If we use the same valuation methods for AI large model companies as we did for internet companies in the past, we may miss out on an entire era.

However, this does not mean that all large model companies deserve premium valuations.

For a long time, we have been accustomed to benchmarking startup tech companies against leading firms and then labeling them accordingly. For example, in the chip sector, some startup chip companies are easily compared to NVIDIA and are given dream-like valuations that far exceed their actual revenue levels.

In reality, the gap in terms of company quality is quite significant. NVIDIA, as a global leader in chips, has products that are easy to sell and delivers results easily, both from a demand and market perspective. The high valuations the market assigns are often digested within a year or two.

Therefore, despite NVIDIA's market capitalization surging more than tenfold in recent years, its dynamic price-to-earnings ratio stands at only 46.5 times as of now. Meanwhile, some chip companies benchmarked against NVIDIA, which lack strong products and revenue, are given valuations exceeding 100 times, creating a clear and absurd gap.

Moreover, from a product perspective, as leading tech companies increase their capital expenditures, the gap between them and mid-to-lower-tier companies is widening rather than shrinking. With faster product iteration from leading companies, the market will favor higher-quality products. Under these circumstances, stories of small companies making comebacks will become increasingly rare.

Take the current AI industry, for example. The market does not believe in comebacks from small companies' products. The underlying competitiveness of AI companies often lies in computing power, algorithms, and models—areas where startups struggle to compete with giants. Currently, in the global AI landscape, capital expenditures from leading tech companies like Google, OpenAI, Meta, and Amazon are rising yearly, and computing power is generally in excess—an advantage that small companies lack.

It should be noted that today's AI competition requires substantial, tangible investment with no shortcuts. Therefore, while AI is currently the world's biggest trend, it does not mean that small startups can break through.

Right now, the market is arbitrarily assigning fantastical valuations to some startup AI companies, especially those with revenue not exceeding the billion-dollar mark being valued in the tens of billions. This is undoubtedly an irresponsible approach because valuations ultimately need to be justified by performance. If results are not delivered, the consequences for both the market and investors could be disastrous—a point investors should be wary of!

Doubling Down! All-In

The competition in AI has reached a fever pitch.

The increasing capital expenditures of leading tech companies indicate that, for the giants, there is no turning back in this AI race. As a result, some leading companies are even taking on massive debt to fuel their expansion.

Media reports indicate that Google's parent company, Alphabet, plans to raise $20 billion through its dollar bond issuance, exceeding the initially projected $15 billion. Sources familiar with the matter said that Alphabet's dollar bond issuance has attracted over $100 billion in orders.

Additionally, Alphabet is also promoting its inaugural bond issuance plans in Switzerland and the UK to investors. The UK issuance may even include a rare 100-year bond.

It is reported that this will mark the first attempt by a tech company to issue such long-dated bonds since the dot-com bubble of the late 1990s.

Notably, leading tech companies continue to ramp up their efforts. According to statistics, the combined expenditures of just four tech giants—Google, Meta, Microsoft, and Amazon—are expected to reach $650 billion this year, with funds primarily allocated to expanding AI infrastructure.

As tech companies place crazy bets, the capital market has not stood idly by.

In the third quarter of 2025, Berkshire Hathaway, a company not typically interested in tech stocks, took a significant stake in Google, making it one of its top ten holdings. Some domestic private equity giants have also begun to heavily invest in the sector.

Recently, Dan Bin's Oriental Harbor Overseas Fund submitted its 13F filing for the fourth quarter of 2025 to the U.S. Securities and Exchange Commission (SEC). The filing revealed that Dan Bin liquidated positions in several stocks, including Alibaba, Coinbase, Netflix, Astera Labs, BitMine Immersion Technologies, Broadcom, and TSMC, during the fourth quarter. His holdings became further concentrated in tech giants, particularly with a substantial increase in his stake in Google, which became his top holding.

By the end of the year, Oriental Harbor Overseas Fund held approximately $406 million worth of Google shares, accounting for about 31% of its total U.S. stock holdings—significantly higher than NVIDIA's 18% weighting. The fund held 1.2935 million Google shares at the end of the period.

Notably, Dan Bin not only bought Google shares aggressively but also significantly increased his position in a leveraged ETF that doubles exposure to Google, with an increase of 117.61%, further boosting his Google holdings through leverage. It is reported that Oriental Harbor Overseas Fund managed approximately $1.316 billion at the end of the period, up slightly from $1.292 billion at the end of the third quarter of the previous year.

Another private equity giant, Jinglin Asset Management, also substantially increased its stake in Google. According to the U.S. stock holdings data submitted by Jinglin Asset Management (Hong Kong) Limited, Jinglin's overseas entity, to the SEC as of the end of 2025, the private equity fund significantly increased its holdings of nearly 930,000 Google shares in the fourth quarter of 2025, making it its new top holding. The position's market value reached $842 million at the end of the quarter, accounting for over 20% of the portfolio.

As of the end of 2025, Jinglin Hong Kong's top five holdings were Google, Meta, Pinduoduo, NetEase, and Manbang Group.

Google's Bid for the Global AI 'Throne'

Regarding heavy investments in AI, Dan Bin stated in a media interview: "Looking at a longer time horizon, this year may simply be a year of accumulation. We still firmly believe that artificial intelligence will drive a decade-long trend, and the broad direction of AI will continue this year."

Kan Jian Finance believes that with the rapid approach of the AI era, AI is quickly becoming a market consensus.

Notably, Duan Yongping also bought stakes in leading AI companies in the fourth quarter. According to his holdings, Duan not only holds a significant position in NVIDIA but also increased his stake in Google.

He also stated that in the new year, he needs to seriously learn how to use AI, hoping to reach a point where he understands it well enough to make heavy bets.

It should be noted that being optimistic about the future does not mean there is no fear of the present. Currently, with the surge in leading AI companies' stock prices, market sentiment toward these companies has become divided. Additionally, their massive capital expenditures have raised some concerns in the market.

In early February, Google's parent company, Alphabet, released its fourth-quarter earnings report, showing revenue of $113.828 billion, up 18% year-over-year, exceeding Wall Street's consensus estimate of $111.4 billion. Diluted earnings per share were $2.82, also surpassing market expectations of $2.63. Net income for the same period reached $34.455 billion, up nearly 30% year-over-year.

From a full-year perspective, Alphabet's revenue exceeded $400 billion for the first time, reaching $402.836 billion, with net income of $132.17 billion—multiple core metrics hit all-time highs.

The cloud business, which attracted the most market attention, reported revenue of $17.66 billion in the fourth quarter, up 48% year-over-year, significantly higher than analysts' expectations of approximately $16.2 billion. Google attributed this growth to enterprise-grade AI infrastructure demand, AI-native solutions, and sustained growth in core GCP products.

Notably, Google Cloud's operating profit reached $5.313 billion in the fourth quarter, more than doubling year-over-year. The cloud business's "breakthrough" indicates that after crossing the profitability threshold, it will substantially contribute sustainable cash flow to the company.

The cloud business's outperformance has also emboldened Google, leading to increasingly large capital expenditures. Alphabet expects capital expenditures to reach $175 billion to $185 billion in 2026, with a midpoint near $180 billion (approximately RMB 1.2435 trillion), significantly higher than Wall Street analysts' previous expectations of around $120 billion and nearly double the 2025 level.

Addressing market concerns about massive capital expenditures, Alphabet CEO Sundar Pichai stated at a recent AI summit in India, "In some contexts, people talk about this as an industrial revolution, but it's happening 10 times faster and on 10 times the scale."

He added, "In the cloud business alone, the backlog doubled year-over-year over the past year, reaching $240 billion. This demonstrates the potential for returns on another front. Therefore, we are investing to meet this demand."

Previously, Pichai also said that AI investments and infrastructure construction are "fully driving revenue and growth," and that increasing capital expenditures in 2026 is to meet existing customer demand rather than prematurely tapping into potential expectations.

Kan Jian Finance believes that as AI competition enters deeper waters, the era of storytelling is over. In the future, the market's scrutiny of AI products will become even more ruthless. Whether they can gain market and user favor will be key to market valuations. Moreover, with the accelerating pace of AI iteration, market eliminations will continue to speed up. In this context, if startups fail to quickly capture the market and deliver results ahead of leading AI companies, they will be swiftly left behind by the giants.