may not necessarily be those with the biggest data centres or the most advanced models. They could be the companies that turn AI adoption into recurring revenue, higher margins and sustainable free cash flow.

For discussion purposes only, not financial advice.