Wire flash
Li Bei: Real estate opportunity upgraded to once-in-20-years, AI bubble risk persists
Editorial responsibility
- No named human review is recorded for this page.
- Source reporting is collected, normalized, translated or condensed automatically when needed.
- Automatically published source-backed update
In an interview with Tencent Finance, Li Bei, founder of半夏投资 (Banxia Investment), provided a detailed analysis of China's real estate sector and global AI risks. She argued that the new policy promoting presale of completed homes (现房销售) will lead to a more thorough industry shakeout, with up to 98% of developers exiting. This, she claims, upgrades the real estate opportunity from 'once in a decade' to 'once in 20 years,' benefiting a handful of top state-owned developers with strong financing and product capabilities. She forecasts that some of these firms' stock prices could rise 5-10 times over several years. Conversely, Li Bei warned that AI capital expenditure growth is likely to peak in mid-2025, as model revenue growth has already slowed significantly. She described the recent AI stock correction as a valuation adjustment, predicting a 'second wave' of declines when capital spending actually peaks and profits fall. She concluded that as the AI 'fever' subsides and the US economy slows, Chinese consumer and real estate stocks, currently at historically low valuations, could become a 'desert oasis' for global capital.
Source report
Author: Chen Zhaoyu | Editor: Liu Peng
Real Estate Opportunity Upgraded from "Once a Decade" to "Once in Two Decades"
Banxia Investment founder Li Bei recently stated in a dialogue with Tencent Finance that while the new policy on presale of completed homes has caused a significant short-term negative impact on real estate stock prices, the level of opportunity in the sector has actually increased. She believes the subsequent bull market in real estate stocks will be prolonged, and valuations can be viewed more optimistically.
In her view, under the influence of the new real estate regulations:
- Over the next 1–2 years, the industry will undergo a more thorough cleanup, with more companies exiting.
- In the long term, industry entry barriers will rise, widening the gap between companies.
- A small number of leading developers with both product strength and financing advantages may become beneficiaries under the new rules.
- Stock prices of some companies could rise 5–10 times over the next few years.
The Presale of Completed Homes Policy: A Key Variable
Li Bei identifies the new policy requiring presale of completed homes as a key variable reshaping industry dynamics.
For homebuyers:
- Purchasing completed homes allows deferring down payments and monthly installments.
- It avoids risks of unfinished projects and uncertain delivery quality.
- She estimates completed homes can command a price premium of approximately 3% to 5% over presale homes.
- Developers can share part of this premium, while buyers benefit from lower capital costs and reduced risk.
For developers:
- Market concerns focus on extended turnover cycles and compressed profits.
- However, Li Bei's calculations show that if land payments can be made in installments, project capital turnover time may increase by only 20% to 30%.
- Combined with modest price increases and lower land costs, net profit margins could actually improve by 20% to 30%.
Long-Term Industry Outlook
Li Bei predicts a significant increase in industry concentration:
- The proportion of exiting companies may rise from 90% to 95%, or even 98%.
- The number of companies capable of continuous nationwide land acquisition and expansion may fall to single digits.
Key barriers for new entrants:
- Land payment requirements, sales fund supervision, and the lead bank system raise entry thresholds.
- Financing capability and parent company guarantee capacity become core competitive advantages.
- Under the financing cost gap, central state-owned enterprises (SOEs) will further widen their advantage over private enterprises.
- Even when the industry reaches a high level of prosperity, few new players will consider entering, leading to a prolonged highly oligopolistic structure.
AI Bubble Risk: "First Drop Is Valuation, Second Wave Still to Come"
Beyond real estate, Li Bei also warned of risks in the AI sector.
Key judgments:
- Overseas AI capital expenditure is still growing, but quarter-on-quarter growth may peak around mid-next year.
- Cloud providers' investment increases were based on the assumption of rapid linear growth in AI model revenue.
- However, model companies' revenue growth has already slowed significantly.
- The adjustment in AI upstream-related stocks since July mainly reflects valuation contraction, not an immediate downward revision of profit expectations.
She believes the "second wave of decline" in the AI sector may occur when capital expenditure truly peaks and profits begin to fall.
Real Estate and Consumption: "Desert Oases" by Next Year
Li Bei argues that when the AI "fever" subsides, the U.S. economy slows, U.S. Treasury yields decline, and the dollar depreciates, China's real estate and consumption sectors may become "desert oases" for global assets.
On consumption:
- Consumption has undergone five to six years of adjustment.
- The steepest decline in residents' marginal propensity to consume may have passed.
- Even if the economy takes another step down next year due to declining real estate investment and exports, consumption should remain relatively stable.
- Many essential and some discretionary consumption leaders are at historically low valuations.
- Combined with dividends and share buybacks, some companies offer attractive static returns.
- She describes Chinese consumption stocks as "gold everywhere."
- When AI capital expenditure cools and domestic demand policies intensify, these companies may enter a main upward cycle.
On real estate:
- Previously, developers sold presale homes, with profits only recognized after completion and settlement.
- Profits this year reflect sales from two years ago.
- New home sales began improving in the second half of 2024, with quality projects achieving net profit margins of around 10%.
- This improvement will start showing in financial statements from the second half of this year.
- By next year, leading companies' reported net profit margins on development could rise from 1–3% to 5–7%, and potentially 7–10% the year after.
- This improvement comes from both the time lag between sales and settlement, and the clearing of impairment burdens.
Full Dialogue (Edited Excerpts)
01 Real Estate Opportunity: From Once a Decade to Once in Two Decades
Tencent Finance: How do you view the impact of the new presale of completed homes policy on real estate?
Li Bei: Previously, 90% of companies might exit, leaving 10%. Now, it could be 95% exiting, leaving only 5%. Presale of completed homes will further intensify the reshuffling. This needs to be viewed from both long-term and short-to-medium-term perspectives.
Current industry stage: The real estate sector is in the late stage of a downward cycle, entering a bottom-divergence phase.
- New home sales have improved since the second half of 2024.
- Quality companies' new projects can achieve nearly 10% net profit margins.
- Compared to second-hand homes, new "fourth-generation" homes offer higher floor area ratios, better design and quality.
- Many projects sell out upon launch.
- Second-hand home prices are still declining slightly nationwide, but the decline has narrowed significantly year-on-year.
- Shanghai has started to rise, and over a dozen second-tier cities are stabilizing and recovering.
Company level:
- 90% of companies have already exited.
- New home supply has dropped significantly and continues to decline.
- New construction starts are about one-quarter of peak levels; new launches are about one-fifth.
- However, surviving companies with strong product capabilities and financing have seen significant operational improvement over the past two years, though this is not yet reflected in financial statements due to reporting lags and ongoing impairment charges.
Impact of the new policy:
Many believe presale of completed homes will significantly extend developers' turnover cycles and increase financial costs, thereby sharply reducing ROE. Li Bei disagrees, citing three factors:
- Home prices can rise.
- Land prices will fall.
- Land payments can likely be made in installments.
On home prices:
- For presale homes, buyers pay a down payment now and service the mortgage for two years before delivery.
- For completed homes, buyers only pay a 5% deposit now, with down payment and mortgage starting two years later.
- Most buyers would prefer the latter.
- Therefore, completed homes can command a 3% to 5% price premium and still be more cost-effective than presale homes, while eliminating risks of unfinished projects and delivery uncertainty.
On land prices:
- Recent land auctions under the new policy show that without installment payments, land prices can be about 30% lower than comparable plots before the policy.
- If installment payments are allowed, local governments effectively share some of the capital turnover burden.
New financial model (steady state after one year): For a leading developer:
- Assuming 50% down payment for land, with the remainder paid after presale of completed homes.
- Capital turnover extends by only 20% to 30%.
- With modest leverage from group support, sales scale can be maintained.
- Home prices rise about 3%, land prices fall 10–20%.
- Net profit margins may actually increase by 20% to 30%.
Long-term results:
- More thorough cleanup: Exiting companies may rise from 90% to 95% or even 98%. Companies capable of continuous nationwide land acquisition and expansion may not exceed single digits.
- Higher entry barriers: New policies require full or at least half land payment upfront, plus sales fund supervision and lead bank systems. The old model of using minimal equity to acquire land and then leveraging through loans is largely unworkable.
- Financing capability as core competitiveness: The gap between private enterprises and central SOEs is enormous. Even mixed-ownership companies may face financing costs nearly 100 basis points higher than central SOEs, leaving little competitive advantage.
02 AI: First Drop Is Valuation, Second Wave Still to Come
Tencent Finance: How do you view the AI bubble debate?
Li Bei: AI capital expenditure is growing strongly this year and will continue next year, but quarter-on-quarter growth may peak around Q2 next year, possibly Q3 or Q4.
Cloud providers raised capital expenditure after AI models succeeded in coding, with Anthropic's ARR rising rapidly. After Q1, cloud providers significantly increased capital expenditure in Q2, assuming linear extrapolation: model revenue could reach about 500 billion by year-end, exceeding 1 trillion in subsequent years. If revenue truly exceeded 1 trillion, it would match the current 1 trillion-plus capital expenditure, making it reasonable rather than a bubble.
However, ARR growth slowed significantly in Q2. Li Bei repeatedly warned of this risk in May-June. The current market adjustment reflects this reality.
From April to June, the market significantly raised profit expectations for 2027 due to higher capital expenditure, and valuations based on 2027 also increased. The current decline is not a "Davis double-kill" but a valuation contraction: the market still recognizes significant profit growth in 2027, but if ARR cannot sustain, investment levels are unsustainable. Capital expenditure likely peaks in 2027 or 2028. Without end-user revenue growth, industry chain investment cannot be maintained, and high valuations cannot be justified.
Li Bei believes the second wave of AI decline will occur when capital expenditure truly peaks, possibly around mid-next year. At that point, profit expectations will begin to decline, triggering a second drop.
03 Real Estate and Consumption: "Desert Oases" by Next Year
Tencent Finance: What is the logic behind investment opportunities in consumption and real estate?
Li Bei: Some may ask: if the economy is weak, why would consumption and real estate have opportunities? Because their profits will no longer decline.
Consumption:
- Consumption has experienced five to six years of decline.
- Consumer sentiment has stabilized; the steepest decline in marginal propensity to consume has passed.
- After the shock of the pandemic and housing price drops, panic savings have subsided, and sentiment is stabilizing.
- Essential consumer goods are less affected by economic fluctuations.
- Both essential and some discretionary consumption leaders are at historically low valuations.
- Examples: Dairy leader (40-50x PE in 2021, now 10x); frozen food and condiment leaders (three-digit PE in 2021, now around 10x).
- Dividend yields plus buybacks offer 4% to 7% static returns for many companies.
- Once the AI narrative collapses or domestic demand policies are introduced, these stocks will be discovered as "gold everywhere."
Real estate:
- Previously, developers sold presale homes, with profits recognized only after completion and settlement.
- This year's profits reflect sales from two years ago.
- New home sales improved in H2 2024, with quality projects achieving ~10% net profit margins.
- This improvement will appear in financial statements from H2 this year.
- By next year, leading companies' reported net profit margins on development could rise from 1–3% to 5–7%, and potentially 7–10% the year after.
- This improvement comes from both the time lag between sales and settlement, and the clearing of impairment burdens.
By this time next year, real estate companies' financial statements will reflect 2025 sales, which were actually quite good in H1 2025. Real estate and consumption will emerge as "desert oases."
Source
腾讯财经Eastern