#AI股持续上涨还有哪些投资机会
Four directions, from low risk to high risk:
1️⃣ Untapped parts of the AI industry chain (the safest)
Data centers/IDC: PE percentile 59%, PB 73%. Runze Technology’s PE is 20.4x, +28% year-to-date, and +9.56% over the past 5 trading days; for computing power leasing, the PE percentile is only 34%, and cash flow can be verified. Cloud computing: PE/PB/PS percentiles are all within 41%~44%. The Cloud Computing ETF from E Fund (sh516510) has a size of 2.53 billion and is up +5.85% over the past 5 days; the risk-reward looks the cleanest. Communications ETF from Cathay (sh515880) has a size of 59.7 billion, +37.8% year-to-date, and the PB percentile is observable.
2️⃣ Broad-market funds + dividends to hedge (for those already heavily positioned in AI)
Dividend-low-volatility ETF from HuaTai-PineBridge (sh512890): +11.25% over the last 60 days, with a negative correlation to AI. Targets: sh512890, sz159201 (free cash flow), sh563020. Underlying dividend yields: China State Construction 6.32%, China Telecom 4.11%, China Merchants Bank 4.94%. Price appreciation over 60 days: +15%~28%—slow but dependable.
3️⃣ Discount in Hong Kong tech stocks
Hang Seng Tech ETF from Huaxia (sh513180): size 39.8 billion, -23.77% year-to-date, with a 37.7% drawdown from the peak. For the Hong Kong Connect Internet ETF, Fullgoal (Fuguo) is -34.5% year-to-date. The same batch of AI stocks shows a systematic discount in Hong Kong; over the past 5 days, it rebounded +2.6%~2.8%.
4️⃣ AI applications (high risk)
AI marketing led with +8.35% over the past 5 days, but the sector’s PE is 437x. Performance realization is questionable, so it’s only suitable for investors who can tolerate a drawdown of 30%+.
⚠️ Three points to note
1. Watch out for the “new fund trap.” A crowd of new issues for the computing-power ETFs on the ChiNext: e.g., E Fund sz158050, Huaxia sz158041, and others (9 in total). Small size, highly homogeneous exposures, management fee 0.5%. Huaxia’s same product shows a turnover rate of 99.21%. The chips rotate day by day—more like a speculation vehicle than an allocation/positioning one.
2. Choose older funds, not new ones. AI ETF from E Fund (sz159819) was founded in 2020; size 24.7 billion; fee 0.15%—far better than Ping An’s sh512930 (size 3.3 billion) and Fullgoal’s sh589380 (size 0.168 billion). With small funds, the bid-ask spread and the risk of liquidation are both higher.
3. Look at the last 60 days, not the last 20. CPO and AI chips have still been down 27%~30% over the last 60 days; the recent rebound is just a repair. Meanwhile: PetroChina +28%, Shenhua +25%, ICBC +14.85%, CCB +15.84%. The medium-term strength is on the dividend side. Use “older broad-market/dividend” exposures to dampen volatility in your AI position—don’t chase new launches, and don’t treat a rebound as a reversal. $BTC $ETH
Four directions, from low risk to high risk:
1️⃣ Untapped parts of the AI industry chain (the safest)
Data centers/IDC: PE percentile 59%, PB 73%. Runze Technology’s PE is 20.4x, +28% year-to-date, and +9.56% over the past 5 trading days; for computing power leasing, the PE percentile is only 34%, and cash flow can be verified. Cloud computing: PE/PB/PS percentiles are all within 41%~44%. The Cloud Computing ETF from E Fund (sh516510) has a size of 2.53 billion and is up +5.85% over the past 5 days; the risk-reward looks the cleanest. Communications ETF from Cathay (sh515880) has a size of 59.7 billion, +37.8% year-to-date, and the PB percentile is observable.
2️⃣ Broad-market funds + dividends to hedge (for those already heavily positioned in AI)
Dividend-low-volatility ETF from HuaTai-PineBridge (sh512890): +11.25% over the last 60 days, with a negative correlation to AI. Targets: sh512890, sz159201 (free cash flow), sh563020. Underlying dividend yields: China State Construction 6.32%, China Telecom 4.11%, China Merchants Bank 4.94%. Price appreciation over 60 days: +15%~28%—slow but dependable.
3️⃣ Discount in Hong Kong tech stocks
Hang Seng Tech ETF from Huaxia (sh513180): size 39.8 billion, -23.77% year-to-date, with a 37.7% drawdown from the peak. For the Hong Kong Connect Internet ETF, Fullgoal (Fuguo) is -34.5% year-to-date. The same batch of AI stocks shows a systematic discount in Hong Kong; over the past 5 days, it rebounded +2.6%~2.8%.
4️⃣ AI applications (high risk)
AI marketing led with +8.35% over the past 5 days, but the sector’s PE is 437x. Performance realization is questionable, so it’s only suitable for investors who can tolerate a drawdown of 30%+.
⚠️ Three points to note
1. Watch out for the “new fund trap.” A crowd of new issues for the computing-power ETFs on the ChiNext: e.g., E Fund sz158050, Huaxia sz158041, and others (9 in total). Small size, highly homogeneous exposures, management fee 0.5%. Huaxia’s same product shows a turnover rate of 99.21%. The chips rotate day by day—more like a speculation vehicle than an allocation/positioning one.
2. Choose older funds, not new ones. AI ETF from E Fund (sz159819) was founded in 2020; size 24.7 billion; fee 0.15%—far better than Ping An’s sh512930 (size 3.3 billion) and Fullgoal’s sh589380 (size 0.168 billion). With small funds, the bid-ask spread and the risk of liquidation are both higher.
3. Look at the last 60 days, not the last 20. CPO and AI chips have still been down 27%~30% over the last 60 days; the recent rebound is just a repair. Meanwhile: PetroChina +28%, Shenhua +25%, ICBC +14.85%, CCB +15.84%. The medium-term strength is on the dividend side. Use “older broad-market/dividend” exposures to dampen volatility in your AI position—don’t chase new launches, and don’t treat a rebound as a reversal. $BTC $ETH