This is the part of earnings season I love. Within roughly 48 hours, we get results from: Microsoft. Alphabet. Amazon. Intel. Strategy. Coinbase. SanDisk. But I would make a mistake if I treated all seven the same. Some I would be comfortable holding. Some I would deliberately reduce. And for some, I would rather own the stock that reacts to their earnings than gamble on the reporter itself. My rule: The bigger the information gap, the smaller my pre-earnings position. Here is how I am approaching them. 🟢 1. GOOGL: MY FAVOURITE MEGA-CAP ASYMMETRY Alphabet enters tonight's report from a very different position from Microsoft. GOOG has already corrected from roughly $338.50 toward $314-$315. That matters. Expectations have come down. I want three things: Search resilience Google Cloud accele

Intel reports Thursday after the bell. The stock has already rallied sharply, closing around $93.74 Tuesday and trading around $95 pre-market. The new Classiq acquisition gives investors another exciting narrative: Intel + quantum computing. But I would not buy Intel at $95 because of quantum. Quantum may matter enormously one day. Thursday's earnings are about something much less glamorous and much more important: Can Intel turn booming AI server demand into enough profit to finance its manufacturing comeback? That is the real turnaround test. 🔵 INTEL NOW HAS TWO COMPANIES INSIDE ONE STOCK This is how I think about Intel today. Company 1: The improving semiconductor business Data-centre CPUs. AI inference. Server demand. Client processors. Pricing. Product margins. This side can generate

Earnings season always creates the same dilemma: Hold overnight and risk the gap? Or sell beforehand and risk watching the stock explode without you? I think that is the wrong question. On Nov 4, two stocks on my radar report after the bell: Circle (CRCL) Beyond Meat (BYND) Both can move violently. Both have passionate bulls and bears. But I would not manage them remotely the same way. Why? Because overnight risk should not be determined by how exciting a stock is. It should be determined by: How much of the thesis can earnings actually prove? 🟢 CRCL: I WANT EXPOSURE, BUT NOT MAXIMUM EXPOSURE Circle's earnings are much bigger than whether EPS beats consensus. The market needs answers to three questions. 1. Is USDC still expanding? This is the engine. More USDC circulation means a larger re

Monday looked reassuring. Nasdaq +0.82% S&P 500 +0.70% Dow +0.28% After two difficult weeks for technology, AI buyers returned. But I would not interpret one green session as an all-clear. This week has an unusual setup: Fed + Microsoft + Alphabet + Amazon + memory earnings That means Tuesday is less about chasing what already moved and more about identifying where Wednesday and Thursday could create the next asymmetric trade. Here are the five pressure points I am watching. 1. MICROSOFT: THE AI ROI TEST MSFT enters earnings around $529-$533, close to recent highs. That makes the hurdle high. I want to see: Azure acceleration AI revenue growth Capex discipline Free cash flow Microsoft does not need to prove AI demand exists. It needs to prove the billions spent satisfying that demand a

$Alphabet(GOOG)$   Alphabet and Microsoft report Wednesday after the bell. On paper, investors will compare the usual numbers: Google Cloud vs Azure. AI growth. Capex. Margins. Free cash flow. But I think the more interesting trade is hiding in the charts. Alphabet has fallen from roughly $338.50 to $314, about a 7% reset. Microsoft is around $529-$533, close to the upper end of its recent range. That means these two companies enter the same AI exam with very different expectations already embedded in their stocks. And that matters. 🔵 MICROSOFT HAS TO PROVE THE PREMIUM Microsoft's AI story is arguably cleaner. Azure gives investors a relatively direct way to measure whether enormous AI infrastructure spending is

This may be the most important mega-cap earnings week of the quarter. Microsoft. Alphabet. Amazon. Intel. Together, they sit at the centre of an extraordinary AI capital cycle. But after the market's recent reaction to Meta, I think the question has changed. It is no longer: “Who is spending the most on AI?” It is: “Who is getting paid back fastest?” That distinction could decide not only which mega-cap wins earnings week, but whether the broader AI trade gets its next leg higher. 🔵 AI HAS ENTERED PHASE 2 Phase 1 was about building. Buy GPUs. Build data centres. Secure power. Train bigger models. Spend aggressively. For a while, announcing larger AI capex was itself bullish because it signalled confidence in future demand. But eventually every capex cycle faces the same test: Where is the

Memory is waking up again. SK Hynix +4.05% Friday. Micron +2.46%. And Monday pre-market: MU ~$901 SNDK ~$1,267 But this week's earnings are about something much bigger than which memory stock beats estimates. For months, the memory bull case has rested heavily on one product: HBM. AI accelerators need enormous amounts of high-bandwidth memory, supply remains constrained, and SK Hynix and Micron have benefited enormously. But for the next leg of the memory trade to become broader and more durable, I want evidence that strength is spreading. HBM → DRAM → enterprise SSD → NAND → storage This week may finally tell us whether that is happening. 🟢 SK HYNIX: THE HBM BENCHMARK SK Hynix reports first. I care less about whether EPS beats consensus and more about three questions: 1. Is HBM demand sti

This is one of the strangest biotech reactions of the week. Moderna's personalised cancer vaccine, intismeran, combined with Merck's Keytruda showed in more than 1,100 melanoma patients: 42% lower risk of recurrence or death 38% lower risk of distant metastasis or death No new safety signals Those sound like excellent Phase 3 results. Yet MRNA collapsed roughly 17.6%, briefly touching $183 before recovering toward $200. So what went wrong? I think the answer is simple: The science beat the control arm. It did not beat the market's expectations. 🟢 WHAT THE TRIAL ACTUALLY PROVED The biggest positive is not merely the 42% number. Moderna has now taken personalised mRNA cancer therapy through a large Phase 3 study and demonstrated meaningful clinical benefit. That matters because the entire po

Bloom Energy closed at $193.30, down 11.24%, despite numbers many companies would love to report: Revenue: $1.24B, +138% Non-GAAP EPS: $0.94, above expectations FY revenue guidance: $4.4B-$4.6B Gross-margin guidance: 34.5% So why is BE falling? Because at roughly 250x trailing earnings, investors are no longer asking: “Can Bloom grow?” They are asking: “When does extraordinary growth become extraordinary cash flow?” 🟢 WHY I AM NOT BEARISH ON THE BUSINESS Bloom sits directly inside one of AI's biggest bottlenecks: POWER. AI data centres need enormous amounts of reliable electricity, and grid connections can take years. Bloom's fuel-cell systems offer something extremely valuable: power that can potentially be deployed faster and closer to where data centres actually need it. The AI infrastr

$Circle Internet Corp.(CRCL)$   Circle closed at $85.91, but the headline number hides something interesting. CRCL has spent most of October absorbing repeated selling without losing the $80 area. It fell from nearly $100. It briefly flushed to $75.27. It recovered. And now it is compressing around $85-$87 with RSI almost perfectly neutral. That does not look like a stock investors have abandoned. It looks like a stock waiting for its next catalyst. And there are only seven trading sessions before Circle reports Q3 earnings on Nov 4. My base case? CRCL tests $90-$92 next week, then makes an attempt at $95-$100 immediately before earnings. But there are two very different forces that will decide whether that happens. WEEK 1: OCT 26-30 The macr

Tech sold off. Meanwhile: PG +5.58% KO +2.87% LLY +2.83% WMT +2.82% COST +1.99% BRK.B +1.32% At first glance, the message looks obvious: AI out. Defensives in. I think that interpretation is too simple. The market is not suddenly abandoning technology to buy toothpaste, Coke and supermarkets. It may be doing something much more important: changing the price it is willing to pay for uncertainty. And that could affect how I position for the rest of 2026. 🔵 WHAT ACTUALLY CHANGED? For much of the AI rally, investors rewarded companies for what their earnings could become. AI infrastructure spending accelerated. Revenue expectations exploded. Multiples expanded. Future cash flows were pulled forward into today's stock prices. That works beautifully while expectations keep rising. But once valua

The market just gave us an interesting contradiction. The reported Tesla-SpaceX merger plan initially sent both names higher. Now the excitement is being tested hard: SpaceX: ~$145, down 9.1% Tesla: ~$375, down 3.1% That reversal matters. Investors initially saw the dream. Now they are starting to price the bill. And I think the real question is much bigger than whether Tesla shareholders get some SpaceX exposure. Would combining Tesla and SpaceX actually create value, or merely combine two enormous capital requirements under one roof? 🔵 THE BULL CASE IS MUCH BETTER THAN “TESLA GETS STARLINK” Think about what a combined Musk ecosystem could theoretically contain: SpaceX: launch + Starlink + satellite infrastructure Tesla: EVs + energy storage + autonomy xAI: models + AI compute Optimus: ph

$Micron Technology(MU)$   Micron has gone from $936 to about $846 in five trading days, including a 6.26% fall on Wednesday and another slide after hours. That is roughly a 10% correction in less than a week. The easy conclusion is that something must have broken inside Micron. I think that conclusion is premature. What has changed is the market's willingness to pay for the memory story. And right now, investors are mixing up three very different risks: conventional DRAM supply, AI infrastructure spending, and HBM demand. Separating them is where I think the opportunity lies. 🔴 WHY MU IS ACTUALLY FALLING There is no single smoking gun. First: CXMT has changed the scarcity narrative. The market had priced memory as

$Meta Platforms, Inc.(META)$   Meta delivered what investors wanted on the top line. Q3 revenue beat expectations, growth remained around 26% year over year, ad impressions increased 11%, ad pricing increased another 11%, and Q4 guidance came in slightly ahead of expectations. Yet META is indicated around $678 pre-market, down roughly 3%. Why? Because investors are no longer asking whether Meta's AI strategy is working. They are asking how much it will cost. 🟢 THE AI MONETISATION ENGINE IS WORKING This is the part I would not lose sight of. Meta has one of the clearest AI monetisation loops in the market: Better AI recommendations → more engagement → more ad impressions → better targeting → higher advertiser ret

Nvidia rose 2.54% after Jensen Huang outlined a much bigger AI economy spanning energy, compute, AI factories, models and applications. $NVIDIA(NVDA)$ But one part caught my attention: Telecom networks. The first AI infrastructure boom concentrated enormous computing power inside hyperscale data centers. The next phase could push intelligence outward, closer to where data is actually created and consumed. Phones. Factories. Robots. Autonomous machines. Enterprise networks. 5G infrastructure. If that happens, Nvidia's opportunity may become considerably larger than simply selling more GPUs to hyperscalers. 🔵 WHY TELECOM AI COULD MATTER Traditional cloud AI works largely like this: Device → Network → Data Center → AI processing → Net

Tonight's earnings arrive at an interesting moment. Tesla is around $353, while Meta is around $704. Both are AI stories, but the market is asking completely different questions. For Meta: Is AI generating enough advertising revenue to justify enormous capex? For Tesla: Can autonomy, Robotaxi and robotics eventually become large enough to justify a valuation that already looks far beyond a conventional automaker? For me, Tesla is the more fascinating setup because tonight's EPS is almost secondary to what management says about 2027. 🔵 TESLA: I AM WATCHING 4 THINGS 1. Automotive margins This is still the foundation. Consensus expectations around $0.45 EPS and $28B revenue matter, but I care more about whether automotive economics are stabilising. Tesla can talk about robots and autonomy all

The AI chain finally gave memory investors something more tangible than optimism. Super Micro raised Q1 FY2027 revenue guidance from $14.5-$15.5B to $18.5-$19.5B after hours. Even more important, gross-margin guidance increased from 10.4%-10.8% to 12.0%-12.5%. SMCI surged more than 12% after hours, while Nvidia gained nearly 2% and semiconductor exposure strengthened. That combination matters. More AI server demand eventually means more GPUs, more networking and more memory. So is this finally the floor for MU and SanDisk? I think it strengthens the case considerably, but there is one important catch. 🟢 WHY THE SMCI REVISION MATTERS Look beyond SMCI itself. A revenue midpoint moving from roughly $15B to $19B is a huge change in expected server activity. If AI infrastructure demand were gen

If I had to choose one ETF to DCA consistently, my answer would be QQQ. Not because it is the safest ETF. Not because it will outperform every year. And definitely not because I expect technology stocks to go up in a straight line. I choose it because I want my long-term money exposed to companies benefiting from some of the biggest structural trends of the next decade: AI, cloud computing, semiconductors, automation, digital advertising and software. And DCA solves one of my biggest problems as an investor: I do not need to predict every bottom correctly. 🔵 WHY QQQ OVER VOO OR SPY? VOO and SPY are excellent core ETFs. If someone prioritises diversification and smoother compounding, I would happily recommend starting there. But personally, I want more growth exposure. QQQ gives me heavier

The recent tech correction has been brutal. Memory stocks were crushed, high-beta AI names sold off, inflation revived higher-for-longer fears, and investors started questioning whether massive AI capex could actually generate equally massive returns. Then something changed. Semiconductors finally attracted buyers again, and some tech names spiked after hours following strong earnings. I do not think that proves the bottom is in. But it may be telling us which stocks the market wants to own if the correction is ending. 🟢 AFTER HOURS GAVE US A CLUE The important development is not simply that tech bounced. It is why certain stocks bounced while others did not. Companies showing strong AI-related revenue and credible earnings are being rewarded. Companies disappointing on forward expectation

Nvidia rebounded 1.38%, with Marvell +4.06%, AMD +1.76% and Broadcom +1.78% as attention shifts to Jensen Huang's 21 October GTC Berlin keynote. NVDA is around $217. But I do not think another impressive GPU announcement is enough to move the stock materially. The market already believes Nvidia has great chips. What it now wants is proof that AI demand is broadening beyond hyperscaler training into inference, agents, enterprise AI and physical AI. That is what could reignite the trade. 🔵 THE NUMBER BEHIND THE KEYNOTE I am watching for evidence that falling inference costs are creating more total demand, rather than simply making existing AI workloads cheaper. That is the next Nvidia flywheel: Better compute → cheaper inference → more AI applications → more tokens → more infrastructure dema