$Circle Internet Corp.(CRCL)$   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

Memory stocks finally bounced. MU rose 2.97%, SanDisk 5.53% and SK Hynix 4.67% ahead of Micron's 21 October technology webinar. But MU is already indicated around $883 pre-market, down roughly 2.4%. So Wednesday's webinar arrives at an important moment. The market is asking whether the recent memory selloff was an overreaction, or the beginning of a genuine repricing of the cycle. My answer: HBM still looks structurally strong. Conventional memory is where I would watch the risk. 🔵 WHAT WOULD ACTUALLY MOVE MU? Micron is expected to discuss its multigenerational roadmap across DRAM, HBM and NAND, plus customer co-design and the memory requirements of next-generation AI infrastructure. I would ignore generic statements about AI demand. I want three things. 1. HBM roadmap Anything reinforcing

The market has finally given investors something we have not had much of this year: Better prices. But a stock falling 10% does not automatically make it 10% more attractive. My rule for this dip is simple: Buy companies where the price has deteriorated faster than the business. Right now, my shopping list is MU, NVDA, META, CRCL and QQQ, but at very different levels. 🥇 MU: MY HIGHEST-CONVICTION DIP BUY Micron has been hit by the combination of memory-cycle fears, CXMT concerns and the broader technology selloff. But the question is whether HBM demand and earnings power have deteriorated as quickly as the share price. I do not think they have. 🚦 My levels 🟡 $860-$890: Starter position 🟢 $800-$840: Preferred Pick 🟢 $740-$790: Strong Pick 🔴 Below $740: Reassess the thesis If MU reaches $800-

QQQ fell 2.50% Friday, while the S&P 500 lost 1.11%. The obvious explanation is hotter inflation plus Fed uncertainty. But I think the more useful interpretation is this: The market is not pricing a guaranteed hike. It is repricing the probability that rates stay restrictive for longer. That distinction matters enormously for what I would buy next. 🔵 WHY WARSH'S SILENCE MATTERS Warsh declined to pre-commit ahead of the 27-28 October meeting. That is not automatically hawkish. But markets dislike uncertainty, particularly when valuations are elevated. Hotter CPI and PPI have already weakened the case for easier policy. If bond markets begin pricing another hike, long-duration technology gets hit first because its valuation depends heavily on future earnings. That explains why QQQ fell m

After the recent tech selloff, this earnings week arrives at exactly the right moment. SpaceX reports Tuesday, followed by Meta and Tesla. The market already knows these companies can spend enormous amounts of money. Now it wants evidence that those billions are producing returns. My ranking going into earnings: 🥇 META: strongest proof of AI monetisation 🥈 SpaceX: highest upside surprise potential 🥉 TSLA: highest execution risk 🔵 SPACEX: THE FIRST TEST At roughly $145, SpaceX goes first. Consensus is around $12.96B revenue and $0.139 EPS, but those headline numbers are not what I would trade. I am watching: Starlink growth → launch economics → margins → AI infrastructure opportunity. The bull case becomes much stronger if Starlink growth remains powerful while operating leverage improves.

CoreWeave fell 11.91% Friday to roughly $63, and I think investors focusing only on GPU demand are looking at the wrong variable. Demand is not CoreWeave's immediate problem. CoreWeave reported a $104 billion revenue backlog at the end of Q2, before more than $25 billion of additional early-Q3 commitments. Recent capacity has also been contracted at higher prices. The problem is what it costs to finance all that growth. CoreWeave bonds that were issued around a 9.7% yield in June recently traded near 15%. That changes the equation. AI demand can remain spectacular while the equity performs terribly if the cost of funding the infrastructure rises faster than its economics improve. 🔴 THE NUMBER I CARE ABOUT IS NOT GPU ORDERS CoreWeave needs enormous capital before much of its contracted reve

Apple is entering a category it has avoided for years: foldables. At $1,999, the iPhone Duo does not need to become Apple's highest-volume iPhone to succeed. It needs to do three things: Raise ASP → pull premium users into another upgrade cycle → deepen the Apple ecosystem. That is why I think investors watching only preorder numbers may miss the bigger story. 🔵 WHAT I AM WATCHING FIRST: DUO ORDERS Apple closed around $342 after initially responding positively to the launch. For the next few weeks, the first question is simple: Does a $1,999 iPhone sell without heavy discounting? A strong launch would demonstrate something increasingly valuable in a mature smartphone market: Apple can still manufacture an upgrade cycle through form factor, not merely faster processors and better cameras. B

$Rezolve AI(RZLV)$   Rezolve AI closed at $1.75, down another 5.91%, after touching $1.74. This is getting frustrating for bulls. RZLV has produced partnerships, customer wins and an increasingly ambitious AI-commerce story. Yet every promising rebound seems to attract another wave of sellers. The five-day chart captures the problem perfectly: $1.63 → $1.88 → back to $1.75. For next week, I am no longer asking whether Rezolve can produce another good announcement. I am asking: When will the market finally pay for the story? 🔴 THE PROBLEM IS PRICE ACTION RZLV bounced aggressively from around $1.63-$1.65 on 15 October. That was encouraging. But it could not hold the $1.85-$1.90 region. Thursday then erased much of the recovery. That tells me tw

$Beyond Meat, Inc.(BYND)$   Beyond Meat closed at $6.85, down 4.46%, after trading between $6.71 and $7.19. The five-day chart tells the story clearly. BYND fell from around $7.75, found buyers near $6.60, bounced sharply toward $7.20, then gave most of that rebound back. So I am changing my framework slightly. This is no longer a stock I want to predict purely from fundamentals. It is a high-volatility trading vehicle, and next week I want price to prove itself first. 🔴 WHAT WORRIES ME The failed rebound around $7.20 matters. BYND has now shown us that sellers are still willing to appear quickly whenever the stock rallies. More importantly, Beyond Meat still has the fundamental problems that make this trade dan

$Micron Technology(MU)$   Micron closed Friday at $878.84, down another 5.63%. From the five-day chart, MU has fallen from above $1,030 to below $880. This is no longer a routine dip. It is a full repricing of the memory trade. But there is an important contradiction: the stock is pricing a substantial deterioration in the memory cycle while the evidence that Micron's AI-memory economics have actually collapsed is still much weaker. That makes next week very interesting. 🔴 WHY MU KEEPS FALLING Three fears have collided. First, CXMT. Investors are worried its progress in server DRAM eventually introduces enough supply to weaken pricing power. Second, the broader AI trade has suffered a confidence shock. Nasdaq fell