C for me: long-term Treasury yields. I agree investors need to look beneath headline numbers. NFP can look strong while revisions, participation, hiring breadth and duration of unemployment tell a more nuanced story. August payrolls rebounded strongly, but longer-term unemployment remains a concern. Right now I am watching long yields most closely. The 10Y has already tested 5%, while fiscal deficits, Treasury supply and inflation expectations can keep long-term borrowing costs elevated independently of the Fed's next move. That matters directly for equity valuations, mortgages and corporate financing. The Fed just hiked to 3.75%-4.00% and its projections remain hawkish, but the bond market may tell us more about financial conditions than simply guessing the next FOMC decision. So yes, NF

I don't think the market has fully accepted the second hike yet. The 25bp move was largely priced in, but the hawkish surprise was the path ahead. The Fed's September projections show 16 of 18 participants expecting rates to end 2026 above the new 3.75%-4.00% range, with 12 clustered around a 4.00%-4.25% target range. The lack of a stock rally despite an expected hike suggests investors are still digesting "higher for longer". Treasury yields reinforce that pressure, with the 2Y around 4.67% and 10Y around 5.00%. For equities, I think the next CPI and jobs data matter more than the dots themselves. Strong earnings can support the market, but if inflation stays sticky enough to make another hike increasingly credible, high-valuation growth stocks face a tougher discount-rate environment. So

A, META caught my attention most. JPMorgan's case makes sense to me because Meta already has huge distribution, a strong ad business and multiple ways to monetise AI if Muse gains real adoption. The main question is whether the enormous AI capex eventually produces enough incremental earnings. B, NBIS is the one I'd question most. I like the AI infrastructure growth story, but a $355 target after such a huge run requires a lot of future growth to go right. At that valuation, execution, utilisation and capex discipline matter just as much as revenue growth. LAC is interesting too, but that call depends heavily on where lithium prices go, while Morgan Stanley's caution on NVO looks more understandable given slowing growth and increasing GLP-1 competition.

The stablecoin story can keep growing, but the valuation story just became harder. CLARITY failing does not stop USDC adoption. Circle still reported USDC circulation +19% YoY and on-chain volume +151%. What disappeared is the near-term regulatory catalyst that could justify institutions committing more capital with confidence. So I’d separate adoption from the stocks. Stablecoins can keep expanding while CRCL and COIN multiples compress. The next leg needs evidence: continued USDC growth, institutional adoption and eventually a durable regulatory framework. Until then, regulatory uncertainty deserves a discount.

The market has taken the AI slowdown call seriously, but so far it is pricing risk to the pace of compute growth, not the end of the compute boom. That explains why chips and other “picks and shovels” were hit hardest while hyperscalers held up much better. The real confirmation would be cancelled GPU orders, delayed data centres or actual capex cuts. With hyperscaler capex still projected around US$795B this year and US$1.08T in 2027, we have not seen that yet. Tuesday’s chip rebound suggests Monday may have been an initial risk repricing. I’d watch capex guidance and orders next. If those get cut, the slowdown trade becomes much more fundamental.

I think the banks are right about one thing: the 25bp hike itself is largely priced in. With markets putting roughly 90%+ odds on it, the bigger risk is not Wednesday’s hike but what comes next. If the Fed signals this is a limited adjustment, earnings and growth can probably keep supporting equities. But if oil, inflation and yields force markets to price a longer hiking cycle, “priced in” gets recalculated very quickly. With the 10Y around 5%, I’m watching the Fed’s message more than the 25bp headline.

B. USD 22.19. Using a 360-day basis: USD 10,000 × 7.99% × (10/360) ≈ USD 22.19. The 7.99% is an annual rate, so for 10 days we only pay the corresponding fraction of the annual interest.

I buy the thesis, but not necessarily the price after a 14% jump. AI creates a strong structural tailwind for cybersecurity. More autonomous agents mean more identities, endpoints, cloud workloads and attack surfaces to monitor. Security is also one of the harder IT budgets to cut when the threat itself is getting stronger. But cybersecurity cannot simply replace the semiconductor trade. The addressable spending pool is much smaller, and after CRWD and PANW jumped 13%+ in one session, a lot of enthusiasm has been pulled forward. So I would buy the theme, not chase the spike. CRWD and PANW are my preferred names on a pullback. The real confirmation comes when AI-security fears translate into sustained ARR growth, larger contracts and higher guidance. More dangerous AI = more security spendi

I don't think the market has fully reacted yet. The 10Y briefly crossing 5% matters, but the S&P 500 falling only 0.48% while semis plunged nearly 6% suggests rotation rather than broad risk-off selling. The key is whether 5% becomes a ceiling or a new floor. If yields settle back below 5%, equities can probably absorb it. But if the 10Y holds above 5% and keeps climbing, valuation pressure should spread beyond chips into other long-duration growth stocks. With markets now pricing roughly a 90%+ chance of a 25 bp Fed hike, the hike itself is largely expected. I think the bigger catalyst is what the Fed signals about further hikes. For now: rotation, not capitulation. But sustained 5%+ yields would make me considerably more cautious.

My pick is $BlackRock(BLK)$. I prefer its combination of scale, recurring fee income and long-term growth exposure through ETFs and private markets. The US$5.73 dividend is attractive, but I would rather own BLK for business quality and compounding than chase the much higher but more cyclical shipping yields. Analyst consensus is also constructive, with an external S&P Global consensus target around US$1,321. I would cross-check the exact target under Tiger Trade's Analysis tab since its figure may differ. For this week, $UnitedHealth(UNH)$ interests me. It goes ex-dividend on 14 Sep at US$2.32/share. I would view the dividend as a bonus rather than the main reason to buy. For dividend investing, underlying earnings and dividend sustainability matter more to me than simply picking the

My pick is $BlackRock(BLK)$. I prefer its combination of scale, recurring fee income and long-term growth exposure through ETFs and private markets. The US$5.73 dividend is attractive, but I would rather own BLK for business quality and compounding than chase the much higher but more cyclical shipping yields. Analyst consensus is also constructive, with an external S&P Global consensus target around US$1,321. I would cross-check the exact target under Tiger Trade's Analysis tab since its figure may differ. For this week, $UnitedHealth(UNH)$ interests me. It goes ex-dividend on 14 Sep at US$2.32/share. I would view the dividend as a bonus rather than the main reason to buy. For dividend investing, underlying earnings and dividend sustainability matter more to me than simply picking the

My pick is $BlackRock(BLK)$. I prefer its combination of scale, recurring fee income and long-term growth exposure through ETFs and private markets. The US$5.73 dividend is attractive, but I would rather own BLK for business quality and compounding than chase the much higher but more cyclical shipping yields. Analyst consensus is also constructive, with an external S&P Global consensus target around US$1,321. I would cross-check the exact target under Tiger Trade's Analysis tab since its figure may differ. For this week, $UnitedHealth(UNH)$ interests me. It goes ex-dividend on 14 Sep at US$2.32/share. I would view the dividend as a bonus rather than the main reason to buy. For dividend investing, underlying earnings and dividend sustainability matter more to me than simply picking the

B. A USD 300 loss. You short-sell 10 shares at USD 100, receiving USD 1,000. When the price rises to USD 130, buying back those 10 shares costs USD 1,300. Loss = USD 1,000 − USD 1,300 = −USD 300. Borrowing the shares does not protect you from losses. A short seller profits when the share price falls and loses when it rises. This also highlights the key risk of short selling: the potential loss is theoretically unlimited because a stock price has no fixed upper limit.

B. A USD 300 loss. You short-sell 10 shares at USD 100, receiving USD 1,000. When the price rises to USD 130, buying back those 10 shares costs USD 1,300. Loss = USD 1,000 − USD 1,300 = −USD 300. Borrowing the shares does not protect you from losses. A short seller profits when the share price falls and loses when it rises. This also highlights the key risk of short selling: the potential loss is theoretically unlimited because a stock price has no fixed upper limit.

At WTI above US$100, I would rather manage energy exposure than chase the spike. The diesel crack above US$110 is particularly important because it suggests the pain is moving beyond crude into refined products, which can feed directly into transport costs and inflation. I would keep some energy exposure as a geopolitical and inflation hedge, but favour profitable producers and integrated majors with strong cash flow rather than high-beta names that need oil to keep rising. After a 7% one-day move, the risk-reward for adding aggressively looks poor. My approach: hold the hedge, take some profit into strength, and keep dry powder for a pullback. If US$100 becomes a durable floor rather than a temporary geopolitical premium, energy earnings estimates may still have room to rise.

Burry has a point, but I would separate $NVDA from $PLTR. Nvidia’s valuation looks far easier to grow into. AI demand remains powerful, while NVDA trades around 24x forward earnings. That is not obviously bubble territory if earnings continue compounding strongly. Palantir is the harder call. Q2 revenue surged 93% YoY and margins expanded impressively, but PLTR still trades around 87x forward earnings and ~65x sales. At that price, excellent execution is already expected. So I agree more with Burry on PLTR than NVDA. Great company does not automatically mean great stock at any price. My pick: NVDA can grow into its valuation; PLTR needs near-perfect execution to justify its own. I would not short either aggressively, but PLTR has much less room for disappointment.

I favour the uranium supply chain for durability. AI data centres may accelerate nuclear demand, but uranium benefits from the broader reactor fleet and fuel-security needs, rather than depending on any single SMR design reaching commercial scale. BE is interesting because fuel cells can address the nearer-term problem: data centres need reliable power before new nuclear plants can realistically arrive. SMR offers the biggest upside if deployments scale, but also the greatest execution, financing and regulatory risk. My ranking: uranium for the strongest long-term risk/reward, BE for the nearer-term AI power bottleneck, and SMR as speculative optionality. The AI electricity shortage looks structural, but that does not make every power stock a structural winner. I would rather own the bott

B. GOOGL for me. Meta may have the larger social distribution opportunity, but Google already sits much closer to commercial intent through Search, Shopping, YouTube and its broader ecosystem. If Gemini agents can move users from “search” to “act and transact”, Google has a clearer path to monetisation. META could ultimately surprise if it converts its huge user base into commerce, but that requires changing user behaviour and proving transaction economics. For now, I would rather own GOOGL for the AI-agent trend. It has both distribution and an existing monetisation engine to build upon.

I favour the uranium supply chain for durability. AI data centres may accelerate nuclear demand, but uranium benefits from the broader reactor fleet and fuel-security needs, rather than depending on any single SMR design reaching commercial scale. BE is interesting because fuel cells can address the nearer-term problem: data centres need reliable power before new nuclear plants can realistically arrive. SMR offers the biggest upside if deployments scale, but also the greatest execution, financing and regulatory risk. So I see it as: uranium for the durable structural thesis, BE for nearer-term AI power demand, and SMR as the higher-risk optionality. Given the sector's high beta, I would expect plenty of momentum-driven volatility even if the long-term power-demand thesis remains intact.

B. GOOGL for me. Meta may have the larger social distribution opportunity, but Google already sits much closer to commercial intent through Search, Shopping, YouTube and its broader ecosystem. If Gemini agents can move users from “search” to “act and transact”, Google has a clearer path to monetisation. META could ultimately surprise if it converts its huge user base into commerce, but that requires changing user behaviour and proving transaction economics. For now, I would rather own GOOGL for the AI-agent trend. It has both distribution and an existing monetisation engine to build upon.