To The Moon

Microsoft vs Alphabet: AI Spending Is Growing. Where Is the Cash Return?

The latest earnings from Microsoft and Alphabet highlight an important question for the entire AI trade: how much profit can companies generate from the enormous amounts they are spending on AI infrastructure?

Microsoft delivered a strong quarter, with revenue of $92.3 billion and adjusted earnings per share of $4.89. Azure revenue grew 48% year over year. Yet the market’s reaction was not simply a celebration of those numbers.

Alphabet reported $127.58 billion in revenue, but its results missed analyst expectations on key measures. Its shares came under pressure after the announcement, while Microsoft’s shares moved higher in the post-earnings reaction.

To me, the divergence shows that investors are becoming more selective. They are no longer rewarding every company that mentions AI. They want evidence that AI investment can produce enough incremental revenue, profit and cash flow to justify the cost.

1. Microsoft: strong cloud growth, but watch the cost of scaling

Microsoft’s Azure growth is an encouraging signal. Demand for cloud computing, AI services and enterprise infrastructure is giving the company a substantial opportunity to expand its business.

The company also reported commercial remaining performance obligations of $677 billion, up 73% year over year. That indicates a large pool of contracted future business, although it should not be treated as revenue that will all be recognised immediately or converted into profit at the same margin.

Microsoft has several advantages: an established enterprise customer base, recurring software revenue and the ability to sell cloud services alongside products that businesses already use. These factors may help it monetise AI more effectively than a company relying on a single new product.

But there is another side to the story.

Building and operating data centres requires substantial spending on servers, networking equipment, power and facilities. These assets also generate depreciation expenses over time. Even when demand is strong, the returns on new capacity depend on utilisation, pricing and the revenue earned from each unit of infrastructure.

My key question is whether Azure’s growth can continue at an attractive margin as Microsoft expands capacity. Strong demand is valuable, but the investment case becomes much more convincing when incremental revenue eventually translates into free cash flow after capital expenditure.

2. Alphabet: a reminder that even a strong business can disappoint

Alphabet’s results provide a useful contrast.

Its advertising business remains a major source of cash generation, while Google Cloud gives it another route to monetise enterprise AI demand. However, investors must weigh the potential benefits of AI against the cost of developing models, building infrastructure and competing in cloud services.

There is also a strategic challenge. AI-powered search and conversational tools could create new products and revenue opportunities, but they may also change how users interact with traditional search. Alphabet needs to show that it can capture the upside of AI without undermining the economics of its existing business.

A revenue miss does not automatically mean the long-term thesis is broken. Equally, a company with strong cash generation is not immune to valuation risk or execution mistakes.

I would watch Google Cloud growth, operating profitability, advertising performance and capital expenditure together. If AI investment supports cloud demand while the core advertising business remains resilient, the company has several ways to fund its ambitions. If spending accelerates faster than monetisation, investors may become less willing to give it the benefit of the doubt.

3. Amazon: AWS growth must justify the infrastructure bill

Amazon is an important next test because AWS is a major cloud infrastructure business and the company is investing heavily to meet demand.

For its next results, I would focus on three areas:

* AWS growth: Is demand accelerating, and can Amazon convert that demand into sustained revenue?

* Operating income: Is the cloud business generating more profit as sales expand?

* Free cash flow after capital expenditure: How much cash remains after Amazon pays for the infrastructure needed to support future growth?

The third measure is especially important. Revenue and operating income can look strong while heavy investment in property and equipment absorbs a large amount of cash.

That does not automatically make the spending unattractive. Infrastructure built today may generate returns over several years. But investors need to assess the expected payback period, capacity utilisation and whether customers are committing to enough business to support the investment.

I would be more comfortable with higher capital expenditure if AWS growth, operating profitability and demand visibility were strengthening together.

4. Intel and SanDisk: two different ways to test the AI hardware boom

Intel and SanDisk offer another perspective, but they should not be analysed as if they have the same business model.

For Intel, I would look for evidence of improving execution, stronger product competitiveness and a credible path to sustainable operating profitability and cash generation. AI demand alone cannot solve manufacturing challenges or guarantee that Intel will regain market share.

For SanDisk, the focus is NAND flash and storage, including enterprise SSDs used in data centres. AI workloads create demand for storage capacity, but the investment case also depends on NAND pricing, product mix, supply discipline and margins.

A strong SanDisk report would show more than higher revenue. I would want to see whether pricing improvements are translating into stronger profitability, whether demand appears durable and whether management’s forward outlook supports the current earnings expectations.

Both companies could benefit from AI infrastructure spending, but the risks are different: Intel must demonstrate operational progress, while SanDisk must show that the memory and storage cycle can sustain attractive earnings.

5. What would convince me that AI spending is paying off?

I would look for three signs across the sector.

First, revenue growth must be supported by real customer demand. Contracted business and repeat spending are more reassuring than optimistic forecasts alone.

Second, profitability must improve alongside sales. Growth that requires increasingly expensive infrastructure is less attractive if margins keep weakening.

Third, free cash flow must hold up over time. Heavy capital expenditure can depress cash flow in the near term, so I would assess it against the expected returns from new capacity rather than treating one quarter’s cash flow as the final verdict.

I would also compare companies using consistent accounting measures. Adjusted earnings can help investors understand operating trends, but they should not replace reported results or cash-flow analysis.

My takeaway

I remain interested in AI infrastructure, but I would be selective about where I invest. Microsoft’s Azure growth is encouraging, Alphabet’s results show that execution and expectations matter, and Amazon, Intel and SanDisk each face a different test of whether AI demand can translate into sustainable returns.

The companies that ultimately benefit most may not be those spending the most. They may be the ones that convert each additional dollar of investment into durable revenue, healthy margins and cash that remains available after funding the next stage of growth.

For me, the next phase of the AI trade is about proving the return on investment, not simply announcing a bigger AI budget.

Which company do you think is best positioned to turn AI spending into sustainable cash flow: Microsoft, Alphabet, Amazon, Intel or SanDisk?

$Microsoft(MSFT)$  $Alphabet(GOOGL)$  $Amazon.com(AMZN)$  $Intel(INTC)$