Tencent: The Art of Boring Brilliance

Tencent earnings Q2 2025 WeChat ads games cloud AI

Tencent has become so predictable, so relentlessly competent, that it’s almost dull to analyze them quarter after quarter. But in China, where management missteps are a national sport and regulatory tripwires are everywhere, that kind of dullness is a superpower.

As always, I’m not here to dissect every revenue line or margin delta. I try to focus on what genuinely stood out to me—especially things that, in my view, are under-articulated or buried beneath the usual surface-level commentary.

Anyway, here are their revenue, earnings, and margin developments in case you haven’t seen them yet.

So let’s actually take a step back.

What Makes Them So Predictable?

Tencent’s management is outstanding. Ok, everyone knows this—ask any serious China watcher to name exceptional management teams, and Tencent always makes the short list. But I still don’t think people grasp how unusual they are. I certainly keep underestimating them.

That became obvious again while reading other Chinese tech earnings reports. Most of them are exercises in filtering signal from noise. Management teams openly mislead, present problems as if they were positive developments, or just waste everybody’s time.

Then there’s Tencent. Their calls actually teach you something. They’re not trying to mislead.

I actually look forward to their earnings calls. That alone tells you everything.

Playing the Government Game Perfectly

Want to see management strength in action? Look at how Tencent handled Beijing’s tech crackdown.

When the government made clear that tech companies shouldn’t be making too much money, Tencent didn’t push back. They scaled back quietly. Revenue growth even turned negative. And during that pause, they got to work fixing their business fundamentals.

They slashed low-margin cloud contracts. Phased out unprofitable low-margin revenue streams. While growth slowed, margins improved dramatically.

They didn’t whine. They didn’t fight it. They didn’t try to lecture the central bank about risk management. (Hi, Jack.) They didn’t go on a multi-billion RMB shopping spree for random failing businesses. (Hi, Daniel.) They didn’t light money on fire launching a new price war in the middle of involution when the government is carefully looking. (Hi, Richard.)

They just stayed out of trouble. And in China, that’s a strategic asset.

Promises Made, Promises Kept

What continues to impress me is how precisely Tencent follows through on what they say.

About two years ago, they began talking about margin improvement. And the results are now showing up. For example, Cloud margins went from the low 30s to the low 50s.

Then, after successfully juicing operating leverage and buying back shares, they shifted again—this time toward reinvestment. When the government gave the green light for tech to start growing again, Tencent immediately began spending again, especially in AI, which—helpfully—is exactly what the government supports.

They’re not just building AI with huge Capex spending to sell it. They’re building AI that massively benefits their own operations.

Games: Quietly Gaining Power

I haven’t talked about games much in recent months, but this quarter is worth highlighting.

What stood out to me is how incredibly self-confident they’ve become. A few quarters ago, they bluntly talked about firing the monetization teams of certain underperforming games. And now? Monetization improved. Clearly.

They figured out the playbook to turn games into evergreen franchises. They’re no longer panicking over weak quarters. They simply adjust and fix the problem. It reminded me of what Garena did with Free Fire—slump, reset, comeback. Tencent seems to be playing that same kind of long game now, only at greater scale and with better tools.

But what’s most exciting is that games now benefit from several major tailwinds at once:

• Regulation is shifting in favor of game developers—Tencent said that app store commissions in China are tilting more in favor of developers now, which will boost margins. International lawsuits also tilt toward developers.

• AI has a huge impact. It is improving player engagement and monetization. With the help of AI, development cycles are getting shorter, and development costs are falling.

The combination is powerful—and still underappreciated, since much of it hasn’t flowed through the P&L yet.

Ads: Growing Fast Without Raising Ad Load

Let’s talk advertising. Tencent is playing the same long game here too—and doing it with their usual discipline.

Management is deliberately under-earning. Not because they’re incompetent, but because they’re owners. The kind that would rather own a thriving business five years from now than impress a UBS intern with a beat-and-raise quarter.

Their video accounts currently show ad loads in the mid-single digits. For comparison, Douyin and Kuaishou are comfortably in the low to mid-teens. And still, Tencent grew ad revenue by 20% year over year.

That growth didn’t come from slapping more ads in people’s faces. It came from better targeting, driven by AI. Click-through rates are up.

During the call, management said that if AI investments ever became too costly, they could always increase the ad load and start printing cash. They’re not doing it now. But they could. That kind of optionality, just sitting there unused, is rare.

Also worth noting: even when Alibaba, JD, and Meituan reduce ad budgets during their latest voucher-burning bonanza, Tencent’s ad business holds up. It’s now so large and diverse that it absorbs those cuts easily.

Optionality. Resilience. Owner mindset. All present.

The Quiet Construction of a Mini-Shop Empire

Tencent said very little this quarter about mini-programs or mini-shops. Which, of course, is the real signal. Because everything they laid out in earlier quarters is unfolding—just with their usual silence and precision.

They’re building this thing slowly, methodically, brick by brick. No fireworks. Just execution.

Latest update: merchants can now import SKU libraries directly from mini-programs into mini-shops. Sounds trivial. But it’s another puzzle piece locking into place.

They already own the infrastructure. All of it.

• They have WeChat—the most intimate social graph in China. They know who talks to whom, when, how often, and what gets forwarded.

• They have Video Accounts—they know what people are watching, rewatching, pausing, skipping, and sharing.

• They have Mini-Programs and Mini-Shops—they see what users browse, linger on, compare, abandon, revisit, and finally buy.

• They have WeChat Pay—they see where the money goes, how frequently, and for what.

• They have tentacles in music streaming, long-form video, news feeds, and more. Every swipe, scroll, and skip is data.

What they’re quietly building is a commerce engine with full-spectrum visibility—upstream intent, midstream browsing, downstream transaction. No other platform in China has that kind of end-to-end coverage.

Now add in social gifting and group commerce mechanics, and suddenly the whole thing starts to resemble early-stage Pinduoduo—only this time with 1.3 billion users already inside the ecosystem.

Tencent won’t shout about it. But they’re laying foundation after foundation. And they’re not in a rush, because they know exactly what they’re building. I’m still waiting for the day Pony Ma says in the earnings call, “Release the Kraken.” Until then, they will be quiet.

Cloud and CapEx: Smart Spending, Not All-In

Two quarters ago, Tencent said GPUs would come online in Q2. They did. The company is now renting out computing power and expanding its international cloud business.

Capital expenditure rose 149% year over year, which may look aggressive at first glance. But sequentially, it actually declined by 32%. This isn’t uncontrolled spending. It’s part of a staged rollout aligned with internal needs.

Management had already cautioned last quarter that much of the AI-related investment may take up to two years to translate into visible financial returns. That said, some of the benefits are already showing up, as I discussed in the ads and games sections.

Some have suggested this might become Tencent’s “Meta moment”—a broad, AI-led transformation across its social and content products. That may turn out to be accurate. But the company’s approach is clearly more measured. As management put it:

We have to spend smartly—rather than just saying, ‘Oh, we’re going all in’ and spend on buying a lot of chips, hiring a lot of people, and doing a lot of marketing.

Tencent Q2 2025 call

So yes, CapEx is rising. But the real story is in how carefully that money is being put to work. Ambition is clearly there—but it’s paired with control.

Just a final reminder: we shouldn’t assume that China’s cloud business or LLM monetization will follow the same path as the West. Take long-form video as a warning. In China it’s a structurally bad business. Look at iQIYI, Baidu’s spin-off. Over 100 million subscribers, yet a market cap of just $1.9 billion. Netflix, with around 300 million subs, is valued at $525 billion. The gap isn’t just scale. It’s that one business works, and the other never really did. Okay, it’s also not helpful that this is a Baidu spin-off. As always, Baidu is excellent at seeing opportunities but lacking the operational skills.

Summary:

So those are my takeaways from this quarter. Tencent isn’t making noise—they’re just executing. The WeChat ecosystem keeps opening up new monetization channels, from ads to e-commerce, and AI is quietly improving every major business line. It’s all moving in the right direction. No need to say more.