Artificial Intelligence: Google Spends $51 Billion in 3 Months as Amazon Burns Cash Despite Record Cloud

Artificial intelligence is driving cloud sales and attracting enterprise customers, but it is now mobilizing substantial cash outlays.

The results published by Alphabet, Microsoft and Amazon this week elicited a very different reaction from investors: growth is no longer enough; it is now necessary to demonstrate that datacenters, chips and computing capacity can be financed and profitable.

The contrast is striking. In the third quarter of 2026, Alphabet invested $51.4 billion and posted negative free cash flow of $9.1 billion. Amazon reported negative free cash flow of $42.4 billion over the trailing twelve months, despite the strongest cloud growth in nearly five years.

Microsoft, for its part, shows $18.6 billion of positive free cash flow in its latest fiscal quarter. These data do not all cover the same period: comparing them requires a method.

Google: spectacular growth, but $51.4 billion in investments in three months

Alphabet generated $127.6 billion in revenue in the third quarter of 2026, a rise of 25% year over year. Advertising and Google services produced $96.3 billion in revenue, up 11%. Revenue from Google Search and related activities advanced by 12%, a pace that worried investors when compared with earlier periods.

Google Cloud growth is, however, exceptional: $30.7 billion in revenue, +103%, driven notably by AI infrastructure, enterprise services and TPU systems. Yet this strong demand comes at a cost. Alphabet spent $51.4 billion on investment outlays during the quarter; its free cash flow was -$9.07 billion, versus -$5.9 billion in the prior quarter.

Caution is warranted: negative free cash flow is not a non-cash loss. Alphabet remains profitable for the quarter. Management also explained that a $5.2 billion payment related to an old European fine and TPU inventories had weighed on cash. Markets therefore assess both the scale of the financial effort, the speed of return on investment, and the risks facing traditional research.

Amazon: cloud activity accelerates, but free cash flow swings negative

Amazon reported on October 29 quarterly revenue of $201.9 billion, up 12%, and operating income of $28.1 billion. Its Amazon Web Services (AWS) division posted $46 billion in revenue, +39.3%, its fastest growth in almost five years. AWS operating income reached $17.8 billion.

The red flag lies in the capacity to fund expansion. In the twelve months ending September 30, Amazon generated $168.3 billion of operating cash flow, but its net capital expenditures after disposals and aids reached about $210.7 billion. The balance yields free cash flow of −$42.4 billion, vs +$14.8 billion a year earlier. Amazon expects around $220 billion of cash capital expenditures in 2026. This figure remains an annual forecast, not a quarterly outlay.

Thus, the group does not “lose” $42.4 billion in the sense of the income statement: it posted a quarterly net profit of $20.7 billion. It does, however, deploy more cash to build its infrastructure than it currently frees up from its operations over the considered year. That timing mismatch is what the market punishes.

Microsoft: the comparison that unsettles its rivals

For its first quarter of fiscal 2027, ended September 2026, Microsoft reported $92.3 billion in revenue, +19%, and $36.6 billion in net income. Azure grew 48% year over year and the Microsoft Cloud division 30%. Capital expenditures reached $45.2 billion, but the group posted $18.6 billion in positive free cash flow.

This difference does not mean that Microsoft is insulated from a datacenter overcapacity issue. It shows that as of the results date, the group’s existing and new activities still generate enough liquidity to absorb quarterly investments according to the disclosed metric. Microsoft’s high spending, like that of its rivals, remains tied to long-term hardware commitments.

Why the comparison of the three giants must remain rigorous

FCF stands for “free cash flow”: cash available after investment expenditures according to the definition published by each company. Accounting scopes may vary. The table does not rank the companies from the “best” to the “worst”: it shows distinct financing tensions and different horizons.

The three unknowns of the AI economy

First unknown: the utilization of capacity. Server rooms occupied by paying customers have a different economic value from capacity built in advance. Second unknown: the lifespan of equipment. Chips evolve rapidly, while datacenters and their electrical contracts are long-term commitments. Third unknown: pricing. If AI models become cheaper or more efficient, the price charged per query can fall at the same time as infrastructures must be amortized.

The cloud business model also has a multiplier effect: increasing the number of users does not mechanically translate into an equivalent rise in earnings. One must account for energy, cooling, chip availability, network costs, maintenance, and potential price declines in case of competition.

What lesson for Tunisian businesses?

For a small or medium-sized enterprise, a bank, an industrial company, or a Tunisian media outlet, the financial news from the digital giants offers a practical lesson: buying or developing AI does not in itself constitute a result. The project must measure a tangible gain, such as time saved on an operation, a reduction in error rate, a decrease in production costs, or an increase in the number of cases processed without a drop in quality.

A serious experiment must compare a reference process with an AI-assisted process, include licensing costs, API consumption, integration, human oversight and data security costs, and then measure real effects. For media, the evolution of online search also raises a distribution question: if assistants respond directly to users, how to preserve the value of content, its visibility and the relationship with the reader?