In the past six months, the team has mainly researched the #AI and #RWA sectors, and the most frequently asked question is: ‘Has #AI already passed the hottest "hype period"?’ This is because we see many #AI crypto projects or stocks seemingly experiencing slow growth or no growth at all!

My view is—far from it. The current state of #AI is roughly equivalent to the internet stage in 1995 and 1996. At that time, people were just starting to go online, browsers had only been around for a few years, and companies like Amazon and Google had not yet truly exploded, but the trend was already very clear. #AI is now in a similar situation, just entering an accelerated growth phase.

🧐 Why do I say this?

1️⃣ Computing power remains a bottleneck and the biggest opportunity. To develop AI, large models need to be trained, which requires massive computing power. In the U.S., there’s Nvidia, and chips are sold out. In China, due to U.S. export controls, a domestic industry chain has been forced to emerge. For example, companies like Cambricon have seen their stock prices and market values multiply several times. The underlying logic is simple: without chips, AI cannot progress. This is the most fundamental "shovel business." When I invest in stocks, I pay the most attention to this kind of "certain demand." Regardless of how well AI applications perform, computing power is always essential.

2️⃣ The "Buy China" logic in China. The U.S.-China relationship is actually a "double-edged sword." The U.S. not selling top-tier chips to China seems like suppression, but conversely, it has become a super opportunity for domestic alternatives. Without the blockade, Chinese AI companies might still be using Nvidia's GPUs and would lack the motivation to develop their own chips. Now, it has instead forced out a complete domestic ecosystem. When I look at Chinese semiconductor and computing power-related companies, this is the core logic: self-sufficiency is a national strategy, the market is large enough, and demand is rigid.

3️⃣ The government acts as a super venture capital. If you look at China's model, it is very similar to the new energy vehicles from ten years ago. At that time, the government subsidized batteries and car purchases, effectively pushing companies like BYD and CATL to prominence. Now, AI and robotics are the same; local governments provide subsidies and invest in funds, acting as "early VCs" to lay the foundation for the industry. The rest is left to entrepreneurs to compete in execution. As an investor, my judgment is: the government direction determines the sector, and entrepreneurs determine the leaders.

When evaluating investments, we look at an AI or robotics project, and the two key questions are:

✅ Can it truly solve a pain point? For example, for robots, can they genuinely assist the elderly at home or move goods in a warehouse, rather than just being visually appealing or a big gimmick?

✅ Does it have product-market fit (PMF)? In other words, is this product a necessity or just a nice-to-have? If it only "looks cool" but no one is willing to pay for it, then it won't work. The same goes for AI models. There are models everywhere now, but those that can truly run a business model either meet enterprise needs (like code automation, office efficiency) or consumer needs (like smart assistants, entertainment). When I invest in projects, I am most wary of "show-off" companies that seem technically impressive but lack practical applications.

In terms of long-term trends, I predict that in the next 3-5 years, the demand for AI computing power will be 10 times or even 100 times what it is now. This means:

· Upstream chip and computing infrastructure companies will still hold the most stable "meat-eating" position.

· Midstream large models will become increasingly differentiated; those that can be implemented will stand out, while those relying solely on money and computing power may be eliminated.

· Downstream application scenarios (like robotics) may seem early now, but once a breakthrough similar to electric vehicles is found, growth will be astonishing.

So our investment thinking in the #AI sector is:

First layer: Computing power/chips → The strongest certainty, supported by national strategy, with unlimited demand (the logic is similar for crypto; distributed GPU computing power still has room).

Second layer: Large models → Competing for resources and practical scenarios, careful selection is necessary, as gaps will quickly widen.

Third layer: Applications (like robotics) → High risk, but if you hit the leader, it could yield a hundredfold return.

So when I look at AI now, it’s like looking at the internet and new energy vehicles back in the day. We are far from the point of bubble burst; instead, we have just entered the "accelerated runway." 🧐

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。