I read it as tone improvement, not policy shift. At best, the visit delivers: supply chain stability language clearer licensing for “China-compliant” chips from NVIDIA But not: relaxation on leading-edge AI exports meaningful reopening of China revenue at scale So this is friction reduction, not a chip policy win. For NVDA: +2.2% pre-market + $315 PTs = expectations already elevated Earnings risk is asymmetric: strong beat → limited upside; any miss → sharp pullback “Buy the rumour, sell the news” is plausible. A true second leg higher needs actual policy text change, not delegation optics. Base case: Trip: positive headlines, limited substance Earnings: strong but high bar Positioning: avoid chasing, manage risk into event

The market reaction is telling you something important: symbolism has been priced in, policy change has not. NVIDIA having Jensen Huang at the table matters, but only at the margin. It shifts the agenda, not the constraints. Here is how I would frame it. What his presence can realistically deliver It elevates AI chips from a regulatory issue to a trade negotiation variable. That alone is progress. It increases the probability of incremental concessions: slightly relaxed thresholds, licensing clarity, or tacit tolerance for “China-compliant” SKUs. It gives policymakers better visibility into industry consequences, especially on supply chains and US firms’ competitiveness. What it is unlikely to change Core export controls are driven by national security, not trade balance. That sits above a

I think the market is treating Alibaba Group less like an e-commerce company now and more like a “China AI infrastructure + sovereign cloud” proxy. That is the real rerating driver. The bullish case is not hard to understand: Cloud revenue +38% AI-related revenue still growing triple digits Management saying AI products are already ~30% of external cloud revenue and could exceed 50% within a year “No GPU sits idle” implies utilisation is extremely high, which matters because idle GPUs destroy ROIC in AI infrastructure businesses  But the market is also glossing over something important: operational profitability is ugly right now. Adjusted net income collapsing toward near-zero while capex explodes tells you Alibaba is still in the “build first, monetise later” phase.  The key qu

If you strip away the optics, the sequencing usually follows what can be signed quickly versus what requires regulatory clearance or political capital. 1) BA / GE – most immediate (highest probability) Aircraft orders are the cleanest “headline deliverable”. China can announce bulk orders for Boeing with engines tied to General Electric (GE Aerospace). These deals are politically symbolic, commercially straightforward, and have precedent during state visits. Expect this first, possibly even during the visit. 2) MU / ILMN – medium-term (policy signalling first, fundamentals later) Micron Technology easing is plausible as a goodwill gesture. But actual earnings impact depends on procurement recovery, which takes quarters. For Illumina, any thaw is slower. Genomics sits closer to national sec

Micron Technology (MU) still has a strong structural bull case. HBM is effectively sold out into 2026–2027, hyperscaler AI capex remains aggressive, and memory is shifting from “commodity” toward strategic infrastructure. That is why analysts are talking about a “virtuous cycle” instead of a normal DRAM boom-bust phase.  But after a near-parabolic move toward $800, risk/reward is no longer clean. Options markets are already pricing large swings, and psychologically round numbers often attract profit-taking.  Personally: Long-term investor → scaling in slowly still makes sense. Short-term trader → chasing here feels late unless momentum stays extreme. $770 support is more attractive because it gives better downside control while still respecting the AI memory thesis. If MU loses t

1. Middle East demand helps yields, especially premium and cargo, but cannot fully offset oil at US$120. Fuel remains the dominant cost driver, so margins likely compress despite stronger traffic. 2. FY net profit likely down YoY. Demand is resilient, but higher fuel costs plus Air India losses and softer interest income are key drags. 3. Air India is a long-term strategic asset, not a near-term earnings driver. It gives exposure to India’s structural growth, but is currently dilutive with execution risk. I would price it as a long-duration option, valuable if turnaround succeeds, but a balance sheet drag for now.

RKLB is not quite the “SanDisk of space”. The bottleneck narrative is similar, but SanDisk had proven cash flows. Rocket Lab USA, Inc. (RKLB) is still scaling. At ~US$45–50B vs ~US$600M revenue, valuation already prices in strong execution. It is one of the few credible space players, but upside now depends heavily on Neutron and defence growth. AST SpaceMobile, Inc. (ASTS) after -10% is not automatically cheap. It remains pre-scale, with delays and weak earnings. Treat it as venture-style. Only buy small if you believe in long-term execution. Terafab looks like a Musk-style moonshot. Strategically logical, but extremely capital intensive. Likely cash-burning near term, with uncertain timelines.

Watching: $NVDA$ still leads the AI infrastructure trade, but expectations are now dangerously high. One weak guidance line could shake the whole sector. $MU$ remains one of my favourites because HBM demand still looks structurally supply-constrained, though memory cycles can reverse brutally. $CRCL$ is interesting if stablecoin regulation truly goes mainstream, but momentum is getting crowded fast. Avoiding: Overextended “AI story” stocks with weak actual earnings leverage. Narrative alone is no longer enough. Positioning: Not all-in, not all-cash. This still looks like a liquidity-driven AI bull market, but chasing vertical candles after euphoric runs usually ends badly. I’d rather hold quality names, trim into spikes, and keep dry powder for pullbacks.

Circle Internet Group has evolved from a “crypto proxy” into a regulated digital dollar infrastructure story. The market is no longer valuing CRCL purely on current earnings. It is pricing in the possibility that USDC becomes part of mainstream financial plumbing. The bullish thesis strengthened materially after Q1: USDC circulation grew 28% YoY to $77B On-chain transaction volume surged 263% to $21.5T Reserve income still rose despite lower yields Circle is expanding beyond reserve income into payments, AI-agent infrastructure, and tokenised finance services  The CLARITY Act matters because it removes a major regulatory overhang. If formally enacted, institutions that previously avoided stablecoins for compliance reasons may finally participate at scale. That could justify another va

Lumentum is no longer just an “index inclusion trade.” The Nasdaq-100 addition accelerated flows, but the real debate now is whether the company deserves to remain valued as a structural AI infrastructure winner after passive buying fades. The bullish case is still fundamentally credible: AI clusters are increasingly bandwidth-constrained, pushing hyperscalers toward optical interconnects and photonics solutions.  Lumentum’s recent results showed ~90% YoY revenue growth with expanding operating margins, suggesting this is not purely speculative narrative inflation.  NVIDIA’s strategic investment and partnership materially strengthened the market’s confidence that LITE sits inside the next-generation AI networking stack.  But the stock is now entering a harder phase: Index in

Micron Technology is no longer trading like a traditional cyclical memory stock. The market is increasingly pricing it as a core AI infrastructure supplier because HBM demand for AI accelerators remains supply-constrained, with reports that parts of Micron’s HBM capacity are already effectively booked ahead.  That said, chasing near psychologically important levels like $800 carries elevated volatility risk. Options markets are implying very large near-term swings, roughly ±8-9% this week alone.  My view: Long-term bullish thesis: still intact Near-term risk/reward after a vertical move: less attractive Best setup: partial entry now + add on pullback Key levels: $800 to $820: breakout momentum zone, but also profit-taking territory ~$770: first meaningful support ~$727 to $730: d

I lean infrastructure-heavy but balanced overall. Most upside from here may still be in Advanced Micro Devices and Micron Technology because the market is repricing the actual bottlenecks of AI scaling: HBM memory, advanced packaging, networking, storage throughput, not just GPUs alone. NVIDIA remains dominant, but expectations are already enormous. Meanwhile, SanDisk could still have strong upside if AI storage demand becomes structurally persistent rather than cyclical. I do not think this rotation is just a short-term blip. Markets are shifting from “who has AI exposure?” to “who controls constrained infrastructure capacity?” That said, after such violent rallies, risk management matters more: trim parabolic moves, keep core winners, avoid low-quality AI hype names. Hardware still look

Lumentum Holdings joining the NASDAQ-100 does create a genuine short-term technical catalyst because passive ETFs and benchmark-tracking funds are effectively forced buyers. That can extend momentum into and shortly after inclusion. But historically, index inclusion rallies often become “buy the rumour, sell the news” events once passive positioning is completed. If valuation stretches too far ahead of fundamentals, post-inclusion volatility can arrive quickly. Meanwhile, Applied Optoelectronics showed the market is no longer rewarding “AI exposure” alone. Expectations in optical networking have become extremely demanding: Investors now expect visible hyperscaler orders, sustainable margins, and strong forward guidance simultaneously. Within opticals, I would separate the sector into two g

I do not think the entire H2 upside has been fully priced in yet, but expectations are now extremely elevated. Micron Technology and SanDisk are benefiting from something larger than a normal memory rebound: HBM demand tied to AI accelerators remains supply constrained. AI servers consume far more DRAM and NAND per rack than traditional servers. Hyperscaler capex has shifted from “testing AI” to infrastructure scaling. That is why markets are willing to pay higher multiples versus past memory cycles. Still, the market is beginning to price in a “perfect scenario”: sustained HBM shortages, disciplined supply growth, and continued hyperscaler spending into 2027. The biggest risk is exactly what you highlighted. If Samsung Electronics or SK Hynix aggressively expand HBM/DRAM capacity faster t

I would lean toward staying invested, but trimming selectively rather than going “all in” after six consecutive weekly gains. The structural AI thesis still looks intact: Hyperscaler AI capex remains enormous and multi-year in nature. Memory/HBM, networking, power, cooling and semiconductor supply chains are still capacity constrained. Earnings revisions for major AI beneficiaries are still trending upward rather than downward. That said, valuations are undeniably stretched. Historically, when the NASDAQ-100 enters upper percentile valuation zones after a rapid melt-up, markets become highly sensitive to: earnings misses, weaker forward guidance, capex ROI doubts, or macro surprises from rates/inflation. Next week matters because: Alibaba Group and Tencent test whether China AI/cloud deman

My take: Goldman vs hedge funds? Both can be right. Goldman Sachs is looking 6 to 12 months ahead on earnings and liquidity. Hedge funds are trimming because positioning is crowded after a vertical run. Strategic bull, tactical caution. AMD / ARM upside? Advanced Micro Devices still has runway if AI GPU share keeps rising, but valuation is no longer cheap. Arm Holdings has strong AI CPU tailwinds, though much optimism is already priced in. Upside remains, but multiples are stretched. If Iran deal + Fed cuts? Money likely rotates from crowded leaders like NVIDIA into: 1. small caps 2. financials 3. REITs 4. industrial cyclicals 5. software / AI infra second-derivative plays Next leg broadens, not narrows.

I think HBM still has upside, but easy money is no longer easy. Bull case, cycle not done AI accelerator demand remains structurally strong, led by NVIDIA, custom ASICs from hyperscalers, and broader enterprise AI adoption. HBM is not commodity DRAM. It is high-complexity, packaging-constrained, and qualification-heavy. Supply cannot ramp overnight. Beneficiaries remain clear: Micron Technology, SK hynix, Samsung Electronics, plus ecosystem names like SanDisk on broader memory repricing. Bear case, valuation is running ahead This week’s sharp move likely pulled forward part of 2H expectations. Once Samsung and SK hynix commit materially more wafer capacity, markets will start pricing the next oversupply phase before it arrives. Memory stocks historically peak when sentiment is strongest. M

Yes, but only if Intel executes flawlessly. This is potentially transformational, but not automatically so. Why this matters: 1. Validation effect If Apple trusts Intel Foundry, every major chip designer will pay attention. Apple is arguably the hardest customer on process, yield, power efficiency, and supply reliability. Passing Apple’s bar is equivalent to a global seal of approval.  2. Intel becomes more than a CPU company The strategic shift is from: > x86 chipmaker → advanced manufacturing platform That puts Intel into competition with: Taiwan Semiconductor Manufacturing Company in foundry Samsung Electronics in manufacturing scale Nvidia / Advanced Micro Devices in ecosystem influence 3. Financial model improves massively Apple volume means: fab utilisation rises fixed costs

My call: stay invested, but trim selectively. This rally is not purely speculative. It is backed by real capex, strong earnings breadth, and a sharp repricing in AI infrastructure names like Nvidia, Advanced Micro Devices, Micron Technology and Sandisk. The semiconductor index is up sharply, and earnings have broadly beaten expectations.  That said, six straight weekly gains + elevated valuation percentile = thinner margin of safety. Forward returns from here are likely more volatile. My positioning: Core compounders (60 to 70%): hold High-beta runners (20 to 30%): trim into strength Cash (10 to 20%): rebuild for pullbacks Next week matters. If Alibaba Group, Tencent and Cisco Systems confirm AI monetisation, networking demand, and enterprise spend, the bull case extends. If not, the

$CoreWeave, Inc.(CRWV)$ The AI infrastructure thesis still holds, but CoreWeave’s equity thesis has become more fragile. Why the stock dropped despite a beat: 1) Guidance missed, and that matters more than backward-looking revenue Q1 revenue was US$2.08B (+112% YoY) and backlog reached US$99.4B, both signalling massive demand. But Q2 guidance missed consensus, which told markets growth may be more uneven than “hyperdrive” headlines imply.  2) Scale is coming with brutal capital intensity Operating expenses doubled, net loss widened to US$740M, and interest expense was US$536M in one quarter. That is the hidden cost of building an AI utility at hyperscale.  3) Leverage risk is real If utilisation slips, pricing softens, or financing cost