"Bullish" Tom Lee: The Biggest Short-Term Winner of AI Productivity is the Financial Services Sector, Raising S&P 500 Target to 8000 Points

Wall Street's well-known bull Tom Lee has raised his year-end target for the S&P 500 to 8000 points, primarily due to the consensus expectation for earnings per share in 2027 rising to $400, making valuations cheaper than at the beginning of the year. He is optimistic about the financial, healthcare, and technology sectors as the biggest beneficiaries of AI productivity gains, and expects the S&P to experience a sharp pullback in the fall due to new Federal Reserve communication mechanisms and the unlocking of IPOs like SpaceX. However, as long as the economy does not deteriorate significantly, it will still show a "V-shaped" rebound.

Written by: Zhang Yaqi, Wall Street Watch

Wall Street's well-known bull and co-founder of Fundstrat Global Advisors, Tom Lee, has raised his year-end target for the S&P 500 to 8000 points, identifying the financial services sector as the biggest short-term beneficiary of AI productivity gains. He believes that earnings growth will continue to dominate the market, but investors must be wary of multiple risks in the second half of the year.

In a recent podcast, Lee stated that as we enter the second half of 2026, his core logic for maintaining an optimistic stance is based on the continued upward revision of earnings estimates: the market's consensus expectation for S&P earnings per share in 2027 has risen from $350 at the beginning of the year to $400, while the corresponding forward P/E ratio has actually decreased from 19.4 times to 18.4 times. This means that despite the index rising nearly 9% this year, the market is actually "cheaper" in terms of valuation than it was at the beginning of the year. He bases his 8000-point target on a 2027 earnings per share of $400 multiplied by a 20 times P/E ratio.

However, Lee also warns that the S&P index may experience a sharp pullback similar to a bear market in the fall, triggered by factors such as changes in the leadership of the Federal Reserve, supply pressures from large IPOs being unlocked, disruptions in shipping through the Strait of Hormuz, and historically high levels of margin leverage. He maintains his baseline judgment of a "V-shaped rebound," believing that any pullback will quickly reverse as long as the economic fundamentals do not show significant deterioration.

Earnings Driven Remain Solid, but Quality is Questionable

Lee attributes the potential for a "fourth consecutive year of double-digit growth" in 2026 to three overlapping lines of earnings growth: AI and energy infrastructure construction, the return of manufacturing, and the continued effects of government infrastructure spending. He points out that 76% of large-cap growth fund managers have underperformed their benchmarks this year, mainly because they missed the rally in the semiconductor and DRAM sectors, and this group of laggards will provide additional upward momentum in the second half of the year.

However, he also admits that the issue of earnings quality cannot be ignored. Large tech companies are accounting for valuation increases from private equity investments in their earnings, there is a "bullwhip effect" driven by price increases in the semiconductor supply chain, and ultra-large cloud service providers are maintaining capital expenditures through public market offerings and debt financing—these factors all suggest that applying a higher valuation discount to current earnings growth is reasonable. He notes that credit spreads are currently at abnormally narrow levels, which is his primary reference indicator for observing any cracks in the stock market.

AI Productivity Gains: Financial Services Sector is the Biggest Short-Term Winner

Lee believes that AI has proven two things in practice: it makes capable people more efficient and it generates real output from previously "latent idle" work hours. In knowledge-intensive industries, for example, the actual productive time in a 40-hour work week may not exceed 6 hours, and AI is filling these gaps and converting them into actual capacity.

Based on this logic, Lee identifies the financial services sector as the most certain beneficiary of AI productivity gains in the near term, along with the healthcare and technology sectors. He is also optimistic about AI downstream software stocks (IGV) and the seven tech giants (Mag 7)—the rationale being that the valuation downgrades these stocks have experienced have made their risk-return ratios reasonable, and their compound growth potential as beneficiaries of AI has not yet been fully priced in by the market.

Looking further ahead, he believes that industrial robots will first transform warehousing logistics and then penetrate the residential construction industry—leading to a structural leap in semiconductor demand, as the chip usage density of a single robot is about 50 times that of an iPhone. Amazon, with its scale of nearly a million industrial robots, will be one of the core beneficiaries of this trend, in his view.

OpenAI Delays IPO: A Narrative Crack or Strategic Consideration?

Lee finds it "thought-provoking" that OpenAI and Anthropic have both delayed their IPOs, but does not believe this is directly related to SpaceX's successful listing. He speculates that the U.S. government's tightening of scrutiny on new models may be one reason, but admits he does not have inside information.

He believes that if the two companies choose to go public, the issuance itself would not be difficult—private markets have been quite smooth in meeting their financing needs. Investors assessing OpenAI and Anthropic will not simply apply a subscription business valuation framework, but will need to see a path to realizing free cash flow. He compares this to Meta's mobile transformation and SpaceX's creation of satellite spectrum as core assets, suggesting that these two companies have a non-zero probability of achieving "home run" level success—but this essentially bets on the personal judgment and execution ability of the founders, which carries a high degree of key person risk.

Four Major Tests and V-Shaped Rebound Logic in the Second Half

Lee outlines four main risks facing the market in the second half of the year: first, the new Federal Reserve Chairman Waller plans to restructure the monetary policy communication mechanism, including canceling fixed-frequency press conferences and forward guidance, and the market will need to look for alternative policy signals; second, the lock-up periods for large IPOs like SpaceX will gradually expire in the fall, significantly increasing the supply of circulating market capitalization; third, shipping through the Strait of Hormuz has not yet returned to normal, and the global supply gap for petroleum products continues to widen; fourth, margin leverage has increased by 55% year-on-year, which is the fifth highest level in nearly 70 years of history, and is highly correlated with future market pullbacks.

Regarding the judgment that these risks will ultimately lead to a "V-shaped rebound" rather than a sustained decline, Lee's basis is that as long as corporate credit spreads and yield curves do not signal significant economic deterioration, the market has a foundation for rapid recovery. He points out that the market pullback earlier this year related to the conflict with Iran has validated this framework.

Semiconductors: Cyclical Stocks or Structural New Story?

Lee holds a relatively optimistic view on whether semiconductors can maintain their strength. He points out that in every semiconductor cycle over the past 50 years, the total market size itself has not experienced a leap in expansion, but this time the large-scale application of robots will bring a whole new level of demand—each autonomous robot's chip demand is about 50 times that of smartphones. The special performance requirements of space application scenarios will also create differentiated demand for semiconductors. He believes that at least within the visible cycle from 2026 to 2027, the impact of the "big cycle" and "structural transformation" narratives on market direction will not differ much, and visibility has become quite clear recently.

The following is the full interview:

Host: At the end of last year, a guest made a very bullish prediction for 2026 on the show. Fast forward to now, the U.S. stock market has indeed performed well, with an increase of nearly 9% this year, but there are still many unresolved issues in the market. Now that the first half of 2026 has come to a close, we want to invite this guest back to discuss: does he still remain bullish? What are investors currently underestimating? What direction will the market take in the second half?

Today, we welcome Tom Lee, co-founder, managing director, and head of research at Fundstrat Global Advisors. Tom, it's great to have you back on the show.

Review of the First Half of 2026: Leading and Lagging Sectors

Host: We would like to first ask you to review the market performance in the first half of the year, and then talk about your outlook for the second half. The S&P index rose nearly 9% in the first half, the Dow Jones rose 8%, and the Nasdaq rose 11%. The differentiation between sectors is also quite evident—this year, the "seven tech giants" (Mag 7) have performed relatively under pressure, and we will also discuss the situation of cryptocurrencies later. Please share your overall feelings about the first half of the year.

Tom Lee: 2026 is expected to be the fourth consecutive year of double-digit growth. This may surprise some viewers, but historical experience shows that when the market has strong growth for three consecutive years—like we saw in 2023, 2024, and 2025—the fourth year often remains robust. This is one of the reasons we maintained an optimistic outlook at the beginning of the year.

The initial judgment was that earnings growth would become the main driver of the market, and this has indeed been the case. At the beginning of the year, the market's consensus expectation for S&P earnings in 2027—close to our prediction—was $350, and it has now been raised to $400, an increase of $50. The P/E ratio corresponding to 2027 earnings was 19.4 times at the beginning of the year and has now dropped to 18.4 times.

This may surprise everyone: although the index has risen by 9%, the stock market is actually "cheaper" than it was in January. I believe there are ample reasons to remain optimistic now, as I truly believe that U.S. corporate earnings still have room for further upward revision. The core factors driving earnings growth remain in place—partly due to AI and energy infrastructure construction, partly due to the trend of onshoring manufacturing, and the government’s infrastructure spending also has some continued effects. These are all tailwinds supporting spending growth, and overall, investor sentiment has not yet become overly optimistic.

However, now that we are at mid-year, there are two factors worth noting. First, margin debt levels are much higher than at the beginning of the year, with a year-on-year increase of 55%, which is the fifth highest year-on-year increase in nearly 70 years of history. Historical experience shows that this is often related to the "ammunition running out" of those trading on borrowed money.

On the other hand, from this year's fund managers' performance, 76% of large-cap growth fund managers have underperformed their benchmarks, which is quite rare; the proportion for large-cap balanced funds is 60%, which, while not as extreme, is still significant. This means that growth fund managers have largely missed the rally in the semiconductor and DRAM sectors this year, and I believe they will chase this rally in the second half of the year, which is also one of the reasons I tend to remain bullish.

Summary:

Host: I have quite a few concerns about the current market. For instance, looking at the Shiller PE ratio, it is now at an extremely high level, basically approaching the range seen during the internet bubble. I am gradually feeling that the distinction between forward earnings and trailing earnings is becoming very important—much of the earnings we see now are based on contracts signed with these AI companies, and whether these AI companies will actually be able to fulfill these contracts in the future is something I believe is entirely questionable.

For example, the spending plans of OpenAI and Anthropic, as well as SpaceX, etc. Meanwhile, we also note a recent study confirmed by Goldman Sachs: a significant portion of the earnings reported by many large tech companies actually reflects the increase in valuations of their private equity investments in AI companies, and due to relevant accounting standards, this valuation growth will be included in the actual earnings figures.

What I want to say is that while the earnings growth numbers look strong, I actually have some contradictory feelings about it, especially when they say "earnings will reach such and such in the next 12 months." I would love to hear your thoughts on my skeptical attitude.

Tom Lee: I basically agree with your view because what you are raising is actually a question of earnings quality. I think there are several factors that indeed warrant skepticism about earnings quality. First, the book gains from investment valuation increases on the balance sheet are not the same as true operational earnings.

The second difference is that there is currently a certain degree of "price increase" factor—some companies in the chip supply chain, due to the inability to expand wafer fab capacity or the long expansion cycle, are turning to price increases to achieve growth, which means more earnings will directly reflect on the profit side. However, this can cause a "bullwhip effect" because we know that the supply chain will eventually catch up with demand, and prices will fall back.

At the same time, we also know that hyperscalers are writing huge checks for construction, and now they are starting to raise funds from the stock market to support these expenditures—this is why Google has established an ATM issuance mechanism, Meta may also take similar actions, and of course, there is SpaceX's IPO. These are all efforts to raise funds from the public market.

The last point is that there is currently a phenomenon of credit consumption—meaning that companies are utilizing another part of their capital structure (debt) to finance these expenditures. So I believe that for these reasons, we should indeed demand a higher standard for the valuation multiples applicable to earnings growth.

That said, I think these activities are currently concentrated in four countries or regions—so we must seriously consider what this means. These four countries or regions are, of course, the United States and China, and I group South Korea and Taiwan together as partners in AI infrastructure, possibly adding Japan. So, while we can maintain a certain degree of skepticism, the story that is unfolding is that AI is currently clearly benefiting only a few countries or regions.

Is the Market Overheated? Debate on the Shiller PE Ratio

Host: Speaking of the Shiller PE ratio, I think the real question for us right now is: is the market overheated? Interestingly, there seems to be no consensus on this issue—everyone is looking at different indicators, trying to determine whether the market is truly in a bubble, or just "overheated," or neither. What do you think about this? I believe earnings quality should also play a role in this discussion.

Tom Lee: I did such analysis when I was at JPMorgan, and now at Fundstrat we continue this practice—you can calculate the Shiller PE ratio by industry, which allows for a more "apples to apples" comparison. For example, if you calculate the Shiller PE ratio for tech stocks alone, you can compare it to 1929. Of course, the so-called "tech companies" of that era are completely different from now—back then, tech companies were radio manufacturers, and later included microwave manufacturers, etc.

But measured this way, the current Shiller PE ratio is not that extreme because the proportion of tech in the earnings composition is continuously increasing. The tech sector—I don’t have precise numbers at hand, but it probably accounts for about 60% to 70% of total earnings growth, while its share of total earnings is only about 40%. This was not the case in 1999—back then, tech stocks did not account for a high proportion of total earnings, but they contributed significantly to the valuation multiples.

So I think from the perspective of earnings composition, this time is different. The ISM data also reflects the same trend—if you look at manufacturing and services separately, over the past 50 years, we have shifted from "manufacturing being the mainstay of economic activity" to manufacturing only accounting for about 30%. So I believe the Shiller PE ratio actually has room to rise further.

However, I want to add a fair question: is the credit spread underestimating risk? Because the yield on 10-year Treasury bonds has been rising, I am actually surprised that the spread between high-yield bonds and investment-grade bonds is maintaining such a tight level. Logically, considering geopolitical risks and the capital cost pressures brought by high interest rates, the spread should widen, but in fact, it hasn’t. If you ask me, before observing whether the stock market will show cracks, I would first focus on the credit market—but I believe the signals currently being sent by the credit market indicate that liquidity in the market is still excessive.

Second-Order Effects of SpaceX and Mega IPOs

Host: Regarding the impacts or second-order effects of these large IPOs, there seem to be contradictory views in the market—one viewpoint suggests that these mega IPOs will siphon off a lot of IPO funds from the market; another viewpoint suggests that this will actually have a positive effect on the IPO market, essentially declaring that "the IPO window has reopened." What do you think? What do you believe the second-order effects of these large IPOs will be?

Tom Lee: I think the groups affected by IPOs can be roughly divided into three categories: first, the issuers themselves; second, the private shareholders of these companies before they go public; and third, the broader overall market. SpaceX is a great example—it has an IPO scale of only $75 billion, while the overall valuation of the company exceeds $1.5 trillion, so the current float is only about $90 billion.

From a market cap perspective, SpaceX, as a stock, actually has a float market cap smaller than most components in the Nasdaq 100 index, ranking roughly in the top 50 of the total market cap of Nasdaq, which is also why it performs well in trading. However, these shares will be released in phases, and by the end of this year, the total market cap of the released float should exceed $1 trillion. I believe this represents a significant supply increase for the public market and the overall market. I expect the supply pressure effect from SpaceX will gradually become apparent by the end of this year.

But before that, this has also created a huge wealth effect—because SpaceX has only raised $18 billion throughout its development, and now its valuation has reached $1.5 trillion. So those shareholders who held shares during SpaceX's privatization phase have created enormous wealth, which will actually bring about a real economic stimulus effect, as banks will be willing to lend against the shares they hold as collateral, so I believe this will actually boost GDP.

So you are right, Scott, there is indeed a balancing force here. My judgment is that the overall economy will benefit from these IPOs because they create enormous paper wealth and locked assets; the stock market will perform well before the lock-up expires because once the lock-up begins, the market will need to digest this new supply; and the issuers themselves will perform very well because they now have a channel to raise funds from the public market, which can accelerate their expenditures. So I believe the issuers will fully benefit from their IPOs.

Cracks in the AI Narrative? Delays from OpenAI and Anthropic

Host: Let’s continue discussing capital expenditures and the second-order effects of these AI companies—these companies have made extremely large commitments in capital expenditures, and I believe this is one of the reasons many related stocks have been pushed up. I want to propose a hypothesis: it seems that OpenAI may be putting its IPO plans on hold, and I can’t help but think this means that market momentum has shifted, or growth expectations have not met the anticipated levels.

Are you worried that we are starting to see—I don’t want to use the term "AI bubble"—but at least some cracks in the AI narrative, and that these cracks will affect the entire market? Or do you not worry too much about this issue?

Tom Lee: The AI narrative indeed has many potential pitfalls, Scott. First, as you know, we need to build a massive amount of power and infrastructure to support all of this, and building all of this is itself a huge project. To be honest, we are not even clear what impact this will have on residential electricity prices, what environmental damage it might bring, or what the quality of life will be like if you happen to live near a hyperscale data center.

I think we are not entirely clear why both Anthropic and OpenAI have postponed their respective IPO plans, which I find quite intriguing because whichever side goes public first can benefit from being "first to market"; no one wants to be the "last to go public."

I suspect this may be related to the U.S. government tightening approvals for new models—you know, for example, the Mythos model underwent extensive scrutiny, and the Fable model was once taken down. Is it possible that the U.S. government is saying, "We now need to review all your models"? And perhaps as a result, they are being asked to tighten their plans? I’m not sure, but to me, this is indeed an intriguing phenomenon. However, I believe SpaceX is a huge success story, so I don’t think there is a direct relationship between SpaceX's success and the postponement of IPOs by OpenAI and Anthropic.

Why Hasn't OpenAI Gone Public Yet?

Host: I want to dig deeper into this question—why hasn’t OpenAI gone public yet? What are they worried about, what are they anxious about? I would love to hear your thoughts.

From a financial perspective, OpenAI’s current situation doesn’t seem too good. Frankly, from a profitability standpoint, they lost nearly $40 billion last year; if we look at operating profit, they probably lost about $21 billion last year. Currently, our understanding of the AI industry is that revenue growth is very rapid, usage is also continuously increasing, but the costs of running and training these models are extremely high, and we have not yet seen the AI business truly achieve profitability, and it is still in some sort of "experimental phase."

So I am very curious, is it because ------ we have recently seen their financial data leaked, and the market reaction has not been very good? My personal intuition is that these financial data are indeed unexpectedly poor from a profitability perspective. I wonder if investors might not be able to bear such financial conditions, so they feel the need to sort out the business model before going public? I would love to hear your thoughts on whether this factor has played a role, as well as your views on these companies ------ these companies carry such high expectations, but their business models seem not to have really taken off yet.

Tom Lee: Let me first clarify that I do not have all the insider information; I am just sharing my personal views. I believe that if OpenAI or Anthropic goes public, their IPOs would be very successful because their stories are relatively easy to understand ------ they are not the kind of conglomerates with complex businesses. OpenAI and Anthropic are clearly at the forefront of building complex reasoning models, which will ultimately become our agents. Currently, the public cannot invest at all, and institutional investors are also basically unable to participate.

I think they have no problem raising funds in the private market ------ in fact, they have been very successful in raising tens of billions or even hundreds of billions of dollars. So this might be one of the reasons for their delay in going public, but to me, it remains a thought-provoking phenomenon. However, I believe that when investors look at OpenAI and Anthropic, they do not see them as a subscription business; they need to see free cash flow and the companies' willingness to continuously invest and recruit talent to maintain their leading positions ------ because both companies are very unique. Of course, I am not an insider, so I do not know the real situation.

Does Tom really believe in the AI deal?

Host: As an investor, do you personally really believe in this story? This is exactly what I want to clarify ------ because I feel that my thoughts are similar to yours: AI is undoubtedly an extremely important historical moment for the market, and these two companies are the leaders in this field. If the opportunity is right in front of us, why not bet on it? But I do believe that the issue of the business model remains a huge unresolved question. I would love to hear your personal views on this.

Tom Lee: This is still a story being written in the "future tense"; we are at most in the "first chapter" and can only make some speculations through analogy.

For example ------ although it may be a bit off-topic, to me, SpaceX's greatest achievement is that they turned satellite spectrum resources into reality ------ SpaceX (and Starlink) completely relies on satellite spectrum operations, and this spectrum resource was almost worthless compared to terrestrial spectrum before they entered the market, yet they turned it into the most valuable spectrum resource in the world. Musk acquired global satellite spectrum at almost "zero cost," while in today's America, if you want to obtain 20 megahertz of cellular spectrum, it costs about $200 per person. This is a world of difference ------ he almost got the satellite spectrum for free, and now it has become the most valuable spectrum resource in the world.

Meta is a similar example: they initially had a free, user-generated content-based business ------ essentially just an "electronic yearbook," but they turned it into one of the largest monetization businesses in history. I remember when Meta first launched its mobile business, many people were not optimistic, thinking that Facebook would have no advantage on mobile because the desktop version already had such a rich content experience. But it turned out that mobile was the key to their "hijacking" the entire advertising market.

So what about OpenAI and Anthropic? With their ability to build complex reasoning models, who knows which industry they will ultimately disrupt? Maybe we think they are just doing advertising business; perhaps we are oversimplifying their true business model. Will they build the future biotechnology laboratories? Will they create the future workforce? I believe this is still a story being written, but if we refer to the achievements of the most valuable companies like Meta, or how Google elevated the search business to a completely different level, or how SpaceX turned satellite spectrum into a core asset ------ I think this is why I feel that these companies' IPOs have a non-zero probability of becoming truly "home run" level success stories.

Host: I completely agree and am glad you mentioned this. But are you also concerned about the risk ------ what if companies like OpenAI do not succeed and do not hit this "home run"? What if something goes wrong, like Sam Altman makes a wrong strategic decision that leads the company to collapse in some way?

In my personal view, this non-zero probability risk should actually be priced in ------ not just reflected in OpenAI's own valuation, but also in the overall market valuation, because the entire market has become highly dependent on OpenAI's success and whether OpenAI can continue to invest heavily in computing power and chips.

Tom Lee: There is indeed an interesting paradox here ------ the more successful OpenAI is, the more it indicates two things: first, Sam Altman himself becomes increasingly important because, after all, it is him as a "person" making various decisions. I believe he would not actually type in ChatGPT asking, "What should I do next" ------ although he might actually do that ------ but his strategic "self-actualization" highly depends on an extremely talented team. So this is largely a "human story"; to build two such remarkable companies relies on people.

You are right; this is essentially betting on Sam himself and his vision, while also betting on the Anthropic team and their vision ------ this is actually a "two-horse race," and both companies could potentially succeed in different directions.

Host: It is this kind of "betting" that makes me worried about the current market ------ too much is riding on his personal capabilities, on whether he can truly realize the AI story. If it ultimately does not materialize, in my view, the nearly 9% increase in the S&P 500 this year, most of the growth did not come from large tech stocks themselves, but from semiconductor and DRAM memory companies ------ that is, all those chip-related stocks that are "selling shovels" for the AI boom.

If this story ultimately does not materialize, if things do not develop in the direction everyone expects, the market could very likely experience a very sharp and severe downturn. Of course, none of this has happened yet; the story is still unfolding, but I guess this should also be on your radar.

Tom Lee: I think one thing is worth noting ------ to some extent, we are discussing from a narrative perspective. Because even if OpenAI and Anthropic are indeed crucial, if they really encounter significant setbacks, the overall earnings level of the S&P would not deviate too much from the expected $400.

But at the same time, we must also realize that since the 1940s, there has always been a person in every historical period ------ whose every move is crucial to the market, and that is the Chairman of the Federal Reserve. This is a single individual, which itself is a huge "key person risk." Fortunately, we have never really encountered a situation where the Federal Reserve Chairman suddenly cannot fulfill their duties ------ I think this is somewhat a "miracle" of capitalism, because it can be said that Kevin Warsh (the current Federal Reserve Chairman) is now one of the most important figures in the world.

Host: I agree with your view ------ the market does have vulnerabilities. The S&P index is now around 7300 points; I remember when the market bottomed in 2009, it was around 600 points. So it has indeed come a long way, but fortunately, today's valuation multiples are actually lower than they were in 2009.

Tom Lee: You are right, but one very different point is that the growth rate of the U.S. economy is suddenly accelerating, which is the root of all the phenomena you just described ------ because all of this is basically driven by AI. This is also why we can feel a certain comfort while also needing to remain highly vigilant: on one hand, this prosperity is mainly concentrated in a few countries, which is somewhat reassuring; but on the other hand, as you said, it has also created a strong "path dependency" on a few individuals.

Host: This has almost become a question: in the face of this vulnerability, what exactly should you do? Should you choose to "stay out and watch" because it is fragile, or should you also realize, as you mentioned, that there are indeed "home run" level opportunities, and are you willing to miss out? This is probably the core issue. Next, let's have Scott talk about your predictions for the second half of the year.

AI Productivity Boom: Who Will Benefit

Host: Tom, you previously mentioned that AI is creating a productivity boom, and we have indeed seen productivity improvements, not just a technological boom. Which industries do you think will benefit the most from this wave of productivity enhancement?

Tom Lee: This is actually a rather difficult question to quantify precisely. I believe AI has proven a few things ------ although I can only provide some anecdotal examples, the first point is that it has indeed made those who are already capable even more efficient ------ because I have witnessed this firsthand at Fundstrat Capital and within Fundstrat, where we essentially deployed an "army of researchers," but this army is actually our Claude agent.

At the same time, it has also revealed the nature of work. Most people would say, "This is a 40-hour work week," and from an economic statistics perspective, we would say, "You worked 36 hours," or "The weekly working hours are 40 hours," because that is how the clock-in records are. But we all know that during those 40 hours, the actual time spent working efficiently may only account for a portion of that ------ could it be that in a week of 40 hours, there are only 6 hours of actual productive time? Very likely, the rest of the time people might just be browsing information, having lunch, or doing other things.

What AI does is actually fill in those originally "blank" times, creating more actual output ------ I think this is the manifestation of productivity improvement. I believe many white-collar jobs, even in the healthcare, financial services, and technology industries, fit this characteristic ------ AI is making all these practitioners more efficient, even if they theoretically are still working within that original "effective working hours ratio," but now they can accomplish far more than before.

In a few years, we should soon see robots equipped with high skills and flexibility. Many people think this will only appear in warehouses and factories, and I believe that is correct. However, I also think that in a few years, the residential construction industry will be completely transformed. For example, future homes may be built by robots with exquisite craftsmanship. We could carve your house just like we carved the Louvre out of stone, and the price would be similar to your current house. We could completely replicate the stunning architecture of Europe—something that is impossible to achieve in today’s America, but with the help of robotic technology, it could become a reality. Therefore, I believe that once robots possess sufficient capabilities, it will lead to a productivity leap of this magnitude.

Robots, Residential Construction, and Amazon's Logic

Host: I want to delve deeper into this topic because every year we try to pick a large tech company that we believe will outperform the market. While this carries some risk, the company I am most interested in this year is Amazon, precisely because of the point you just mentioned: our core logic is that AI can truly create shareholder value that matches market hype in areas such as autonomous driving, especially Waymo; the second is industrial robots, and Amazon currently has about a million industrial robots, while all other companies in the U.S. combined have about 400,000. Do you think Amazon will benefit from this wave of the "robot era" you envision?

Tom Lee: Absolutely. Scott, Amazon—you may know this company better than I do—is essentially a logistics company. They have warehouses everywhere and a vast resource of merchants. Why wouldn’t Amazon participate in the future residential construction sector? Just like Sears Roebuck in the past, they have the capability to "deliver" houses and then let their robots act as carpenters and construction agents to assemble entire homes. Once that happens, their potential market size (TAM) will double instantly, as it will cover the entire residential and commercial property market. As a logistics company, any area involving logistics is essentially their potential market.

Host: I find this logic quite reasonable.

S&P 8000 Points and Four Major Tests Before Year-End

Host: Let’s continue discussing your predictions for the second half of the year. At the beginning of the year, you set a target price for the S&P at 7700 points, and the market trend has generally aligned with that expectation. However, you adjusted your forecast in your mid-year report, now expecting the S&P index to rise to 8000 points by year-end. But there is an interesting premise—you believe the S&P index will experience a bear market-like adjustment in the fall and then quickly rebound. Can you elaborate on why you make this judgment?

Tom Lee: Yes, we did raise our target price to 8000 points—this is basically based on a projected earnings per share of $400 in 2027, calculated at a 20 times price-to-earnings ratio. However, from June to December, I believe the market will face several challenges.

The first and most obvious challenge is that the Federal Reserve is about to welcome a new chairman, Kevin Warsh, who has some quite ambitious goals—he wants to fundamentally readjust the way the Fed operates. He has set up five special working groups, one of which is to redefine the measurement of "inflation," the second is to analyze the Fed's communication mechanisms, the third is about data collection methods, and there are several other directions.

In my view, these are new challenges that the market needs to readjust to, as the market has become accustomed to a certain fixed communication rhythm—such as holding press conferences after the Federal Open Market Committee (FOMC) interest rate decisions. However, before Powell, holding press conferences after interest rate decisions was not the norm. It is said that Kevin Warsh may only hold press conferences when he feels he has something to say, and he does not intend to provide forward guidance anymore, so the market will have to look for other alternative indicators to gauge policy direction—perhaps ultimately relying on prediction markets. Moreover, he may also want to redefine the measurement standards for inflation, but that will be tricky because he himself wants to pull inflation back to the 2% target—but the question is, what statistical basis is "2%" measured by?

The second challenge will be the unlocking of all this IPO liquidity—regarding SpaceX, the unlocking will really start in the fall. The third challenge is the conflict with Iran, which is causing a persistent and expanding gap in the supply of oil products, as shipping through the Strait of Hormuz has not yet returned to normal. Even if the domestic gasoline supply in the U.S. is sufficient, it does not mean that other oil products like lubricants are also abundantly supplied in other parts of the world—put simply, oil is the "lubricant" that keeps the entire economic machine running smoothly, and whether any part of it will encounter problems is highly uncertain in my view.

The fourth challenge is the level of margin leverage—historically, this is usually associated with some form of market correction within the next six months. Therefore, I believe that from now until the mid-term elections, there may be a fifth risk factor, or what we previously discussed as "vulnerability," which may further exacerbate the market correction. And as you know, once a correction truly begins, the market often reacts very violently. So I am not entirely sure what will happen next, nor do I want to label the current position as a "top." We have consistently advised clients to maintain their positions and continue holding, because I believe there is already enough skepticism in the market, but even so, regardless of where we peak, I believe the magnitude of the correction will be quite considerable.

Why This Adjustment Will Present a "V-Shaped" Reversal

Host: Why do you believe the market will rebound so quickly? Because the challenges you just mentioned are precisely the questions I have been pondering—markets seem to naturally go through these tests, and of course, that has never been the mindset for viewing the market or investing. But these challenges do seem very important, and if a correction does occur, why do you think the market will rebound so swiftly?

Tom Lee: Whenever there is a sharp correction in the U.S. market, our consistent stance is that these corrections often present a "V-shaped" reversal—and this judgment has always been accompanied by a lot of skepticism. Take earlier this year, for example; we said that the market correction related to that conflict would be a "V-shaped" reversal.

Host: You did say that.

Tom Lee: Yes, many people did not believe it because they pointed to oil prices and various uncertainties. But I realize that the market often digests negative shocks in advance, which is why I believe that even if a very sharp correction may occur, as long as the economic fundamentals do not truly deteriorate and do not fall into a substantive negative cycle—yield curves and corporate credit spreads will give us the answers—as long as the economy does not truly collapse, whatever kind of correction occurs will ultimately present a "V-shaped" reversal. I know this sounds a bit mechanical, and of course, this judgment will ultimately be tested by the market, but this remains my default judgment at the moment.

Host: Moreover, judging from the speed of recent rebounds, the market indeed seems to be increasingly inclined towards a "V-shaped" reversal. That conflict is a great example—you mentioned at the beginning of the year that we would see a bear market-like correction, and then the market would rebound sharply, which is exactly what we have seen, just presented in the unexpected form of "conflict with Iran."

Looking at the sectors you are currently most optimistic about, you mentioned energy, small-cap stocks, finance, and industrials—these I agree with. Additionally, you also mentioned the tech sector, specifically naming the "Magnificent Seven" (Mag 7) and the software stock index IGV, which includes those software stocks that have experienced the "SaaS apocalypse" and "SaaS apocalypse 2.0". I agree with this, but why do you have a long-term bullish outlook on Mag 7 and IGV?

Tom Lee: I believe they are all beneficiaries of the AI wave—just as Scott mentioned that Amazon is a huge beneficiary of AI, I also believe that the financial services industry will be a huge beneficiary of AI. Therefore, I think investors today are more inclined to buy into those "bottleneck segments," as the growth in these segments is evident. However, over time, those companies downstream of AI will continue to reap the benefits of compound growth—because these software companies are not a monolithic entity; they each have boards, CEOs, sales teams, and engineers, and they are also witnessing all the changes we are seeing and have many ways to benefit from AI.

So for us, it is precisely the "de-rating" that these stocks have experienced that makes us feel that the current risk-reward ratio is quite attractive.

Meta and Microsoft: Performance After Sell-Off

Host: There are two massive companies with strong profitability that have seen their stock prices severely impacted this year. Microsoft’s stock price has dropped 24%, and its price-to-earnings ratio has fallen to 22 times; Meta has dropped 15%, and its price-to-earnings ratio has fallen to 20 times. What is your view on these two companies?

Tom Lee: I have a lot of confidence in both companies. I believe their ability to respond to future situations will be far more adept than the market currently expects. Both companies have a long track record of demonstrating that they understand and can respond to critical strategic turning points in business development that are vital for survival.

So, while it is easy to simply say now that "Meta has problems because they have invested too much in super-scale computing infrastructure and are no longer that cash-rich company," I believe this is still an organization with tremendous talent, and Zuckerberg has repeatedly proven his ability to make very precise strategic shifts. Therefore, I am confident that they will handle the current situation well, even if the current financial data seems to be in some transitional phase.

Microsoft is in a similar situation—Microsoft has made some very wise investments in AI, and because of this, they have achieved a significant scale advantage in their cloud computing business, which itself speaks volumes about their foresight. So in my view, both companies have quite good "dashboards" and "crystal balls," and I am confident they will handle the current situation well in the coming years.

Host: I agree with that. Now let’s take a look at the best-performing sector this year—it’s basically semiconductors. A set of data provided by Apollo’s Torsten Slok is quite interesting: semiconductor stocks now account for 19% of the total market capitalization of the S&P index, meaning that nearly one-fifth of the entire market consists of semiconductor stocks, while this proportion was less than 10% in 2025.

Semiconductors Now Account for One-Fifth of the S&P Index

Host: I would love to hear your thoughts on how the semiconductor story will unfold next, as it is indeed the core driving force of the market right now. We can even look at small-cap stocks, such as the Russell 2000 Index—much of the gains actually come from these niche AI-related stocks, rather than the traditional value stocks that small-cap stocks are typically thought to represent. It can be said that the entire market is currently revolving around semiconductors. How do you think this trend will play out? Do you think there is still room for growth? Why has this round of increases been so intense?

Tom Lee: Historically, semiconductors have always been a very typical cyclical sector. In the past, people usually traded these stocks based on the "book-to-bill ratio." However, in the past few cycles, this cyclical logic has not been as applicable. The question you raised is actually at the core: is this a long cycle, or has something fundamentally changed?

Host: Exactly, that is the question.

Tom Lee: At least for now, regarding 2026, or even possibly 2027, both explanations are not that important because the visibility in the short term is already very clear.

I suspect that semiconductors may be entering a brand new cycle, a brand new story, because—if we look back over the past 50 years, the total addressable market (TAM) for each semiconductor cycle has not fundamentally changed, and the buyer demographics are basically the same. There may have been capital inflows, leading to a "bullwhip effect" that influences everything from semiconductor equipment to semiconductors themselves.

But this time is different. We are now facing robots—robots have an extremely high demand density for semiconductors, far exceeding that of an iPhone. The number of semiconductors you need to invest in a robot—I don’t know the exact number, but I guess it’s about 50 times that of an iPhone. So once you start deploying each new autonomous robot (which can save you labor costs), the entire industry chain behind it is highly reliant on semiconductors. Of course, in the future, if semiconductors are to be sent into space, they will also need to meet very unique extreme environmental adaptation conditions. Therefore, I believe that this time, semiconductors are likely telling a completely new story.