The one-year low in Chinese shares was triggered by the very report that China may allow ByteDance and Alibaba to buy Nvidia's new chips — news that is actually good for Chinese AI builders over time, so the selloff looks like panic rather than deteriorat

Verdict: a real thesis with a confirmed-approval engine — but no entry yet, so wait

The strongest case for this idea is that the selloff happened on constructive news — a Reuters report that China may let ByteDance and Alibaba-affiliated buyers purchase Nvidia's new chips — and that revenue lines are still growing (BABA +8.1%, JD +17.9%, PDD +14.5% year over year), consistent with panic rather than demand collapse. The strongest case against is that the damage is not just sentiment: JD's free cash flow fell roughly 78% to $1.6B, BABA's operating cash flow halved to $11.0B and its dividend was cut to $1.05 per share from $2.00, and ownership filings for the quarter ended June 30, 2026 show net open-market insider selling at both BABA (about -$70.8M) and PDD. The verdict: wait. The setup is live as a watch-list only — the entry rules have not fired, and the binding condition is RSI (14) below 30, with BABA at 41.7 and the others 7 to 15 points away. What would flip it: either an official Chinese confirmation of the Nvidia approvals (which may drive the snap-back that aborts the entry) or a triggered capitulation-and-reclaim across a name, at which point the trade becomes actionable on its own 2.4% stop and 4.7% target terms.

Trade now: four entries armed, none triggered — wait for RSI under 30 and the middle-band reclaim

This is a watch-list setup today, not a signal. The rules were checked on real daily bars across all four names and did not open an entry, so there is nothing to buy yet. Each name needs four things at once: a daily low at or below the lower Bollinger band (20-day, 2 standard deviations), RSI (14) below 30, a MACD histogram cross above its signal line, and a close back at or above the middle band. One of the four is already true everywhere: every ticker closed below its lower band (BABA $109.94 vs. $111.25; FXI $34.28 vs. $34.75; JD $26.74 vs. $27.27; PDD $78.17 vs. $81.12), and the MACD cross is marked as near. The binding condition is RSI: BABA sits at 41.7 (needs to fall below 30, about 11.7 points away), FXI at 44.8 (about 14.8 away), JD at 37.4 (about 7.4 away), and PDD at 38.8 (about 8.8 away). The close at or above the middle band is also unmet in all four cases — that is the capitulation-and-reclaim part, and it deliberately requires a bounce before entry. If an entry triggers, the strategy's own exits frame the trade: a fixed stop at minus 2.4% of position value, a take-profit at plus 4.7%, a fail-safe exit if the close drops back below the lower band, and a 90-day time stop. That is roughly 2:1 reward to risk per position, sized so each position risks 2.4% of equity with a 25% maximum position weight and one trade symbol per rule. "Wait" means something concrete: set alerts on RSI (14) crossing below 30 for each ticker, watch for the first daily close back above the middle band afterward, and do nothing until both happen together with the MACD cross and the band-low touch in the same window. One scope limit to note: the sensitivity evaluation ran out of its time budget, so no nearby-parameter recommendation was established — trade the setup exactly as written. The context supports patience rather than chasing. The idea argues the one-year low was panic selling on news (a reported China approval path for Nvidia chips) that is actually constructive for Chinese AI builders, and that global managers are ending a four-year China underweight. But the same data shows the pain is real: JD's full-year net income fell about 46% year over year and its free cash flow dropped roughly 78% to $1.6B; PDD's net margin compressed to 22.7% from 28.5%; and BABA's operating cash flow halved to $11.0B. Letting the strategy demand a genuine capitulation plus reclaim before entry is the right discipline against catching the first knife.

Growth Is Still Real Underneath the Panic — and Big Money Is Coming Back

The idea argues the one-year low in Chinese shares (per Bloomberg's September 28 piece on chip and optical firms sliding) was triggered by a Reuters report that China may allow ByteDance and Alibaba to buy Nvidia's new chips — a report the thesis reads as long-term positive, not negative. That framing is a genuine contrarian bet: if chip access is confirmed, the AI-exposed names sold hardest have the most snap-back room. The fundamentals show companies that are growing while they invest, which is consistent with a panic-driven selloff rather than demand collapse. Look at the top line. Alibaba grew FY revenue (period ended March 31, 2026) 8.1% to $148.4B, with a 39.8% gross margin that has been stable-to-improving since FY2023. JD grew revenue 17.9% to $187.2B, placing it in the top 14% of its consumer discretionary peer group for growth, and gross margin actually ticked up to 16.0%. PDD grew 14.5% to $61.8B while still posting a 21.6% operating margin and 23.7% return on equity — top-decile profitability among tech peers. These are not deteriorating franchises at the revenue line; they are companies spending heavily into a cycle. Alibaba's balance sheet reinforces the case: $19.1B in cash against a current ratio of 1.28, plus a dramatic share-count reduction — outstanding shares fell roughly 90% year over year in the reported series, an unusually aggressive contraction that mechanically supports per-share value. JD still holds $30.5B in cash and short-term investments against $57.4B in total liabilities, and PDD sits on $60.4B of cash and short-term investments with a current ratio of 2.43 and essentially no heavy fixed assets. None of these balance sheets is a forced seller. The positioning side of the thesis also has support: per Bloomberg's September 28 report, global fund managers have ended a four-year underweight on China, drawn by AI prospects and cheap valuations. Combined with the contrarian logic — the selloff occurring on news that is, if anything, constructive for Chinese AI builders — the setup favors buying dislocation rather than chasing momentum. One process note: the defined capitulation-and-reclaim entry rules (a touch of the lower Bollinger band with RSI under 30, then an MACD crossover and close back at or above the middle band) have not fired in the evaluated daily-bar window across BABA,…

Scores

• Conviction score breakdown: 43
• Thesis support: 62
• Trade readiness: 38
• Risk quality: 45
• Trigger proximity: 30
• Fundamentals trend: 38

Watch items

• BABA — RSI (14)
• BABA — Close vs Bollinger middle (20)
• FXI — RSI (14)
• JD — RSI (14)
• PDD — RSI (14)
• JD — Free cash flow
• BABA — Insider net open-market activity
• BABA — China approval of Nvidia chip purchases