To The Moon

I would watch margins next quarter, while giving Alibaba a modest cloud re-rating. The bullish case is real: Cloud and Compute grew 45%, its strongest growth in 22 quarters, while cloud adjusted EBITA jumped 133% and margin expanded to about 12%. AI product revenue has also delivered triple-digit growth for 12 consecutive quarters.  But I would not fully re-rate BABA on cloud growth yet. The problem is capital intensity. Capex rose 75% to RMB67.7bn, while GAAP net profit fell roughly 75%. Management is effectively exchanging near-term earnings and free cash flow for future AI capacity.  The crucial question is therefore not whether AI demand exists. It clearly does. It is whether cloud revenue and margins can grow faster than AI infrastructure spending. My hierarchy: 1. Cloud gro

I would wait for Warsh’s tone, while keeping a core long-tech position rather than rotating aggressively into rate-sensitive assets yet. The key signal is that Treasury’s intervention only produced a temporary rally. The long end quickly returned to concerns over deficits, inflation and term premium. The 30-year yield has been around multi-decade highs, while the 10-year has remained near 4.7%.  My positioning: Core: Stay long quality tech. AI earnings and structural capex remain powerful, although high long-term yields are the main valuation risk. Nvidia earnings on 26 August could provide another catalyst.  Do not chase rate-sensitive assets yet. Banks, REITs, small caps and long-duration bonds could rally sharply if Warsh signals easier policy, but they could suffer if he emph

My pick: GOOG > PYPL > CRWD > PLTR. GOOG has the strongest risk/reward, combining Search cash flow, Cloud growth and major AI optionality through Gemini and infrastructure. AI may threaten Search, but Alphabet also controls much of the ecosystem needed to monetise AI. PYPL is the contrarian value play. If checkout stabilises and margins improve, upside could be meaningful, though it remains a turnaround. CRWD remains a great business, but valuation leaves less room for error. PLTR has phenomenal growth, but its valuation already prices in exceptional execution. I agree with the downgrade tactically, not necessarily fundamentally. My move: buy GOOG, consider PYPL, and wait for better entry points on CRWD/PLTR.

I would buy SK Hynix on weakness, rather than step away from memory. My preference is SK Hynix > Samsung > avoiding the sector. The key distinction is that SK Hynix's payout is not simply management saying, "we have run out of attractive investments". It is explicitly buying and cancelling 40 trillion won of shares, while raising its target to return more than 50% of 2025-27 cumulative FCF. That is a direct reduction in share count and a strong signal management believes the stock is undervalued.  Samsung is potentially even more interesting as a value + dividend play, but the >100 trillion won figure remains a media report awaiting board approval. The reported plan would allocate 50% of FCF to shareholders, with dividends expected to dominate.  I don't see the payouts a

I would wait for Warsh’s Jackson Hole tone before rotating aggressively back into tech. The Treasury intervention is meaningful, but I would not interpret it as a durable reversal in long-term yields. The 30-year yield had reached about 5.34%, its highest since 2007, before Treasury announced it would at least double long-duration buybacks to $4bn per operation.  The bigger issue is the Fed. July's minutes were more hawkish than the headline "hold" suggests: three officials wanted a 25bp hike, several saw inflation as broad-based, and there was no meaningful discussion supporting a cut. Markets are even assigning better-than-even odds to a hike by October or December.  So my positioning would be: Tech: cautiously add, not chase. Lower yields provide exactly the relief that high-d

I would not chase Moderna at $174.38. I would rank the three choices: 1. Merck: best risk/reward 2. Wait for full data: best disciplined approach 3. Moderna: highest upside, but highest valuation risk The Phase 3 result is genuinely important. INTerpath-001 hit both recurrence-free survival and distant-metastasis-free survival, validating the personalised neoantigen approach in a pivotal trial.  But Moderna has already repriced the success very aggressively. The market is now capitalising not merely the melanoma indication, but the possibility that this becomes a platform across multiple solid tumours. That is where I would be cautious. Full hazard ratios, subgroup consistency, overall survival, durability, manufacturing economics and regulatory details are still needed. Reuters speci

If I had to pick one piece of the AI infrastructure stack for the next six months, I would choose memory/storage, with Micron (MU) as my preferred exposure. AI is increasingly becoming a data-movement problem, not just a compute problem. HBM demand remains strong, while AI servers are also driving significant demand for high-performance SSDs and NAND. Tight supply and improving pricing could provide additional operating leverage. Micron is particularly interesting because it has exposure across HBM4, conventional server DRAM and enterprise SSDs, giving it multiple ways to benefit as AI infrastructure scales. Power could ultimately become the biggest bottleneck, but power-generation and grid projects generally have longer lead times. Chips remain attractive, but valuations and expectations

If I had to choose one for the next 2–3 years, I would pick Marvell (MRVL). The key distinction is that CPO is not simply an “optics boom”. It changes where the value accrues. 1. Marvell: best overall CPO exposure Marvell is positioned across the interconnect stack rather than relying solely on optical modules. Its Celestial AI acquisition gives it Photonic Fabric for scale-up CPO, with management targeting a US$500m annualised run-rate by FY2028 Q4 and US$1bn by FY2029 Q4.  That is potentially a much larger incremental opportunity than merely selling more transceivers. 2. AXT: my second choice, but potentially the biggest near-term torque AXT is becoming a critical upstream bottleneck. Q2 InP revenue hit a record US$30.7m, versus US$13.6m in Q1, driven by AI optical demand.  The

I would pick A. Micron for the next three years. Nvidia remains the strongest AI leader, but expectations and valuation are already extremely high. Micron offers a different way to capture the AI boom, particularly through HBM and high-end memory. AI workloads are becoming increasingly memory-intensive, creating potentially structural demand for faster, higher-capacity memory. The biggest attraction is the possibility that AI demand keeps memory supply tight for longer, allowing Micron to sustain unusually strong pricing and margins. If that happens, earnings growth could significantly outpace the broader market. Berkshire is the safer choice, with diversified businesses, strong cash flow and a huge liquidity cushion. It would probably be my pick if capital preservation were the priority.

Of the four, I would choose Micron for the best risk-adjusted exposure, although SanDisk has the most explosive upside. My ranking: Micron > SK Hynix > SanDisk > Western Digital. Micron: My preferred balance of HBM/DRAM exposure, AI demand and valuation. Druckenmiller's Q2 exit is worth noting, but I would not treat one fund manager's portfolio decision as a fundamental signal.  SK Hynix: Probably the strongest pure HBM beneficiary, but you are paying for that leadership. It is less directly exposed to the SanDisk/NAND thesis. SanDisk: Highest upside, highest risk. The Investor Day genuinely changes the story: eight NBM agreements covering roughly half of FY27 and two-thirds of FY28 capacity provide unusually strong demand visibility. Management is targeting mid-to-high-teens