To The Moon

Speaker: Dr. Franklin Wu @老实人谈美股 (Quantitative Researcher at a financial institution in Shanghai; PhD, University of Chicago — background in physics, quantitative research and trading strategy development)
Live Date: September 2, 2026 (Review Live >>)

In this livestream, Dr. Franklin Wu walked through a repeatable "Macro → Industry → Technicals → Risk" framework for AI and semiconductor investing — covering a hawkish Jackson Hole and the Treasury's "Bessent Put," why the AI compute chain runs far beyond Nvidia, how to turn a bullish view into an actual trade using QQQ and SPCX as case studies, and how (and when) leveraged ETFs like $Direxion Daily Semiconductors Bull 3x Shares(SOXL)$ and $ProShares UltraPro QQQ(TQQQ)$ actually fit into a portfolio.

Want a deeper dive? We broke this session down into 4 full recap articles, each covering a different piece of the framework>

• Live Recap 1: A Hawkish Jackson Hole, the "Bessent Put," and Why Tech Is Watching the 30-Year

• Live Recap 2: Why Semiconductors Are Still the "Hard Currency" of the AI Trade

• Live Recap 3: Turning a Bullish View Into a Trade — Technicals, QQQ and SPCX Case Studies

• Live Recap 4: Leveraged ETFs, the September Playbook, and Q&A Highlights

Prefer to watch the highlights? Catch these key moments from the live session in short clip form>

• 【Livestream Clip 1｜Dr. Franklin Wu: Franklin's 3 Takeaways: AI Is Strong, But 3x ETFs Mislead You】

• 【Livestream Clip 2｜Dr. Franklin Wu: It's Long-Term Yields, Not Earnings, That Hit High-Valuation Tec

• 【Livestream Clip 3｜Dr. Franklin Wu: Why TQQQ Looks Right But Your Account Never Recovers】

• 【Livestream Clip 4｜Dr. Franklin Wu: 4 Checks Before You Trade: Macro, Industry, Technicals, Risk】

🐯💬 Join the discussion: Share your market view or questions below. Every useful and thoughtful comment will receive Tiger Coins!

🎯 5 Key Takeaways

• Fed Chair Kevin Warsh's Jackson Hole message put inflation first, raising September rate-hike odds to roughly 56% from about 35% — pushing up the discount-rate bar that rate-sensitive tech and AI names have to clear.

• The Treasury's expanded long-bond buybacks (per-operation cap raised to about $4bn) act as a cushion on the 30-year yield, but Franklin was clear it's "a cushion, not a superhero" — it doesn't erase underlying deficit and supply pressure.

• AI semiconductors remain one of the strongest US growth themes, but the demand chain runs well beyond GPUs — into memory/HBM, networking/optics, and power/cooling — and Franklin's own pick for the tightest bottleneck right now is memory, citing an expected shortage into 2027.

• $NVIDIA(NVDA)$'s Q2 FY27 print was a blowout ($96.2bn revenue, +106% y/y; $89bn data-center revenue, +117% y/y) with Q3 guidance of ~$108bn — but Franklin's lesson was that fundamentals can be excellent while expectations run even higher, which is why the guide, not just the beat, moves the stock.

• Leveraged ETFs like $Direxion Daily Semiconductors Bull 3x Shares(SOXL)$ and $ProShares UltraPro QQQ(TQQQ)$ target 3× daily returns, not 3× long-term returns — daily resetting creates volatility decay, which is why Franklin suggested strict position limits (roughly under 20% of portfolio for TQQQ, under 10% for SOXL) and short holding periods only.

🏦 A Hawkish Fed, Watching the Discount Rate

Franklin opened with the macro backdrop that sets the "risk temperature" for every AI and semiconductor trade. Fed Chair Kevin Warsh's Jackson Hole remarks treated the 2% PCE target as fixed and described financial conditions as not broadly restrictive — a message the market read as hawkish, with September rate-hike odds jumping from around 35% to roughly 56%.

Discussion: Do you think the Fed's inflation-first stance holds through September, or does incoming data force a softer tone?

📉 The "Bessent Put" — A Cushion, Not a Cure

With 30-year yields near multi-year highs, the Treasury expanded long-bond buybacks to roughly $4bn per operation — nicknamed the "Bessent Put." Franklin's framing: this can slow a yield spike and give valuations room to breathe, but it doesn't make the deficit, inflation, or future bond supply disappear. Around August 25, the 10-year sat near 4.64–4.68% and the 30-year near 5.17–5.21%, both of which he flagged as key variables for tech valuations going into September.

🧮 Semiconductors: The "Hard Currency" of the AI Trade

Franklin's core logic: more AI training and inference ultimately means more data centers, which means more chips, networking, memory and power. Gartner's forecast — cited in the session — has the AI data-center ecosystem's share of semiconductor revenue rising from 36.5% in 2026 to more than 53% by 2030. Rather than memorizing every ticker in the compute chain, his suggested approach is to pick one or two layers you actually understand — core compute, memory/packaging, networking/optics, infrastructure, or application monetization — and ask whether demand is turning into real profit.

💥 NVDA's Blowout Quarter, and Why the Guide Still Matters

$NVIDIA(NVDA)$'s Q2 FY27 revenue came in at $96.2bn (+106% y/y), with data-center revenue of $89.0bn (+117% y/y) and a 75.0% gross margin — with Q3 guidance calling for roughly $108bn. Franklin's read: fundamentals can be excellent, but expectations can be even higher, which is exactly why the guide, not the beat itself, is what moves the stock. He applied the same lens to AMD (whose second-source thesis needs proof via repeat customer orders and share gains, not just being "cheaper than $NVIDIA(NVDA)$") and to $Broadcom(AVGO)$/ $Marvell Technology(MRVL)$ (benefiting from hyperscalers' shift toward custom ASICs).

📐 Turning a View Into a Trade: Three Questions, Two Case Studies

Before buying anything, Franklin's checklist is Direction (are macro and industry trends aligned?), Location (is price near meaningful support?), and Risk (if the thesis is wrong, where does the loss stop?). He applied this to two live case studies: QQQ's late-July pullback into its ~680 support zone, where he'd scale in with roughly a 50% position and add only if the level held on improving volume; and the SPCX ~110 "bottom-fishing" setup, where his rule was small first, confirmation second, bigger later — since "cheaper than yesterday" isn't the same as "cheap."

🎯 Leveraged ETFs: 3× Daily Is Not 3× Long-Term

$Direxion Daily Semiconductors Bull 3x Shares(SOXL)$ and $ProShares UltraPro QQQ(TQQQ)$ both target roughly 300% of their underlying index's daily return — SOXL on the semiconductor index, TQQQ on the Nasdaq-100. Because they reset daily, choppy markets create volatility decay: Franklin's example showed a 3× ETF going +15% then -15% ends near -2.25%, not flat, because the second move compounds off a different base. His guidance: treat these as short-term, high-conviction tools only, with position limits around 20% of portfolio for TQQQ and 10% for SOXL.

⚖️ On Risk Management

Asked how he thinks about risk across the whole framework, Franklin was consistent throughout: never let a big story override a specific stop. Even with a strong long-term thesis on AI, he starts small and adds only after the market confirms the idea — because a large loss is much harder to make small again than a small loss is to grow.

💬 Words from Franklin Wu

"We can fall in love with a person, but never with a company."
"The market setting tells us what we should be careful of."
"Cheaper than yesterday is not the same thing as cheap."
"Never let a big story override a specific stop."

Closing Takeaway

Nothing Franklin discussed points to the AI thesis breaking down — not in the Fed's stance, not in Nvidia's quarter, not in the broader compute chain. What he laid out instead was a discipline for staying in the trade safely: check the macro temperature first, confirm the industry demand is real, find a sensible entry, and define the exit before ever placing the trade. His closing four-step playbook — macro, industry, technicals, risk — was framed less as a forecast and more as a checklist to run before every single position.

Post-Event Resources

Viewers can follow Dr. Franklin Wu's recap and future updates on @TBlive and @老实人谈美股 on Tiger Community. The full livestream replay is available on the Tiger Trade app.

🐯 Your Turn: Join the Discussion

Share your view on one of these questions:

• Do you think September plays out closer to Franklin's base case, bull case, or risk case?

• Was NVDA's post-earnings move justified by the guide, or already priced in before the print?

• Where would you personally draw the line on how much of a portfolio belongs in leveraged ETFs like SOXL or TQQQ?

🎁 Every useful, thoughtful, and well-explained comment will receive Tiger Coins.

Let's compare different views and learn from one another.