Li Bei, founder of Banxia Investment, believes that under the new policy on completed-home sales, the property opportunity has been upgraded from a "once-in-a-decade" event to a "once-in-two-decades" event, with the industry shakeout becoming more thorough, and leading developers' share prices potentially rising 5 to 10 times over the next few years.

At the same time, she flagged the risk of an AI bubble, arguing that the second leg down for AI will have to wait until capital expenditure peaks, which may happen in the second quarter of next year.

When the AI boom fades, Chinese property and consumption could instead become a "desert oasis" for global assets.

"The property opportunity, which I used to call once-in-a-decade, has now become once-in-two-decades," Li Bei said in a recent conversation with Tencent Finance. The new completed-home policy has dealt a significant short-term negative blow to property share prices, but the scale of the opportunity has actually increased, and the bull market in property stocks will last longer and reach greater heights.

In her view, under the impact of the new property rules: over the next one to two years, the industry will undergo a more thorough shakeout, with more companies continuing to exit; in the long run, entry barriers will rise and the gap between companies will widen. A small number of leading developers that possess both product strength and financing advantages may instead become the beneficiaries of the new rules, and some companies' share prices could rise 5 to 10 times over the coming years.

The new completed-home sales policy is the key variable in her judgment of changes in the industry landscape. For homebuyers, buying a completed home allows them to postpone down payments and monthly mortgage payments, and avoids risks such as unfinished projects and uncertain delivery quality.

On this basis, she estimates that completed homes carry a price premium of about 3% to 5% over pre-sale homes, with developers able to share part of the premium and buyers also benefiting from lower funding costs and reduced risk.

On the developer side, the market generally worries that completed-home sales will lengthen turnover cycles and squeeze profits. But her calculations show that if land payments can be made in installments, project capital turnover time may increase by only 20% to 30%, and combined with a modest increase in home prices and a decline in land prices, the net profit margin could actually rise by 20% to 30%.

Over the longer term, she expects industry concentration to rise substantially. The proportion of companies ultimately exiting could rise from 90% to 95%, or even 98%, and the number of companies able to continuously acquire land and expand nationwide may be no more than a double-digit figure.

At the same time, land payment requirements, sales fund supervision and the lead bank system have raised the threshold for new entrants, making financing capacity and major shareholders' guarantee capacity core competitive strengths.

Li Bei believes that given the gap in financing costs, the advantage of central state-owned enterprises over private firms will widen further. Even after the industry reaches a high level of prosperity in a few years, new players will not easily consider entering, and a highly oligopolistic landscape could persist for a long time.

Beyond property, Li Bei also flagged the risk of an AI bubble. She judges that although overseas AI capital expenditure is still growing, quarter-on-quarter growth may peak by the middle of next year. Earlier increases in investment by cloud vendors were based on the assumption of rapid linear growth in AI model revenue; but revenue growth at model makers has slowed markedly, and the correction in AI upstream-related companies since July mainly reflects valuation compression rather than an immediate downward revision of earnings expectations.

In her view, the "second leg down" for the AI sector may have to wait until capital expenditure truly peaks and profits decline. She judges that when the AI boom fades, the U.S. economy slows, U.S. Treasury yields fall and the dollar begins to depreciate, Chinese property and consumption may instead become a "desert oasis" for global assets.

She believes consumption has already undergone five to six years of adjustment, and the steepest phase of decline in households' marginal propensity to consume may be over. Even if the economy takes another step down next year due to declines in property investment and exports, consumption will remain relatively stable; many staples and some discretionary consumption leaders are valued at historical lows, and with dividends and buybacks, some companies' static returns are already attractive, making Chinese consumer stocks "paved with gold."

When AI capital expenditure cools and domestic demand policy is stepped up, these companies could usher in a main upward move. The following is the full text of the conversation, edited for content:

01 The property opportunity, once a once-in-a-decade event, has now become a once-in-two-decades event

Tencent Finance: How do you view the impact of the new completed-home sales policy on property?

Li Bei: Previously it might have been 90% of companies dropping out, leaving 10%; now it may be 95% dropping out, leaving only 5%. Completed-home sales will further intensify the reshuffle, and this needs to be viewed from both long-term and medium-to-short-term perspectives.

First, look at the stage the industry is in. The property industry is now at the tail end of a downturn cycle, in a phase of bottom-level differentiation. In fact, the new-home market began to improve in the second half of 2024, and quality companies' new projects can almost all achieve a 10% net margin. Compared with second-hand homes, fourth-generation new homes have higher usable floor area ratios and better design and quality, so even if they are more expensive, buyers are willing to purchase, and many projects sell out on launch day.

Second-hand homes are convenient, and the national index is still drifting lower, but the decline has narrowed significantly compared with last year. Shanghai has already begun to rise, and more than a dozen second-tier cities are also stabilizing and recovering.

At the company level, 90% of companies have already dropped out, and new-home supply has fallen sharply and is still falling. New construction starts are only about a quarter of the peak, and new project launches are about one-fifth of the peak, still down 20% to 30% year on year. But some surviving companies with strong product strength and financing capacity have in fact seen significantly improved operations over the past two years, though this is not yet visible at the financial reporting level due to lags.

This is because, on the one hand, financial statements lag sales by two years, and on the other hand, there are still legacy burdens being provisioned for impairment. What will happen after the new policy? Many people believe completed-home sales will greatly lengthen developers' turnover cycles and raise financial expense ratios, thereby sharply压低 ROE. I do not think so.

Because on the one hand, home prices can rise; on the other hand, land prices will fall, and land payments can most likely be made in installments. First, look at home prices. If you buy a pre-sale home, you need to pay the down payment now and make mortgage payments for two years, with delivery only after two years; if you buy a completed home, you only need to pay a 5% deposit now, and pay the down payment, arrange a loan and make mortgage payments two years later. For the same home, most people would choose the latter.

Therefore, even if a completed home is 4% to 5% more expensive, based on a mortgage cost of 3% per year, or 6% over two years, it is still more cost-effective than a pre-sale home, and there is no risk of unfinished projects or delivery uncertainty. For the same type of property, completed homes have room to raise prices by 3% to 5% relative to pre-sale homes, and consumers are willing to accept this.

In addition, land prices will also be pushed down. Several land auctions after the new policy show that under the same conditions, if installment payments are not allowed, land prices are commonly about 30% lower than comparable plots before the new policy; if installments are allowed, it is equivalent to local governments bearing part of the capital turnover pressure.

When the new policy was first introduced, I judged there would be two changes: first, home prices might rise; second, installment land payments would most likely appear. Both have now been confirmed: two cities have issued detailed rules on land installment payments, and dozens of projects nationwide have canceled discount offers, equivalent to an indirect price increase.

The reason the pressure from the new policy can be passed on to consumers and local governments is that developers' bargaining power has increased, both with homebuyers and with land sellers. For consumers, developers that can build good homes, avoid unfinished projects and earn recognition for quality are becoming increasingly scarce; for local governments, the number of companies able to acquire land is already limited and will become even smaller in the future.

Looking at the steady-state new financial model one year later, for a leading developer with advantages: assuming 50% of land payment is made upfront and the remainder is paid after completed-home sales, capital turnover would be extended by only 20% to 30%. With moderate leverage supported by the group, sales scale can be maintained. With home prices up about 3% and land prices down 10% to 20%, the net profit margin could actually rise by 20% to 30%. Thus, even looking only at the medium and short term, corporate profits will not fall but rise.

In the long run, there are three outcomes. First, the shakeout will be more thorough. The proportion of companies washed out may rise from 90% to 95%, or even 98%. I judge that no more than a double-digit number of companies will be able to continuously acquire land and expand nationwide in the future.

Second, the threshold for new entrants will rise sharply. Once any industry has high prosperity and high profitability, it will attract new players, but the new policy requires developers to pay land payments in a lump sum, or at least half upfront; at the same time, fund supervision and the lead bank system mean that before project completion, sales funds must remain in supervised accounts. Even if land payments can be made in installments, funds will still be tied up for a long time. The old model of acquiring land with a small amount of own funds and then rolling through mortgage loans is basically no longer viable.

Although some local detailed rules leave openings, they mainly benefit state-owned enterprises backed by large groups. For example, Wuhan's detailed rules allow a minimum 10% down payment for land, and when supervised funds exceed project needs, they can be withdrawn with a group guarantee. But basically only large state-owned enterprises, especially central SOEs, can provide such guarantees, while private firms find it very difficult.

Third, financing capacity becomes the core competitiveness. As more funds are tied up, financing capacity is crucial, and the gap between private firms and central SOEs is enormous. Even for mixed-ownership enterprises, financing costs may be nearly 100 basis points higher than those of central SOEs, leaving almost no competitive advantage. In the future, even if the industry becomes profitable again, when leading central SOEs achieve a 15% net margin, private firms may have only 5% to 6%.

Therefore, the threshold for new companies to enter the industry will be greatly raised, and a highly oligopolistic landscape may persist for a long time.

02 AI is currently falling on valuation, and the second leg down is still ahead

Tencent Finance: Since the beginning of this year, the capital market has shown a "K-shaped divergence," with technology and new productive forces sectors in stark contrast to traditional sectors. But the technology sector has recently pulled back somewhat. There is considerable debate in the market over whether AI is a bubble. How do you view this situation?

Li Bei: AI capital expenditure is indeed growing at a high rate this year and will still grow next year, but from a quarter-on-quarter perspective, it may peak around the second quarter, and even if not in the second quarter, it could be in the third or fourth quarter.

The recent upward revisions in capital expenditure by cloud vendors mainly stem from AI models working well in the coding field and Anthropic's ARR rising rapidly. After this happened in the first quarter, these cloud vendors sharply raised capital expenditure in the second quarter. The assumption at the time was linear extrapolation: based on the first-quarter trend, total model revenue by the end of this year could be around 500 billion (with Anthropic alone at 300 billion), and could exceed 1 trillion next year and the year after.

If revenue could really reach more than 1 trillion, then it would match the current annual capital expenditure of more than 1 trillion, meaning it would not be a bubble but reasonable. But in fact, ARR growth slowed significantly in the second quarter. I repeatedly warned of this risk in May and June, predicting that ARR would slow sharply, and this has indeed been confirmed.

This is precisely why the market has corrected in this round. From April to June, the market sharply raised its earnings expectations for 2027 because capital expenditure was revised up; at the same time, the valuation levels assigned based on 2027 were also higher, because everyone believed this was a sustainable level and could continue to grow.

So this decline is not a "Davis double kill"; what is falling is valuation: the market still recognizes that 2027 profits will grow substantially, but if ARR cannot rise, the investment level is unsustainable, and capital expenditure is very likely to peak in 2027, and even if not in 2027, it will peak in 2028. If end-user revenue cannot pick up, investment across the supply chain cannot be maintained, and high valuations cannot be justified.

My view is that the second leg down for AI will have to wait until capital expenditure truly peaks, possibly by the middle of next year. By then, when everyone sees profit expectations beginning to decline, AI will fall in a second wave.

03 By this time next year, property and consumption may instead become a "desert oasis"

Tencent Finance: What is the logic behind judging that there are investment opportunities in consumption and property?

Li Bei: Some people will ask, if the economy is very weak, why would consumption and property have opportunities? Because their profits will no longer decline. Consumption has in fact already experienced five to six years of decline, and consumers' mindset has now stabilized; the steepest phase of decline in the marginal propensity to consume is over.

Previously, because of the pandemic and the sharp drop in home prices, people engaged in some panic saving and their propensity to consume fell very low. After experiencing phased shocks, people's mindset has begun to stabilize. After the rapid adjustment period passed, the entire industry also re-stabilized. Staples were originally not greatly affected by economic fluctuations, and the key is that these staples and some discretionary consumer goods are now at historically low valuations.

Take two examples: a leading dairy company, as a staple consumer business, has relatively small earnings fluctuations. When consumption was sought after in 2021, it traded at 40 to 50 times PE, and now it is only 10 times. Even more striking are the leading frozen food company and the leading condiment company, both of which had triple-digit PE ratios in 2021, similar to today's technology bubble, and now both have returned to just over 10 times, at historically very low valuation levels.

Moreover, these companies also pay dividends. For example, the company that makes konjac snacks has a dividend yield of more than 7%. On average, this type of company's dividend yield plus buybacks can provide a static return of about 4% to 7%. So they are in fact extremely undervalued; it is just that the current atmosphere is all focused on AI.

Once the AI narrative collapses, or once domestic demand policy is introduced, everyone will discover that this area is paved with gold, and the trend will turn. So I think it is entirely possible now to stay in this sector and wait for the wind to come; it is already too cheap, and it is not sensitive to the economy.

Property follows another logic. Previously, developers sold pre-sale homes, which could only enter the financial statements after completion and settlement, so this year's profits reflect sales from two years ago. As mentioned earlier, new-home sales began to improve in the second half of 2024, and the homes these companies sold already had a 10-point net margin. This part corresponds exactly to the financial statements in the second half of this year; the financial statements in the first half of this year correspond to the first half of 2024, which was precisely the worst period for new-home sales.

In other words, by the second half of this year, the profits of several leading companies will improve significantly, and in the first half of next year they will rise in a step-change manner. Now some leading companies may show a development business net margin of only 1 or 2 points, or 2 or 3 points, on their statements. By this time next year, you may find it at 5 or 6 points, or 6 or 7 points, and by this time the year after next, possibly 7 or 8 points or even around 10 points.

This comes partly from the time lag between settlement and sales, and partly from the clearing of impairment burdens. Over the past two years, they have been continuously making asset impairment provisions to deal with the historical burden of previous price declines. Last year was the peak of this process; this year some companies have already begun to see less of it, and next year basically everyone will significantly reduce it.

So by this time next year, property companies' reported performance will reflect 2025 sales, and sales in the first half of 2025 were actually quite good. You will find that property and consumption may instead become a "desert oasis."

Source: Tencent Finance

Risk warning and disclaimer

Markets carry risks, and investment requires caution. This article does not constitute personal investment advice and does not take into account individual users' specific investment objectives, financial conditions or needs. Users should consider whether any opinions, views or conclusions in this article are suitable for their particular circumstances. Investment based on this article is at one's own risk.