Shyon

I’m watching the US-Canada tariff escalation closely. A 50% tariff on another US$20B of Canadian goods, with Ottawa threatening dollar-for-dollar retaliation, could add fresh inflation and growth concerns if Tuesday’s deadline passes without a deal. For now, it still feels partly like tariff theater, but the risk of a broader trade war is becoming harder to ignore.

On the other hand, Bitcoin breaking above US$75,000 after its strongest week in years is definitely catching my attention. I’m also keeping an eye on gold, oil and AI names this week, especially with Nvidia earnings and Jackson Hole ahead. For me, the key is staying selective and managing risk rather than chasing every headline.

I think $Walmart(WMT.US)Walmart’s 9% drop is a useful warning that the US consumer is becoming more selective, rather than simply collapsing. A 2.6% comparable-sales growth rate, despite a strong earnings beat and raised guidance, suggests that spending momentum is cooling, especially when the profit beat was partly helped by a one-off tariff refund. I’m watching Costco, Home Depot and Amazon closely because weakness across these names could signal a broader normalization in consumer demand rather than an isolated Walmart issue.

For me, the bigger picture remains mixed rather than outright bearish. Long-end Treasury yields moving back up show that investors are still demanding a higher risk premium, while AI-related memory and optical names such as Micron, SK Hynix and Lumentum continue to show strong underlying demand. Alibaba is another good example: AI/cloud growth is accelerating, but heavy capex is compressing near-term profits. I’m therefore staying patient and selective, especially ahead of NVIDIA earnings and Jackson Hole, because the next catalyst could determine whether this pullback is simply a healthy correction or the start of a deeper valuation reset.

$MU 2X Long ETF(MUU.US)

My MUU position has been quite a journey. I started accumulating when the semiconductor and memory trade was under pressure, and for a while, the position was sitting firmly in the red. Then the market turned, $Micron Tech(MU.US) rallied aggressively, and MUU moved with it. My losses eventually turned into meaningful profits, proving once again why I was willing to stay patient with this trade.

But the story has changed again in the short term. After taking some profits, I continued holding and averaging into MUU, and the latest correction has pushed my position back into a minor loss. Honestly, I am not too concerned. I see this as another phase of volatility rather than a deterioration in the fundamental story.

What keeps me patient is the underlying strength of Micron. The company is no longer simply a traditional memory-cycle story. AI is fundamentally changing the amount and quality of memory required per accelerator and data center.

The HBM story is particularly important. Micron has previously indicated that its entire 2026 HBM production is sold out. At the same time, industry data continues to point toward strong AI-driven DRAM demand, with Micron maintaining a significant position in the global memory market.

That is why I am not treating this correction as a reason to abandon the position. The stock can absolutely fall further, and MUU carries considerably more risk than owning MU directly because of its 2x daily leverage. But if Micron's earnings growth, HBM demand, memory pricing and AI infrastructure spending remain strong, the fundamental backdrop can continue supporting another major move higher.

After experiencing MUU go from losses to profits and now back to a relatively small loss, I remain patient. I am watching the fundamentals more closely than the daily price action. If this correction is indeed approaching its final stages, I believe another memory and AI-driven rally could eventually put MUU back on the offensive.

I’d be very cautious chasing Moderna after a 177% one-day surge. The Phase 3 result is genuinely significant and could reshape the long-term cancer treatment market, but the stock has already priced in a huge amount of optimism. I’d rather wait for the initial excitement to cool and see how the commercial opportunity develops.

What interests me more is the semiconductor split. I’m still constructive on memory names like MU and SNDK given the tight AI-memory supply, but the weakness in WDC, COHR and NBIS shows how quickly funding, valuation and profit-taking can change sentiment. For me, this is a market where fundamentals matter more than simply chasing the biggest one-day winner.

I see this semiconductor selloff as a healthy reset, not the end of the AI cycle. Higher Treasury yields are pressuring high-growth and capital-intensive names like $Micron Tech(MU.US), $Sandisk(SNDK.US), $Coreweave(CRWV.US) and $Nebius(NBIS.US), but AI infrastructure and memory demand remain strong.

I’m using the pullback to selectively add rather than panic sell. Memory and semiconductor names remain attractive to me for the long-term AI demand story, although I’m keeping some cash ready in case rates trigger more volatility.

I’ll be watching the Fed minutes closely tonight. If the tone turns less hawkish, this correction could reverse quickly. For me, pullbacks are opportunities to accumulate quality AI exposure, not reasons to abandon the thesis. 📈

I think the memory rally is more than just a one-day headline trade. Musk calling memory the binding constraint for AI reinforces what we’re already seeing: AI infrastructure demand is increasingly running into tight memory and storage supply. That’s why I remain bullish on $Micron Tech(MU.US), $Sandisk(SNDK.US) and $Western Digital(WDC.US), even with the broader market under pressure.

Druckenmiller exiting Micron right before this move is definitely ironic, but I wouldn’t use one fund’s timing as a reason to change my thesis. Markets can move very differently from institutional positioning in the short term, and the stronger signal to me is still the supply-demand imbalance in memory. If demand continues outpacing supply, these companies could still have room to make new highs.

I’m also watching Hong Kong and Singapore closely because the divergence from Wall Street is interesting. With FOMC minutes, Walmart earnings and Jackson Hole ahead, volatility could stay elevated. For me, the strategy remains simple: stay selective, keep accumulating quality AI beneficiaries on pullbacks, and let the long-term semiconductor cycle play out. 🚀

$Direxion Semicon Bull 3X(SOXL.US)

I’m continuing to average up my position in SOXL as the semiconductor sector recovers from the sharp pullback a few weeks ago. The correction created plenty of fear around chip stocks, but I see it as a healthy reset rather than the end of the semiconductor cycle. With AI infrastructure demand, memory strength and continued investment in data centres supporting the broader sector, I believe the long-term growth story remains intact.

For me, the recent rebound shows why I was comfortable accumulating during the weakness. SOXL offers leveraged exposure to semiconductors, so the volatility is much higher, but that also makes pullbacks extremely attractive when I believe the underlying sector is recovering. I’m not trying to catch the exact bottom—I’m building my position progressively and taking advantage of the opportunities the market gives me.

I’m especially glad to see my current SOXL position finally turn from losses into profits after the recent recovery. 🚀 It’s a good reminder that staying disciplined during a sharp pullback can pay off when the underlying trend turns around. I still believe there is more room for the semiconductor sector to run, so I’ll continue to average up selectively and let the recovery play out. More to go! 📈

I think Nvidia’s move is more about discipline than the end of the AI story. Cutting the OpenAI financing guarantee while still investing heavily in SpaceX suggests Nvidia is becoming more selective about where it puts capital, especially as AI infrastructure spending gets increasingly debt-driven.

What stands out to me is that AI demand itself remains strong. Singapore’s 24.2% export growth, driven by a 112% surge in electronics shipments, is another sign that the AI and semiconductor cycle is still feeding through to the real economy. I’d see the recent financing concerns as a reminder to differentiate between genuine AI demand and overly aggressive funding structures.

For me, the bigger opportunity is still in companies with strong fundamentals and direct exposure to AI demand, rather than simply chasing every AI-related name. I’m staying bullish, but I’m also becoming more selective—especially after the huge run-up in semiconductor and AI stocks this year.

Reddit joining the S&P 500 is definitely interesting. I think index inclusion could bring more institutional demand and liquidity, but with RDDT already running strongly, I wouldn’t chase the rally blindly. The meme-stock DNA is still there, so volatility will probably remain high.

In Singapore, I’m paying more attention to the strong earnings from City Developments, CapitaLand Investment and OCBC. CDL’s profit jump is especially impressive, while CLI’s record fee income shows that Singapore’s property and asset-management sectors are still benefiting from a healthier environment. I’m increasingly bullish on selected Singapore blue chips after the STI’s strong run this year.

For US tech, SanDisk’s long-term growth target and the memory-sector momentum caught my attention, especially given my existing interest in the AI infrastructure and memory theme. At the same time, Cisco’s margin pressure is a reminder that AI demand is strong but rising component costs can squeeze profitability. Overall, I’m staying selective rather than chasing every stock that pops after earnings.

I think the latest earnings are a strong validation that AI infrastructure demand is still very real. CoreWeave, Super Micro, Nebius and Lumentum all delivered strong numbers, showing that spending on compute, networking and optical infrastructure is translating into actual revenue growth. To me, this looks more like a fundamental growth cycle than just hype—although valuations are clearly getting stretched.

The interesting contrast is China, where Tencent’s surge in AI capex triggered concerns about cash flow and spending sustainability. That tells me the market is increasingly differentiating between AI companies that can monetize infrastructure demand and those still in the heavy investment phase. Meanwhile, the rising 10-year yield remains a key risk for high-growth valuations.

I’m still bullish on the AI-infrastructure theme, but I wouldn’t chase vertical moves blindly. After such powerful rallies, pullbacks are inevitable, and I’d rather use them to accumulate quality names with strong earnings, backlog and long-term demand visibility. For me, the thesis is intact—the key question now is whether earnings can continue growing fast enough to justify these valuations.

I see Alphabet’s pullback as more about AI spending concerns than weak fundamentals. Q2 revenue grew 24% and Cloud jumped 82%, but with 2026 CapEx potentially reaching $205B, investors want clearer proof that these investments will translate into stronger cash flow and returns.

Meanwhile, CRWV, SMCI and LITE showed that AI infrastructure demand remains strong, with CRWV’s backlog above $100B and LITE saying laser demand far exceeds capacity. So I’m not ready to call the AI trade over, but valuations are clearly facing higher expectations.

I’ll be watching CPI and upcoming earnings closely. From here, beats alone may not be enough — orders, margins and guidance will determine who can keep rallying. I still see room for the AI trade, but the market will increasingly reward companies that can prove real returns from AI spending.