[DigitalToday reporter Chi-gyu Hwang] Can Amazon Web Services (AWS) grow into a $1 trillion revenue business?

Amazon CEO Andy Jassy (앤디 재시) recently assessed after the second-quarter earnings announcement that AWS has the long-term "potential to become a $1 trillion revenue business". He said a ceiling previously seen as hundreds of billions of dollars has now been lifted by at least more than double.

AWS posted an annualised revenue run rate of $169 billion in the second quarter. Revenue growth was 36.7 percent, the highest in 18 quarters. The growth rate jumped to about 37 percent in roughly a year, from 17 percent five quarters earlier.

Tomasz Tunguz (토마즈 턴구즈), founder and managing partner of Silicon Valley venture capital firm Theory Ventures, cited Jassy's remarks on social media platform X (Twitter). "Whether AWS reaches $1 trillion in revenue depends on how much corporate customers use AI," he said. "If demand for AI infrastructure expands as expected, AWS can surpass $1 trillion in revenue. Conversely, if corporate AI transformation slows, massive upfront investment and the burden of borrowing could hold Amazon back," he said.

According to him, AWS is also showing a marked upswing even compared with Microsoft. AWS growth has risen by 20 percentage points over the past six quarters. Microsoft Azure's growth has stagnated at around 40 percent, while AWS is rapidly narrowing the gap.

Profitability is also a strength. "Second-quarter AWS operating profit rose 64 percent to $16.6 billion, and the operating margin reached 39.4 percent, up 6.5 percentage points from a year earlier," Tunguz Partner said. "AWS has the largest revenue scale among the three companies, while sustaining high growth and maintaining high profitability," he said.

In backlog showing future demand, it lagged Microsoft. AWS backlog rose to $496 billion from $364 billion three months earlier, but it was below Microsoft's $678 billion. Tunguz Partner explained that Microsoft, which secured large contracts with OpenAI early, has pre-empted demand for AI infrastructure.

Amazon is also aggressive on AI infrastructure investment. It raised its 2026 capital spending plan to $220 billion from $200 billion, citing reasons including rising memory prices. It is also expanding debt issuance this year to secure funds for investment.

Jassy says the investment now will come back as profit in the future.

He said data centres require capital investment for about 2 years before coming online, but can then be used for more than 30 years, and servers are also operated for several years. If AI demand continues in the long term, the current large-scale investment can return as high profit in the future, he said.

Jassy described the AI market as a "barbell" structure between AI model developers that require large-scale computing and companies pushing to reduce costs. He projected that the work of existing companies in between, which have not yet broadly applied AI inference, could grow into the biggest market in the future.

But it is not certain whether this market will grow as steeply as demand coming from current AI developers. Tunguz Partner reported that Jassy also acknowledged this part.