The overarching theme of the season has been the narrative of productivity-driven growth, which now appears to be entering a phase where its claims are awaiting validation. As we conclude the sixth season of "Fu Peng Talks," the most significant takeaway revolves around the transition from technological advancement to tangible economic application. The central question is no longer about the potential of AI, but about the speed at which this productivity can be integrated into various sectors of the economy.

By the first half of 2026, particularly in the second quarter, the answer to this integration question had become clearer. On the industrial side, major technology companies, which had been the primary sources of capital expenditure over the past few years, saw their free cash flow diminish to zero. This development sent a noticeable shockwave through the market. This period also marked what could be described as the peak of exuberance, with increased volatility in Korean markets, including companies like Samsung and SK Hynix, and retail investor sentiment reaching a cyclical high. Having been a market participant for over two decades, I have witnessed many such cycles, and this level of frenzy often signals a turning point.

This is why I introduced the "peeling the onion" theory, explaining how liquidity impacts markets in the first quarter and how the industrial sector faces pressures in the second quarter. I drew parallels to historical cases where, after a peak in capital expenditure, stress gradually expands outward. It's difficult to point to one specific episode as the most important, as these events are part of a continuous process. Most of our audience understands that we do not produce content for the sake of it; when we do release an update or highlight a specific piece of news, it warrants serious attention.

Market Convergence from June to July: A Cascade from Macro to Micro

Looking back at the past year, the key points were the discussion of liquidity in the first quarter and the repeated emphasis on the free cash flow issue in the industrial sector during the second quarter, alongside the lagging application layer. In this environment, market-level reactions, such as the frenzy following strong earnings from companies like Nvidia, become more pronounced. We had previously highlighted a similar scenario in mid-2024. The pattern is consistent: problems often surface first in the form of high leverage in the market, and historical precedent is highly applicable here.

This isn't about having a perfect track record; it's about not over-communicating when there is nothing new to add. If your investment path has been following the granular breakdown of capital expenditure, you just continue along that line. I don't need to constantly repeat that "the AI era is here." The change in productivity is already happening. However, this process is not the simplistic narrative of "AI is here, so stocks will go up forever." The arrival of this era has a methodology for each stage of development, and highlighting these critical junctures can be valuable. For instance, in June 2026, we cautioned about a potential gap period—massive capital expenditure had been deployed, but applications were not emerging. Any market doubt could flip the game, especially with tightened liquidity and higher interest rates creating an unfavorable macro backdrop. From June to July 2026, we witnessed a complete transmission chain, from the macro and economic layers down to the industrial and market levels, which is worth reviewing. This review traces back to late 2025, where we first discussed liquidity using Bitcoin as an early indicator, and then moved through the economic and industrial layers' free cash flow problems, culminating in the high-leverage clearing at the market level.

AI Industry: Tool Efficiency Realized, Software and Application Layers Still Lacking

The core issue remains AI's struggle to achieve meaningful monetization. There is some, but it is minimal. For example, while China has the world's largest token usage, only about 15% is for coding. AI coding improves efficiency, but humans are still writing the code; it's an enhancement, not a replacement. It's crucial not to confuse the tool layer with the application layer. In the second quarter of 2026, many mistakenly interpreted the surge in coding as an application, but it remains at the tool efficiency level, not a final user-facing application.

The most tangible examples of AI adoption in China are actually AI toys, which are the most direct form. On the consumer side, even the buzz around apps like Doubao is fading; many use it as a search engine, which consumes little computing power and doesn't match the massive upstream capital expenditure. We need to see large-scale, enterprise-level vertical applications. Currently, usage is fragmented across finance, animation, and coding. The largest domestic segment is audio and video production. Overseas, a breakthrough is also lacking in areas like autonomous driving and healthcare. The path to a true software and application layer, where business models mature, must be successfully navigated. We are still mostly in the upstream hardware and midstream efficiency stages, with the software and application layers requiring much more development.

Zero Free Cash Flow: Transition from Spending Your Own Money to Spending Others'

The key development in the second quarter of 2026 is that companies have exhausted their internal funds and must now seek external financing. With high interest rates and high valuations, external financing comes down to equity or debt. Spending your own money invites no scrutiny, but spending someone else's money invites the question: "Why should I let you?" This marks a significant shift in market sentiment. Anyone with business experience understands this. This is why we warned in the first quarter that the free cash flow of major tech firms would turn negative by the second quarter. Investors fixated solely on stock prices may not have understood, but if you've run a business, you know that spending your own money signals confidence. However, when you need to spend others' money, the question becomes about future prospects. If that "hope" weakens and interest rates are high, why would an investor accept the risk when they can get a 5-6% risk-free return? Sentiment turns, the market becomes fragile, and the convergence we saw emerges.

This shift is definitely temporary, which is why we call it a window period. Currently, we don't see the results, but if a breakthrough suddenly appears, investors will likely rush back. This mirrors the sentiment around the release of ChatGPT. In 2022, the US stock market fell over 20%, Nvidia dropped over 60% from its peak, and ARKK fell nearly 70%. But once ChatGPT was released, doubts were immediately dispelled, and capital expenditure on AI began in earnest. The cycle of 2023, 2024, and 2025 followed, which is a normal three-year cycle.

The market's first step is to remove high leverage and let volatility play out. Fortunately, the underlying US economy is relatively healthy, with a stronger old economy and a deleveraged household sector, providing resilience. The future should be a resonance between hardware and software. Hardware runs first, and software may even underperform as investors fear AI will replace it. But ultimately, software will prevail; we will still use software. We are not yet at the stage of a fully autonomous agent like "Jarvis" that can do everything. The progression is from upstream hardware, to intermediate computing and models, and finally to vertical-specific software in various scenarios. The market is currently pulling a distant narrative into the present, which is a sign of a bubble. The software layer needs a restructuring, similar to how we will interact with financial news in the future—not by searching, but by asking questions directly. This transformation and upgrade will occur across every industry, with vertical frameworks being developed over time.

Avoiding Inertia from the Past at Major Cycle Turning Points

Some of our past views, like Japan emerging from its "lost decades," have been misunderstood. Similarly, when I said the US low-interest-rate cycle was over and long-end yields would not be low, possibly around 4.2% on the 10-year Treasury, people applied their old assumptions. These misunderstandings stem from one fundamental issue: holding onto the inertial thinking of a previous dimension at a major cycle turning point. Our discussions about productivity, production relations, and world systems indicate a super-cycle change. If you view the US, the dollar, the financial system, Europe, the UK, or Japan through the old lens, you will be wrong. People's biggest source of experience is their own history. If they've never experienced a market crash, they may believe in perpetual growth, but everything is cyclical. The largest cycles are those involving productivity, production relations, institutional order, and demographics, which are critical.

The benefit of a global perspective is that different economies are at different stages in these cycles, allowing you to learn from history. For example, China, if you look back six or seven years or around the year 2000, without experiencing the real estate or demographic cycles, you couldn't understand their impact. But by observing Korea, Japan, or Hong Kong, or how Americans and Brits speculated in real estate from 2000-2007, you'd see there is no difference. People notice these patterns but believe "this time it's different," yet the outcome is always the same. Experience is the best teacher. To become a seasoned investor, you must go through these cycles yourself. My preference for direct research, such as my trip to Singapore in June 2024 before Nvidia's earnings, revealed high leverage that wasn't in the news. News reports events after they happen; the pre-event analysis requires close research, which isn't widely spread because it hasn't occurred yet. Trading based on news is problematic; news should be used to validate your framework, not to forecast the future. That's why we build frameworks through research and stand on the shoulders of giants, who have better insight and resources. I prefer to ask myself why someone like Bessent is doing research in a particular place and what I might be missing, rather than just accepting their conclusions.

The Essential Skills: Humility, Reading, and Systemic Thinking

If I could pass on only one set of abilities, it would be these: never be arrogant or complacent, always think and engage in systemic analysis. The problem with flash news and short videos is that they are fragmented and don't build a systematic framework. I've always emphasized the importance of reading to form a complete framework and engaging in systemic thinking. Stay humble, avoid unnecessary arguments, and remember that there is always something to learn from others. When someone more experienced or successful speaks or acts, reflect on why they did so, and integrate that into your own framework. This is the best approach.

Preview of the Seventh Season of "Fu Peng Talks"

We have reviewed much of the sixth season, which was built around the framework of productivity, production relations, and world order. The upcoming seventh season promises to be even more compelling. I don't have a pre-written script; I will discuss topics that I find most valuable for you to hear. The core purpose is to accompany you in thinking through problems, not to provide simple answers. Our long-time subscribers understand how to use this content. The response to the sixth season has been the strongest yet, with high engagement and many returning and new members. For the seventh season, we have made significant updates, including five main modules: video columns, a dedicated community, in-depth articles, chart analysis, and exclusive Q&A. We invite everyone to join us for the seventh season.

Risk Disclaimer

Market risk is inherent, and investment requires caution. This content does not constitute personal investment advice and does not consider the specific investment objectives, financial situation, or needs of any individual. Investors should consider whether any opinions or conclusions herein align with their particular circumstances. Any investment decisions made based on this content are at the investor's own risk.