Oil above US$100 changes the market equation for me. The biggest issue is not simply higher petrol prices, but the chain reaction: higher energy and transport costs → higher inflation → higher-for-longer interest rates → pressure on corporate margins and equity valuations. 🟢 Potential winners: Energy Oil producers such as $Exxon Mobil (XOM)$, $Chevron (CVX)$ and $ConocoPhillips (COP)$ should generally benefit if crude remains elevated because higher realised oil prices can translate into stronger cash flow. Refiners may also benefit when refining margins are favourable. We have already seen this rotation: when Brent moved above US$100 on 9 September, the S&P 500 Energy sector gained 1.1% while every other S&P sector declined. 🟡 Technology: Strong fundamentals meet a macro headwind

Markets are heading into a catalyst-heavy week. AI remains the strongest theme, with Wall Street ending Friday higher as Microsoft rallied 3.7% and the major indices all gained. At the same time, I am watching the other side of the equation: oil and Treasury yields. Brent is rising again today as US-Iran tensions remain unresolved, which could revive inflation and rate concerns. My main stock to watch is $Micron Technology (MU)$. Its earnings on 30 September could be an important test for the entire AI memory trade. I will be watching HBM demand, pricing, margins and especially forward guidance rather than simply whether MU beats headline estimates. $Microsoft (MSFT)$ and $Akamai (AKAM)$ are also on my watchlist after renewed AI enthusiasm. Akamai's US$11.6 billion, seven-year Anthropic cl

I don't think the market suddenly discovered dilution a day late. The shares were being sold through an ATM programme, so the dilution was happening gradually and was already known. What changed when the raise was completed was certainty: 29.3 million new shares had actually been issued, raising US$1.94B. That's roughly 5% dilution against Rocket Lab's previous common-share count. Thursday's rally may therefore have been relief that the Iridium financing was secured and the US$3.6B bridge facility could be cancelled. Friday then looked more like the market digesting the other side of that trade: a stronger balance sheet, but more shares dividing the future upside. So I wouldn't say dilution was priced exactly one day late. The market first rewarded financing certainty, then reconsidered w

A cosy Mid-Autumn night under the glowing full moon Tiger and his little bunny friends share mooncakes, warm tea and laughter beneath the lanterns. Sometimes, the sweetest part of the festival is simply having someone to share it with.

Yes, but only slightly until Meta discloses the economics. A transaction fee gives Muse something more valuable than engagement alone: a direct path from AI usage to revenue. But the fee percentage is only one part of the equation. I would want to know transaction volume, average order value, merchant adoption and how often Muse actually completes a purchase rather than simply helping users research one. Even a small cut could become meaningful at massive scale, especially if Muse becomes the interface through which users shop, travel and order services. Conversely, a generous fee means little if merchants resist it or completed transactions remain small. So I would give Muse more strategic value today, but not materially raise my META valuation yet. Show me the take rate and transaction

Alphabet feels it first to me. Expedia may actually benefit if Muse becomes another distribution channel bringing it bookings. Amazon has more protection because it owns the marketplace, checkout relationship and customer data, and it has already shown that it is willing to block Muse from shopping on its platform. Alphabet is more exposed at the starting point of the journey. If users increasingly tell Muse “find me a hotel and book it” or “find this product and buy it”, fewer commercial journeys may begin with a Google search, followed by sponsored links and comparison shopping. That does not mean Google's business is immediately threatened. Muse still needs partners, merchants and payment infrastructure to complete transactions. But the bigger question is whether the AI agent becomes th

A strong Micron print could stop the pullback, but I don't think a headline beat alone will be enough. Expectations are already very high. Micron itself guided to about $50B revenue, 86% gross margin and $31 non-GAAP EPS, so the market will probably care more about what comes next than whether Q4 simply beats those numbers. For me, the real catalysts are strong FY27 guidance, continued pricing power, HBM4 execution and evidence that tight supply persists. If those strengthen the 2027 earnings story, the recent 2–4% pullback could look more like profit-taking than a change in trend. But with expectations this elevated, a good quarter paired with merely cautious guidance could still trigger “sell the news”. Rising Treasury yields are another wildcard because they can keep compressing multipl

For me, the 10-year Treasury at 5.1% is already a level where I start redoing the equity math. At 5.25%, I would become more selective on high-valuation growth stocks, because the risk-free alternative becomes increasingly attractive and higher discount rates put more pressure on valuations. If the 10-year approaches 5.5% without a corresponding acceleration in earnings growth, I would favour stronger cash flows, reasonable valuations and balance-sheet quality rather than chasing momentum. The key question is why yields are rising. Strong economic growth accompanied by stronger earnings is much easier for equities to absorb than a rise driven mainly by persistent inflation, Fed tightening expectations and Treasury supply. So my levels are: 5.25% = reassess; 5.5% = much higher hurdle for eq

For the memory trade to stop needing the next price rise, I would want Micron to show that earnings growth can increasingly come from volume, product mix and structural AI demand rather than just higher DRAM and NAND prices. Strong HBM4 shipments, sustained data-centre demand, firm long-term customer commitments and continued margin strength even as pricing growth moderates would be especially important. If Micron can demonstrate that AI-driven memory demand remains ahead of supply into 2027 while volumes expand, the thesis becomes much more durable than simply betting on another pricing cycle.

B. Cybersecurity & Data Resilience, with CRWD on my watchlist. AI compute is still driving a big part of the rally, but I find cybersecurity more interesting at these levels. More AI agents and automated workflows also mean a larger attack surface, making security increasingly important alongside chips and compute. CRWD is my watchlist pick because it offers exposure to this second-order effect of AI adoption. This week's fresh highs also seem broader than just one theme, with AI, cybersecurity and energy all participating, which is healthier than a rally driven solely by a handful of chip stocks.

B. Cybersecurity & Data Resilience, with CRWD on my watchlist. AI compute is still driving a big part of the rally, but I find cybersecurity more interesting at these levels. More AI agents and automated workflows also mean a larger attack surface, making security increasingly important alongside chips and compute. CRWD is my watchlist pick because it offers exposure to this second-order effect of AI adoption. This week's fresh highs also seem broader than just one theme, with AI, cybersecurity and energy all participating, which is healthier than a rally driven solely by a handful of chip stocks.

B. Cybersecurity & Data Resilience, with CRWD on my watchlist. AI compute is still driving a big part of the rally, but I find cybersecurity more interesting at these levels. More AI agents and automated workflows also mean a larger attack surface, making security increasingly important alongside chips and compute. CRWD is my watchlist pick because it offers exposure to this second-order effect of AI adoption. This week's fresh highs also seem broader than just one theme, with AI, cybersecurity and energy all participating, which is healthier than a rally driven solely by a handful of chip stocks.

I see it mainly as locking in profit, even if it sacrifices some upside. Memory has always been highly cyclical, so securing customers for next year's capacity reduces the risk of being caught with excess supply when the cycle eventually turns. With demand currently so strong, selling too much capacity early could leave money on the table if prices keep rising, but I'd rather see manufacturers protect margins and cash flow than maximise every dollar at the top of the cycle.

$Meta Platforms, Inc.(META)$  I think it makes sense for integrators to rally first. Shopify can show a clearer near-term benefit from Muse through more transactions and Shop Pay usage, while Meta still has to prove how Muse itself will monetise its rapidly growing user base. The partner rallies are encouraging because they suggest Muse is developing into an ecosystem rather than just another AI app. The next step is showing that Meta can capture a meaningful share of the value created by that ecosystem.

I trust the Nasdaq trend, but with caution. The AI and semiconductor trade clearly still has momentum, but a record index alongside a falling Dow and weak market breadth suggests the rally is becoming increasingly concentrated. I would be more confident if the next move brings broader participation from financials, industrials and other sectors. For now, I would stay with the AI trend but avoid chasing it aggressively.

A. Buy AI infrastructure. AI agents may change which apps and services win, but every successful agent still needs compute, memory and data-centre capacity. I prefer owning the infrastructure that benefits from rising AI usage regardless of which application ultimately dominates. The main risk is valuation, so I would still be selective rather than chase every AI infrastructure stock.

A. Buy AI infrastructure. AI agents may change which apps and services win, but every successful agent still needs compute, memory and data-centre capacity. I prefer owning the infrastructure that benefits from rising AI usage regardless of which application ultimately dominates. The main risk is valuation, so I would still be selective rather than chase every AI infrastructure stock.

B. Wait for confirmation. High-beta names can move fast, but one strong session is not enough for me to chase. I would rather see the momentum hold for several sessions, supported by stronger volume and broader participation beyond just AI chips and crypto-related stocks. If the rally broadens while yields and volatility remain supportive, I would be more comfortable adding exposure. Missing the first part of a rally is preferable to chasing a short-lived spike.

B. Wait for confirmation. High-beta names can move fast, but one strong session is not enough for me to chase. I would rather see the momentum hold for several sessions, supported by stronger volume and broader participation beyond just AI chips and crypto-related stocks. If the rally broadens while yields and volatility remain supportive, I would be more comfortable adding exposure. Missing the first part of a rally is preferable to chasing a short-lived spike.

A flat tape on triple witching does not tell me much by itself because expiry-related positioning, hedging and rebalancing can distort the index move. What stands out more is the divergence underneath. Semiconductors gained nearly 3% while several megacaps fell, yet QQQ still finished higher. That suggests capital was rotating within tech rather than simply leaving the sector. I would watch the next few normal trading sessions. If semiconductors continue to outperform after the expiry effects fade, the rotation may be more meaningful. If the divergence quickly reverses, Friday was probably more about positioning and expiry mechanics than a lasting market signal.