Source: "Phil Rosen"

Organized by: Felix, PANews

Macro analyst Jordi Visser recently guest appeared on the "Phil Rosen" podcast, discussing AI, macroeconomic trends, and investment strategies for the coming years. Jordi Visser pointed out that the productivity improvement and profit growth brought about by AI are the core driving forces of the current economy, and traditional GDP metrics fail to reflect the value of AI; he elaborated on the investment logic of Micron and Eli Lilly, believing that Micron has a 60% to 80% upside potential. In addition, Jordi Visser believes that 2027 will be a key year for the explosive commercial application and payment network of crypto digital assets.

PANews has organized the highlights of the interview.

AI is the sole engine driving U.S. Treasury yield and capital expenditures

Host: You have long emphasized that AI is the main storyline of the current macroeconomy in many ways. But even today, many people are still discussing the Federal Reserve's interest rate policy, oil prices, and those traditional macroeconomic indicators. Please share your latest thoughts.

Jordi Visser: My viewpoint has not changed at all. In the current economic system, only one thing truly matters: everything happening in the field of AI. If there is any change, it is that this trend has intensified further.

I can fully argue (and have been emphasizing) that the comprehensive rise in government bond yields this year is clearly due to the growth of nominal GDP, and the growth of nominal GDP is closely related to capital expenditures (CapEx). This includes all the debt generated by the market this year and the enormous capital demand. For example, the investment in AI and data center construction is highly concentrated in only a few enterprises. Due to the urgent need for funds, these companies are rapidly entering the market for financing. This demand exhibits a distinct "front-loading" characteristic: to meet the backlog of construction needs, these companies are scrambling to raise capital. Therefore, in my view, the comprehensive rise in real yields this year is entirely driven by developments related to AI.

Host: So you believe that most people on Wall Street have a biased view of the direction of interest rates. Can you discuss your insights?

Jordi Visser: If you listen to the so-called "AI bubble theorists" on X, everyone will tell you that CoreWeave is about to have problems, Oracle is on the verge of collapse, and so on. But if you listen to the executives truly managing the credit departments in major banks, the story is completely different.

This past first half of the year, insurance companies have been very willing to buy bonds with maturities exceeding 10 years. But soon, they faced extremely serious concentration risks. This is a very unique historical period, as companies heavily leveraged in the market are precisely the largest market cap giants in U.S. equity markets.

This means that if an insurance company or large fund already has a huge exposure to equity in Microsoft or Google, and these giants keep coming to the bond market to raise billions of dollars, the risk control department will immediately alert to concentration risks. If you hold their debt in a fund as well as their stock, it creates serious risk control limit issues.

Therefore, credit officers from major banks have pointed out that concentration is currently the biggest core contradiction in the credit market. That is why you see many large companies' credit default swap (CDS) spreads widening: institutions buy CDS merely to hedge concentration risks and engage in portfolio hedging, which absolutely does not mean that the AI bubble is bursting. It simply means that both the credit market and the stock market have unprecedentedly high concentrations internally.

This situation will eventually stabilize. This week has shown important signals, first, that the 10-year U.S. Treasury yield has begun to decline, and second, the Treasury's toolbox and supply-side adjustments.

Host: You mentioned that many headwinds next year will turn into tailwinds. What specifically do you refer to?

Jordi Visser: Take Oracle as an example. Since September last year, Oracle has become the focus of the market, with its CDS spreads significantly widening because they needed to borrow massive amounts of funds in the equity and bond markets. Over the past 12 months, their equity plus debt financing has reached approximately $60 billion.

However, Oracle's guidance given this week at Investor Day indicates that the company will achieve positive free cash flow by 2028 and generate significant positive free cash flow in 2029, at which point only about $20 billion of new capital will be needed.

Once companies complete these significant initial capital investments, the price of computational power has increased substantially compared to when they originally decided to build. As Gavin Baker has discussed in depth, people overlook a crucial fact: these giants are locking in construction at "yesterday's costs," while the price in the computational market continues to rise. Once monetization starts, they no longer need to borrow as much money.

Oracle is the first company to announce this, but as other large firms roll out their data centers, they will experience the same growth curve. As Gavin said, we may be at the peak of external financing for super cloud providers. Data from Goldman Sachs also shows that the issuance growth rate of investment-grade bonds is about to peak.

In addition, special purpose vehicles (SPVs) are becoming a new weapon for financing giants. For example, Musk completed a $40 billion SPV financing through SpaceX and Apollo. Just like the past commercial real estate cycle, once debt is gradually digested and bad assets are written off or liquidated, the entire market will undergo a new pattern.

Host: Is there anything to worry about with the negative free cash flow shown in Amazon or Alphabet's earnings reports?

Jordi Visser: I have maintained relative caution regarding publicly traded traditional cloud providers throughout this year, and that will not change. I am not optimistic about publicly listed software companies purely based on code because they always face competitive threats from AI.

However, the panic about free cash flow is completely nonsense. If you examine the financial statements closely, you will find that these super cloud providers have very strong balance sheets. When your business revenue is growing at a rapid pace and you are investing cash into future infrastructure construction, what is there to worry about if free cash flow is temporarily negative? Unless you stubbornly believe that there is no real demand for AI and that it is entirely a bubble. Those who shout "AI will not generate any revenue" are usually people who never use AI products.

Microsoft lays off 60,000 employees, more companies will follow suit

Host: Microsoft recently offered a buyout scheme to 60,000 employees (27% of its total workforce). In your recent research report, you wrote: "In past economic cycles, governments often supported the labor force when capital was in trouble; however, in the AI cycle, governments might end up supporting the labor force precisely because capital is too successful." Please help us connect these clues.

Jordi Visser: Think back to the nature of traditional economic recessions: in a credit-debt-based fiat currency system, companies are heavily indebted, and when the market shuts off financing channels, corporate debts explode. The government, after interest rate cuts fail, has to step in to rescue these firms to keep them alive.

Why should we stop recessions? Theoretically, recessions are a process of liquidation and rebirth. If the government keeps pulling out money to rescue the market, the market will be filled with "zombie companies" that cannot survive independently.

But Microsoft's situation is entirely different. In the past, when companies laid off workers, there were always other industries expanding. But when Microsoft offers buyouts to 27% of the workforce, it suggests that Microsoft management clearly forecasts that in two, three, or four years, they will no longer need so many employees.

Meanwhile, Microsoft's revenue is growing rapidly. If all 60,000 accept buyouts, Microsoft’s employee count will revert to pre-pandemic levels from 2020; yet its revenue has skyrocketed by 70% compared to 2020. This means that Microsoft’s "revenue per employee" is showing a parabolic explosion.

This is the true productivity leap brought about by AI, but it poses an unprecedented challenge for the government. First, in the past, government bailouts occurred during economic downturns and soaring unemployment; whereas now, U.S. stocks are at historical highs, and initial jobless claims are at historical lows.

Second, tech giants find that rather than financing at high costs in the market, it is preferable to buy out 60,000 employees. Calculating an average annual cost of $400,000 per person, this could directly save $24 billion in cash flow each year.

Finally, government fiscal revenue is extremely reliant on personal income tax (with personal income tax rates far exceeding corporate tax rates). When large companies significantly reduce high-paying positions and choose to retain profits within the company, the government faces a massive fiscal deficit while the personal tax base is severely eroded. The tension between the government and capital has been pushed to the forefront.

Host: Do you expect more large companies to take similar buyout or layoff actions in the future? Companies like Microsoft, valued at $4 trillion, have virtually no financial pressure; this is entirely due to improving efficiency and momentum. Aren't those firms not performing as well as Microsoft under greater pressure?

Jordi Visser: Highly tech-savvy companies like Microsoft and Block can most quickly integrate AI into their workflows, so they will be the first to lay off employees. For traditional Fortune 500 companies, the process of absorbing AI will be slower, but the trend is irreversible.

Host: If such large-scale buyouts spread in the tech industry, does it mean we are approaching the future that Musk envisions: the government must issue "Universal Basic Income" (UBI) to everyone?

Jordi Visser: I dislike simply labeling it as "UBI," as this term may trigger many people's stereotypes. I prefer to define it as: a profit redistribution from capital to labor.

For example, the government may significantly increase the corporate tax rate on super giants like Microsoft that exceed specific thresholds of "revenue per employee," while reducing the personal income tax rates for the middle and lower classes. This would encourage small businesses and entrepreneurial spirit while curbing the unrestricted expansion of super giants.

Future government intervention in large companies will surely increase significantly. Coupled with what Musk said about the "material abundance brought about by the significant price decline of goods and services due to AI," the overall affordability issues will be resolved through tax restructuring and deflationary dividends.

The GDP accounting method is no longer valid; AI productivity is reflected in corporate profit margins

Host: Many bearish macro scholars often raise a rebuttal: if AI is so impressive, why can’t we see the benefits brought by AI in the macroeconomic "productivity data"?

Jordi Visser: Productivity improvements have long been evident. They are directly reflected in corporate profit margins.

Over the past nearly 20 months, except for the healthcare sector, U.S. large firms have added almost no new hires. In the history of capitalism, it is unprecedented for corporate profits and revenues to grow rapidly while hiring completely stagnates. If there are no productivity improvements, companies must hire to expand revenues. The expansion of profit margins is the most direct evidence of AI productivity.

On the other hand, the GDP calculation formula itself is fundamentally outdated. GDP was designed for the Industrial Revolution, measuring the physical output of tangible goods, and cannot accurately assess intangible assets such as AI, software, and intellectual property.

During the internet boom in the 90s, initial official productivity data also showed no response; later, the government was forced to significantly revise nominal GDP and productivity data upward. GDP simply cannot accurately measure the true value of modern technology. If nominal GDP were revised upward by 20%, the productivity data would instantly become incredibly bright. Therefore, using official GDP productivity data to deny AI is utterly nonsensical. In short, the fact is that profits and revenues of the S&P 500 companies have surged, while they have no need to hire employees.

Host: If corporate profit margins continue to rise, can investors simply buy the S&P 500 index and lie down? However, you earlier mentioned an unwillingness to be blindly optimistic about listed large companies. How do we balance this?

Jordi Visser: Because traditional publicly listed companies are currently in a serious state of workforce surplus. Compare the S1 filings submitted by Anthropic and Google. When Anthropic disclosed its S1, it had an annual revenue of $16.4 billion and a highly streamlined workforce; when Google filed its S1 in 2004, its revenue was only $1.5 billion, and now Google has nearly 200,000 employees.

New AI companies built from scratch have astonishingly high revenue per employee without the historical redundancy of code and personnel burden. In the era of AI, having too many employees presents potential cybersecurity risks.

There is a crucial market warning here: this year, the profit growth for S&P 500 companies is expected to reach 30%, but the S&P 500 index has only risen about 11% this year. On the surface, the market's high-level oscillation appears healthy, but the deeper reason is that corporate profits are skyrocketing while the P/E ratios of the S&P 500 are being compressed.

The market is extremely intelligent; it is pre-emptively warning of potential risks: next year’s focus may no longer be on the Federal Reserve's interest rate hikes or oil prices but on labor market turmoil, political gridlock, bottlenecks in memory chips and data centers, cybersecurity risks, and geopolitical conflicts. The market is applying valuation discounts to these uncertainties.

Optimistic about Eli Lilly and Micron, Micron has a 60%-80% upside potential

Host: In the face of these macro headwinds, you have long been extremely optimistic about Eli Lilly. Please share your latest investment logic on Eli Lilly.

Jordi Visser: Eli Lilly is simply a “super beast” stock. Over the past decade, Eli Lilly has achieved over 30% annualized compound returns and has seen consecutive increases for 10 years. However, in discussions among institutions and retail investors, its popularity is far lower than its performance. The core investment logic for Eli Lilly includes the following points.

Firstly, certain non-cyclical super growth. The obesity and diabetes drugs that Eli Lilly focuses on have massive rigid demand. There are 100 million obese people in the U.S., with a current penetration rate of only about 10%. As oral capsules gradually replace injections, the growth potential is enormous.

Secondly, a very high regulatory moat from the FDA. AI can quickly replicate and disrupt software code, but it is impossible to overturn physical medicine that requires years of rigorous FDA clinical approval overnight. It acts like a "public utility company" that operates monopolistically in a high-growth track.

Thirdly, immune to macro recessions. Historically speaking, whether during the 2000 internet bubble burst or the 2008 global financial crisis, the healthcare sector has always been an excellent safe haven.

By 2030, Eli Lilly will surpass all tech giants, becoming the largest stock by market cap in the S&P 500 index. Nvidia may drop to a 5x P/E due to the cyclicality and terminal value risks of AI hardware; however, Eli Lilly does not face such risk of disruption at all. I firmly believe the two major investment themes for the future are proxy computing power and longevity economics. AI brings medical breakthroughs that allow humans to live longer and healthier, and Eli Lilly is the purest super leader in the longevity economy.

Host: You strongly recommended Micron when it was still at $89 in early 2025; later it skyrocketed 12 times to $1200. You cleared out at over $700, and recently rebuilt your position in Micron at over $800. Please discuss your latest logic.

Jordi Visser: I cleared out of Micron in May/June of this year because the market was extremely crowded at that time. After the past few months of "time and price" digestion, leveraged funds and retail investors have been thoroughly washed out; the stock price has stabilized above the 200-day moving average, with a very clean lineup. The reasons for my renewed substantial position in Micron are:

Firstly, humanoid robots' insatiable thirst for memory (RAM). Musk has clearly stated that the demand for memory from humanoid robots is unimaginable, even calling it "the most scarce resource on Earth." Secondly, the 10GW computing power plan from SpaceX. Musk is promoting the realization of a 10GW computing power construction next year, which requires an immeasurable amount of high-bandwidth memory (HBM) and storage support. Furthermore, Micron and SK Hynix have locked in long-term supply agreements extending to 2030 with customers over the past two quarters. According to reports from the supply chain, such as Corning and SpaceX, there is no bubble in the market, but rather severe supply chain bottlenecks, leading to continuous increases in computing power and storage prices.

Host: As an investor adept at "predicting turning points," what advice do you have for ordinary investors?

Jordi Visser: You must step outside the current noise and seek structural major trends for the next 5 to 10 years. As legendary Wall Street investor Adam Parker said: "The real transformation of the AI industry is currently in the third quarter, while AI stocks are in the fifth quarter."

At this current stage, it is crucial to find targets with huge asymmetric risk-reward ratios. For example, Micron currently has only about a 10% downside potential, while the upside potential is 60% to 80%.

Additionally, longevity economics is a highly certain main line. I'm optimistic about assets related to travel, health management, entertainment, recreational sports, and demographic structural dividends; I am not optimistic about traditional fast food like McDonald's, especially as the popularity of weight-loss drugs like Ozempic grows, people's dietary structure is permanently healthy.

Host: You mentioned earlier that crypto assets are the largest weighted asset in your current investment portfolio; could you elaborate?

Jordi Visser: The reason is simple: among the 100 wealthy individuals we interact with, 95 do not believe in or understand how to allocate into crypto assets at all. When the vast majority of institutional funds have yet to enter the space, that constitutes a massive opportunity. My crypto asset configuration has expanded from merely holding Bitcoin to include Ethereum, privacy security networks, and a digital asset portfolio containing 46 high-quality projects.

As for my optimism regarding crypto assets, there are many reasons: first, the U.S. government cannot reduce its fiscal deficit to 3%. Faced with high interest payments and the younger generation's strong concerns about "housing/living affordability," the government will inevitably ease pressure through financial repression and liquidity injections in the future.

Additionally, with the explosive growth of AI Agents, the traditional M2 money supply will be reshaped by a seamlessly operating 24/7 financial network. RWA tokenization will bring profound changes to the financial system. I believe 2027 will be a key year for the explosion of crypto digital assets in commercial applications and payment networks.

Related reading: Raoul Pal in conversation with Wall Street strategist Jordi Visser: Why now is the best time to invest?

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[email protected]，本平台相关工作人员将会进行核查。