Author: Alex Xu; Source: X@xuxiaopengmint

The AI narrative continues to ferment, but the recent performance of related stocks is like midsummer, while others are like winter.

The former is represented by Circle, the issuer of the largest compliant stablecoin USDC, and the latter is represented by Tencent.

There are many reasons for Circle's rise. For example, when it fell from 300 to 49, it had a deep enough correction and accumulated rebound potential. And recently, the scale of USDC has rebounded and reached a new high again.

But the biggest driving force may be that after the explosion of openclaw, people were shocked by the wave of Agents. This concrete impact led to the popularity of stablecoins that are considered very suitable for Agents settlement, and Circle is currently the only pure-blood stablecoin target among listed companies in the world.

The reason for Tencent’s decline is intriguing. Previously, it was criticized for being slow in the AI era and continued to pull back from a high of 680. After the announcement of an excellent financial report + "reduce repurchases and increase AI investment" the day before yesterday, the stock price did not rise but fell. The market’s concerns about its investment in AI have changed from “insufficient” to “excessive”.

So in the AI era, who is more certain, Tencent or Circle?

Charlie Munger once said that you should not do difficult questions when investing. His original words were:

"One of our secrets is not trying to know too much... I keep putting things into the 'too hard' category. Occasionally an easy decision will come up and I'll make it. That's my system: Most things fall into the 'too hard' category, and there are only a few obvious easy decisions that I deal with immediately."

So, who is the easy question on our investment desk now, Tencent or Circle, and who is the difficult question?

Let’s talk about Tencent first.

The Yuanbao AI currently in WeChat actually plays an experimental form of WeChat’s own Agent in the future:

It can actually complete most of the most frequent tasks that ordinary people use AI to do, namely: query and question answering.

However, currently Yuanbao can only interact with content-based information in WeChat, such as text and picture information, video account content, WeChat public accounts, etc. At this stage, it cannot directly interact with mini programs.

But it is not difficult to predict what form WeChat’s own native agent will actually take: through natural language conversations, we can directly let WeChat’s Agent call mini programs to perform various tasks, including ordering takeout, booking hotels, summarizing videos and recordings, reminding schedules, and reporting work.

Mini programs themselves are a huge application ecosystem. Most of the apps we commonly use have mini program versions: Didi Taxi, Meituan, Pinduoduo, McDonald’s, Lalamove, wps, and even Alibaba’s Xianyu and Ele.me.

More importantly: WeChat agent has very strong system-level permissions for small programs, similar to (or even stronger than) Apple's software in the iOS ecosystem. The development of small programs must follow WeChat's specifications. WeChat's agent will definitely not be directly blocked by major apps like Doubao Mobile Assistant and blocked outside the walled garden. Instead, it will interact with various small programs at the system level. All large and small program apps will also inevitably embrace WeChat's official agent to provide users with better services.

Compared with openclaw, which is still very popular but tortures most early adopters to death and makes a large number of ordinary people give up from getting started, WeChat Agent for 1.4 billion users will definitely have better usability and practicality. Only then will it be truly "national agent promotion". This form may not be so geeky or cool, but it is a product available to 90% of the public.

It is highly certain that this Agent product form will be launched at some point in the future. However, before it happens and before everyone sees it, Tencent’s price is still falling, and Mr. Market is stingy in not pricing it.

But for value investors, Mr. Market is what we want to take advantage of.

We want to invest in things that have a high probability or even will happen, but have not happened yet.

Faced with the highly uncertain AI era, is Tencent’s business certainty that low? I think it's okay.

Is it easy or difficult to buy/hold Tencent at this time?

I think it is much simpler than market sentiment reflects.

Let’s talk about Circle:

Looking at some things in the long run, the certainty is actually rather vague.

Can Circle's future business scale smoothly grow to hundreds of billions, or even trillions?

Optimistic bullish voices think it is possible, and their logic is:

1.ai is the future. In the future, agents will be the main users of the Internet and will dominate the huge scale of online transactions

2.ai transactions must use stablecoins based on blockchain settlement

3. usdc is currently the largest compliant stable currency, and the settlement demand of ai will directly promote the growth of usdc

4. Stablecoins have strong network effects, and scale expansion will be self-reinforcing

5. The scale of usdc will grow simultaneously with the scale of AI transactions to several trillions or more

This derivation chain seems to be perfect, but in fact there are great uncertainties in several important links:

1. In the future, agents will be the main users of the Internet and will they dominate the huge scale of online transactions?

Look at this high-certainty link in a long way.

2. Do you need stablecoins for ai transactions?

Not necessarily. The first one is visible to everyone. Existing payment infrastructure service providers and Internet and AI giants that control users are also pushing their own standards, many of which are not based on blockchain, such as the UCP protocol for agent commerce promoted by Google.

3. Is usdc the largest compliant stablecoin?

Currently, it is, but Tether has great energy, is trying to become compliant, and has good political resources. There are also more trafi financial groups that are eager to try out, as well as payment providers like Paypal.

4. Do stablecoins have strong network effects?

In the face of ruthless agents, the differentiation of stablecoins will be compressed. As long as they can be used for payment, the agent will choose the one with the lowest friction and cost. As for whether it is usd1, usdt, usdc, or a stablecoin issued by a bank? It doesn't matter. You said USDC currently has more merchant-side resources? Payment channel providers such as visa and stripe have much larger merchant networks and channels than Circle, and can also help other stablecoins access it. As long as settlement is convenient and friction is low, merchants don’t care which stablecoin they receive. The network effect of stablecoins may not be that strong in the era of agent-led payments.

5. How many trillions will the scale of usdc grow simultaneously with the transaction scale of ai?

When there is a high degree of uncertainty in the assumptions of 2-4, and when they are added together to derive the conclusion of 5, the uncertainty has become blurred.

Faced with the highly uncertain AI era, is Circle’s business certainty that high? I think it's okay.

Is it easy or difficult to buy/hold Circle at this time?

I think it's far more difficult than market sentiment reflects.

One family’s opinion, for reference only.