NEW YORK, NY – JANUARY 29: Stock numbers for Apple are displayed on a screen at the Nasdaq MarketSite in Times Square, January 29, 2019 in New York City. Apple is set to report first-quarter earnings results after U.S. markets close on Tuesday. (Photo by Drew Angerer/Getty Images)

Getty Images

Artificial intelligence is no longer confined to flashy demos—it’s now a line on the income statement, and an even bigger one on capital expenditure budgets.

In the third quarter of 2025, the world’s biggest tech companies reported robust growth fueled by AI and cloud services, paired with staggering increases in spending on the infrastructure powering that boom. Microsoft, Alphabet, Amazon, Meta, and Apple all posted solid results and double-digit revenue gains, buoyed by demand for AI-enhanced products and cloud computing. In tandem, their capital spending surged into the tens of billions of dollars per quarter as they race to build and rent out the computing “rails” behind the AI revolution.

Yet outside the U.S., two major players stands apart: Alibaba and Tencent. Often called China’s closest analogues to the Western tech titans, their businesses span social media, gaming, e-commerce, fintech, cloud, and digital content—making it a key non-U.S. test of whether the AI super-cycle is truly global. Their latest earnings provide a revealing counterpoint: Can a platform operating under different regulatory constraints and without Silicon Valley-level capex keep pace in the AI era?

Big Tech’s AI Infrastructure Spending Soars

Big Tech’s spending on AI infrastructure shows no sign of slowing, as illustrated by surging Capex across leaders like Amazon, Google (Alphabet), and Microsoft. The Q3 2025 earnings season underscored this arms race in AI and cloud investment. Some highlights include:

• Microsoft: Revenue jumped 18% year-on-year to $77.7 billion on strong demand for its AI-infused Azure cloud services. The company recorded nearly $35 billion in quarterly capital expenditure—a record high—and warned that annual spending will climb further to meet booming AI workloads.
• Alphabet (Google): Quarterly revenue climbed 16% to $102.3 billion, beating expectations, as its core advertising and Google Cloud units benefited from surging enterprise AI demand. Alphabet hiked its 2025 capex guidance to $91–93 billion, up dramatically from about $52.5 billion last year, to build data centers and custom AI chips.
• Amazon: The e-commerce and cloud giant poured $89.9 billion into capex in the first three quarters alone, largely to expand its AI-driven AWS cloud and logistics footprint. Amazon’s CFO projects around $125 billion in full-year 2025 capital outlays – and even more in 2026. So far, the investment is paying off: AWS revenue rose 20% in Q3, its fastest growth in years, as businesses embraced new AI services on the platform.
• Meta (Facebook): Meta reported solid revenue growth (26% YoY), but investors’ eyes are on its aggressive infrastructure push. Q3 capex reached $19.4 billion, and Meta raised its 2025 spending plan to $70–72 billion, with “notably larger” outlays slated for next year to build AI supercomputing capacity. This massive investment in data centers and chips – aimed at powering advanced AI models and services – is pressuring Meta’s near-term margins and stirred debate about returns.
• Apple: Even Apple, traditionally less focused on cloud data centers, delivered a record $94 billion in its June quarter (a fiscal Q3 record for the company) on the back of strong iPhone and services sales. On its earnings call, Apple emphasized it is “significantly” growing AI investments across devices and services. Those AI initiatives are driving a rise in Apple’s own capital expenditures (which hit $3.5 billion in the June quarter, up from $2.1 billion a year prior) – though Apple noted it relies on third-party infrastructure for some needs, keeping its capex growth more moderate than peers.

Across the board, these giants are treating AI and cloud infrastructure as foundational investments. The through-line: their capex profiles now resemble national-scale infrastructure projects, underlining expectations that AI is a long-term growth engine. Each company is effectively laying digital railroad tracks for the AI age, aiming to be a platform others must depend on.

Alibaba: The Chinese e-commerce and cloud giant posted a 5% YoY increase in total revenue to RMB 247,795 million (US$34,769 million) for the quarter ended September 30, 2025. However, this growth came at a significant cost to profitability, with net income decreasing by 53% YoY to RMB 20,612 million (US$2,893 million), and Adjusted EBITA dropping by 78% YoY to RMB 9,073 million (US$1,274 million). Crucially, Alibaba’s free cash flow turned negative, reaching RMB (21,840) million (US$(3,068) million), reflecting substantial investments. This investment was particularly evident in the China E-commerce Group, where Adjusted EBITA decreased 76% due to heavy spending in quick commerce, user experience, and technology. In contrast, the Cloud Intelligence Group saw its revenue accelerate by 34% YoY to RMB 39,824 million (US$5,594 million), with AI-related product revenue showing triple-digit growth, and Adjusted EBITA for this segment increasing by 35% to RMB 3,604 million (US$506 million). This signals Alibaba’s commitment to building out its AI and cloud infrastructure, mirroring the investment-heavy strategy of its Western counterparts despite the immediate impact on overall profitability.

Tencent’s third-quarter 2025 results show a company riding the AI wave, but with a very different strategy. Rather than outspending rivals on data centers, Tencent has focused on weaving AI into its sprawling ecosystem of apps and services. The company’s revenue rose 15% year-on-year to RMB 192.9 billion (about $27 billion), topping analyst forecasts, and net profit jumped 19%. Notably, this growth came amid China’s lower economic expansion and stricter tech regulations, highlighting Tencent’s resilience.

Crucially, AI is now a real contributor to Tencent’s top line, not just a research project. Advertising revenue surged 21% from a year earlier, which management attributed largely to AI-driven improvements in ad targeting and creativity, as well as higher ad impressions from new content formats like WeChat’s Video Accounts (a TikTok-like short video feed). Tencent’s CFO noted that roughly half of the ad upswing came from higher ad pricing (eCPMs) thanks to more precise AI targeting. In other words, better machine learning models helped match ads to users more effectively, boosting advertisers’ willingness to pay without Tencent having to cram in more ads. Meanwhile, user engagement with video content (via Video Accounts) and other services drove the rest of the ad growth by increasing ad inventory.

Tencent has also rapidly deployed AI to enhance its products and efficiency. Over the past year, the company rolled out its in-house large language model (Hunyuan) and a ChatGPT-like assistant named Yuanbao across its platforms. Yuanbao now integrates with popular apps like WeChat (including Channels short videos and official accounts), QQ Music, Tencent Video, and even enterprise tools like Tencent Meeting, where features such as AI-generated meeting summaries have been introduced. This “applied AI” approach – infusing intelligent features throughout social media, gaming, finance, and office apps – aims to deepen user engagement inside Tencent’s huge 1.4 billion-user ecosystem.

On the enterprise side, Tencent is leveraging AI to cut costs and offer new services. Management highlighted efficiency gains from AI in areas like game development workflows and cloud operations. The company has also launched AI-powered solutions for advertisers and businesses (for example, an automated ad campaign platform that optimizes targeting, bidding, and creative content). These moves help Tencent monetize AI quickly across its existing businesses, rather than primarily selling AI computing power to external clients.

Perhaps most striking is that Tencent achieved these gains with far lower capital spending than its American counterparts. Tencent’s capital expenditure in Q3 2025 was about RMB 13 billion (~$1.8 billion), actually down 24% from the same period a year earlier. For full-year 2025, the company expects Capex to increase from 2024 levels (when it spent around $10.7 billion) but remain modest as a percentage of revenue. Tencent has said its AI-focused Capex should reach the “low teens” percentage of revenue – implying on the order of $10–15 billion annually, a world apart from the $70+ billion that U.S. giants are each spending.

Competition, Risks, and Outlook for the AI Boom

This divergence in strategy raises a larger strategic question for the industry: who wins in the long run—the companies building the biggest AI “railroads,” or those focusing on immediate AI-driven product gains? In the near term, U.S. Big Tech firms’ willingness to spend lavishly on AI infrastructure is cementing their dominance. Their cloud divisions are becoming the essential utilities of the AI economy, forcing smaller players and enterprise customers to rent compute power rather than build their own. This could entrench a few platforms as toll-collectors for AI innovation. From a competitive standpoint, it’s an “AI arms race” that newcomers will find hard to match, potentially reinforcing Big Tech’s market power.

At the same time, the heavy spending comes with risks. Investors and analysts are closely watching the ROI on these AI bets. Thus far, market reactions have been mixed: companies like Amazon and Microsoft saw their stock jump as cloud AI growth reaccelerated, whereas Meta’s shares slipped when it signaled even bigger capex next year, reflecting worries about squeezing profits. Some observers warn of a possible AI investment bubble, drawing parallels to the late-1990s dot-com era. The difference now, optimists argue, is that today’s tech giants are highly profitable and AI demand is real – evidenced by tangible productivity gains and new revenue streams for customers. Still, the pressure is on these companies to deliver visible returns on their enormous outlays. Shareholders will tolerate sky-high Capex only as long as it translates into competitive advantage and earnings growth, not just tech glory.

Regulators, too, are sharpening their gaze. With so much computing power and data in the hands of a few firms, authorities in the U.S. and Europe are exploring new rules to prevent monopolistic control in the AI era. Antitrust cases (such as the U.S. government’s lawsuit over Google’s search dominance) and digital market regulations (like the EU’s Digital Markets Act) are increasingly factoring in the influence of AI and cloud infrastructure concentration. The balance between fostering innovation and curbing excessive power will be a key storyline as the AI boom continues.

Looking ahead, the industry may be entering a new phase of the AI cycle. The “super-cycle” of investment has clearly gone global and shows few signs of slowing; even geopolitical barriers haven’t stopped Tencent and others from pushing forward. The next chapters will likely be about monetization and efficiency. Companies that have spent big on AI will need to prove that these investments can drive sustainable growth or margin expansion (for instance, through automation or new AI services). Those taking a more measured approach will aim to demonstrate they can reap AI’s benefits without breaking the bank.

For now, Alibaba and Tencent’s Q3 suggests that the AI revolution isn’t one-size-fits-all. Silicon Valley’s strategy of spending to build AI infrastructure at almost any cost is one path to dominance. They are showcasing another path: leveraging a massive user base and integrating AI features deeply into an ecosystem to generate revenue quickly, all while keeping capital spending relatively lean. If the coming years reveal that AI’s true value lies not just in who owns the biggest server farms but in who best integrates AI into daily life, their model could offer a blueprint for a more capital-efficient AI era. On the other hand, if AI’s future economics favor scale above all else, They may eventually need to ramp up investments or partner externally to stay in the race.