This article source: Gelong City North Xu Gong

Data support: Pythagorean Big Data

AI bubble is becoming the most divisive consensus in the global market. Dalio says the bubble is very high, Huang Renxun says the opportunity has just begun; one sees the overheating of the capital market, while the other sees the beginning of a productivity revolution.

The real issue is not whether there is a bubble in AI, but what remains after the bubble bursts. The internet bubble of 2000 caused the Nasdaq to plummet, companies to go bankrupt, and wealth to evaporate, but it also left behind underwater cables, broadband networks, and cloud computing infrastructure, ultimately supporting Amazon, Netflix, YouTube, and mobile internet.

Today's AI also stands in a similar position. On one hand, hundreds of billions of dollars are invested in data centers, electricity, liquid cooling, optical modules, and GPUs; on the other hand, there is a huge gap where application revenues have yet to be fully realized. The bubble obviously exists, but the underlying productivity is not inflated. When token costs plummet and intelligence is accessed like electricity and water, AI will no longer just be a chat tool but will enter the real workflows of coding, healthcare, finance, law, manufacturing, and research. The market will wash away shell companies and PPT entrepreneurs but will not reverse the direction of AI+. The bubble will burst, but the industry will remain. Below, enjoy:

In recent days, the market has experienced violent fluctuations, and the theory of the "AI bubble" has been widely discussed.

• Bridgewater's founder Dalio said: There is a bubble in the AI market, and the level is "relatively high."

• NVIDIA CEO Huang Renxun said: There is a huge opportunity in AI, and the demand for computing power has just begun to explode.

Who should we trust?

Both of them are not wrong.

Is there a bubble in the AI industry? It undeniably exists.

However, bubbles in the tech sector are often the only way society can pay tribute when faced with disruptive advanced productivity. It is not purely a derogatory term.

In the long run, this is an inevitable phenomenon at the very beginning of advanced productivity.

Many people compare the current situation to the internet bubble of 2000, feeling anxious. The internet bubble at that time did indeed lead to the Nasdaq crashing nearly 78%, with over $5 trillion in wealth evaporating.

But twenty years later, which industry can do without the internet? Today, the value of the internet industry has long exceeded that of the bubble period.

The AI bubble, at least on the surface, looks like a similar situation. The bubble existing in the capital market cannot hinder the fact that almost all industries in society are being actively empowered by AI.

AI+ is the trend of the future. Just as all industries today cannot do without the internet, all industries in the future will also be inseparable from AI.

01 The "IQ tax" that innovation must pay

In that era when companies could go public and raise money just by having a .com in their name, between 1995 and 2000, the Nasdaq soared nearly 600%. Then came a financial storm that lasted for two and a half years.

Those famous names back then, such as the software company MicroStrategy, plummeted 62% in one day due to accounting scandals and over-hyping; Pets.com (selling dog food online) and Webvan (the pioneer of fresh e-commerce) went bankrupt... In panic, almost everyone blamed the internet as a scam.

However, the physical infrastructure that was precipitated by the excessive squandering of speculative capital often nourishes the super giants of the next era at a very low cost. The reason the bubble burst is not due to the internet technology itself but because the physical construction speed of the infrastructure could not keep pace with the market rhythm.

For example, those telecom companies that were flourishing at the time (such as WorldCom, Global Crossing), spent heavily on laying global underwater cables and optical dense wavelength division multiplexing networks that did cause their own bankruptcies, but these cheap "information highways" later became perfect breeding grounds for the rise of Netflix, Zoom, and the mobile internet.

Without the global frenzy of investment in telecom infrastructure around the year 2000, there would not have been the later explosive growth of YouTube's video streaming, nor would there have been the subsequent cloud computing infrastructure.

The most typical example is Amazon. Its stock price dropped from its highest point of $107 in 1999 to $7 in 2001, a decline of over 90%. But it survived because its underlying business logic of "restructuring retail through the internet" was aligned with the direction of advanced productivity.

This is the classic Amara's law: overestimating the short-term impact of a new technology while severely underestimating its long-term impact. At the beginning of a technological revolution, the frenzy of speculative capital inevitably leads to over-investment, creating bubbles. This is the "IQ tax" that innovation must pay. But once the bubble dissipates, what remains will be an unbreakable advanced productivity.

02 Why do AI spending by enterprises not decrease but increase?

Looking back to 2026, the AI industry's bubble seems even larger.

Among the five major cloud service providers, such as Amazon, Google, Meta, Microsoft, and Oracle, capital expenditures are expected to reach $690 billion in 2026, with total AI infrastructure investments projected to reach $5.3 trillion by 2030. Among these, only about 25% will be spent on GPUs, with the remaining 75% going towards physical infrastructure: liquid cooling systems, power transmission, network switches, optical modules, and land.

In terms of revenue, all leading pure AI companies like OpenAI, Anthropic, Cohere, Mistral, and Perplexity are expected to collectively generate no more than $40 billion in total revenue in 2026.

Investing nearly $700 billion in the foundational layer, and getting back only hundreds of millions from the application layer. This serious asymmetry, what else can it be if not a bubble?

One cannot simply come to this conclusion bluntly. There is a key point that cannot be ignored:

• In March 2023, when OpenAI released GPT-4, the mixed cost per million tokens input was about $30.

• By April 2025, with the optimization of model architecture and improvement of reasoning computing power, models of the same intelligence level will see prices plummet to $0.1-0.15 per million tokens.

According to Stanford University's "AI Index Report" and data from TokenCost: AI reasoning costs have fallen over 99.7% in the last two years.

According to traditional linear thinking, if costs plummet, then enterprise AI spending should decrease as well. But in reality, enterprise AI cloud spending tripled between 2024 and 2025.

Why?

Because when the marginal cost of "intelligence" approaches zero, AI is no longer just a simple text summarizer or companion machine, but has entered a new era of intelligent agents and multimodal enhanced retrieval. Enterprises begin to allow AI agents to automatically run thousands of tasks, to write code, to scan millions of legal contracts, and to simulate biological experiments.

Cheap tokens unlock a vast amount of long-tail demand that was previously constrained by costs and unable to be commercialized.

In this regard, if we compare NVIDIA of 2026 with Cisco, the dominant internet hardware company of 2000, we can see hints. The ecological niche of both is extremely similar, but their underlying financial health is worlds apart.

(Hardcore financial comparison between NVIDIA and Cisco)

This precisely confirms the economic "Jevons Paradox": technological advances improve energy utilization efficiency, but instead of reducing energy consumption, cost reductions lead to greater demand.

Even after experiencing the so-called "DeepSeek moment" at the beginning of last year, the market quickly regained clarity in the following months: the better the algorithms are optimized, the lower the threshold for enterprises to adopt AI becomes, resulting in an exponential increase in total computing power consumption.

Because of this, AI has the potential to gradually embed itself into almost all traditional industries. Just as in the past twenty years, all industries have embraced the internet+. From SaaS software to biomedicine, and advanced manufacturing robots driven by embodied intelligence, in 2026 almost all industries are embracing AI+. No one will discuss "whether we should use AI," but rather worry about "Is our data properly cleaned? Do we have enough API call quotas? Is the RAG architecture optimal?"

Currently, there is indeed a bubble in the AI industry. However, for enterprises, if you do not embrace the bubble, you will be crushed by the times. This point has already been proven by nearly two decades of internet history.

03 The deep evolution of the market: from infrastructure to applications

Currently, we are undoubtedly at a very critical juncture in the technology lifecycle: on the eve of the "trough of disillusionment" on the Gartner Technology Maturity Curve, or a turning point in the theory of "Technological Revolution and Financial Capital".

The AI bubble has actually already begun to burst, but many people are not aware of it. Just a few rookies, writing dozens of pages of PPT and wrapping an OpenAI API, can raise money. Now, the tide has receded, and these companies without a moat and only concepts are dying in large numbers.

This is the market purifying itself and a manifestation of the bubble bursting. But this is just the surface. The deeper logic of the market is undergoing three profound transformations:

First, the value shift from CapEx to OpEx

Currently, most of the money has been earned by the shovel sellers, NVIDIA, TSMC, and those selling optical modules and server liquid cooling equipment have captured most of the dividends. But as computing power gradually becomes "infrastructure" like water and electricity, the true excess profits will gradually shift to the application layer. These are the AI native enterprises that can use extremely low-cost tokens to really solve pain points in vertical industries and reshape business processes (OpEx optimization).

Second, valuation multiple compression and performance digestion

The market values of AI infrastructure are overly high, but this does not necessarily mean a crash is imminent. In many cases, the rapid growth of enterprise profits can gradually digest high valuations by exchanging time for space. As long as the revenue growth of cloud computing giants keeps up with the depreciation speed of capital expenditures, this game of passing the parcel can evolve into an unprecedented industrial upgrade.

• For example, global automotive manufacturing giants and chip giants, through the introduction of end-to-end AI twin technology, have shortened the R&D to mass production cycle of new products by 35%, improving the overall efficiency of the entire line by 18%.

• In the financial industry, for example, quantitative trading, risk control, and credit evaluation in 2026 are fully dominated by multimodal agents. AI is not only processing macro expectations with microsecond timestamps but is also deeply involved in asset pricing at every micro level.

• In industries such as law, healthcare, and auditing, highly reliant on seasoned expertise, AI has also transformed from "junior assistant" to "partner-level expert".

ChatGPT, Gemini, and Claude has over 1 billion active users, with a substantial portion using it as an alternative tool for daily high-intensity intellectual labor. Including you and me. All of the above are tangible occurrences that everyone can see.

04 Conclusion

Looking back at the grand history of technology, Schumpeter's idea of "creative destruction" is always playing out.

The capital market is always impatient, hoping to invest $1 today and earn back $10 tomorrow. When nearly $700 billion in infrastructure investment cannot be fully transformed into application-side profits in the short term, the market is bound to welcome a round of brutal reshuffling. Getting rid of those speculative shell companies that only rely on giving PPT presentations, leaving behind those that truly have technological depth and practical applications.

After the reshuffling, those cheap and massive computing centers and highly optimized model algorithms will serve various industries at extremely low prices.

After 2000, humanity entered the digital age where all industries cannot do without the internet. Today, we are also irreversibly moving towards an intelligent prosperity era where all industries are vertically integrated and powered by AI.

Amid the noise of the bubble, the underlying productivity potential is not inflated at all.