rwang07
@rwang07
Past bets that played out
These are the clearest thesis calls with observable outcomes, linked back to the original videos.
Post is a question plus a quoted technical comparison: AMD HBM4 data rate per pin (7.5 Gbps) vs Nvidia HBM4 (10.7 Gbps). No explicit supplier named; no clear catalyst date; implies potential performance/throughput disadvantage for AMD vs NVDA if accurate.
Post is a question plus a quoted technical comparison: AMD HBM4 data rate per pin (7.5 Gbps) vs Nvidia HBM4 (10.7 Gbps). No explicit supplier named; no clear catalyst date; implies potential performance/throughput disadvantage for AMD vs NVDA if accurate.
Post highlights a purported technical disadvantage for AMD in HBM4 (lower per-pin data rate vs Nvidia). If accurate and market-relevant, it supports a near-term narrative tailwind for NVDA and a relative headwind for AMD in AI accelerator competitiveness, but the post lacks context (config, total bandwidth, stack size, validation status), making actionability moderate-to-low.
What this channel is watching now
Recent mentions show where this author is concentrating attention right now.
Latest videos and market context
Recent source posts from this author. Create an account to inspect the complete persisted research trail.
Ray Wang @rwang07 8h Who is the biggest HBM suppliers for AMD? 🤓 Zephyr @zephyr_z9 12h AMD's HBM4 has a much lower da...
Post is a question plus a quoted technical comparison: AMD HBM4 data rate per pin (7.5 Gbps) vs Nvidia HBM4 (10.7 Gbps). No explicit supplier named; no clear catalyst date; implies potential performance/throughput disadvantage for AMD vs NVDA if accurate.
Ray Wang @rwang07 9h 💀 Zephyr @zephyr_z9 12h AMD's HBM4 has a much lower data rate per pin compared to Nvidia Nvidia'...
Post highlights a purported technical disadvantage for AMD in HBM4 (lower per-pin data rate vs Nvidia). If accurate and market-relevant, it supports a near-term narrative tailwind for NVDA and a relative headwind for AMD in AI accelerator competitiveness, but the post lacks context (config, total bandwidth, stack size, validation status), making actionability moderate-to-low.
Pinned Ray Wang @rwang07 1h 1) Most attention today is on leading DRAM and NAND suppliers. But in-depth research (7k+...
Post highlights a potential “biggest memory supercycle” (bullish for memory pricing/earnings) but flags rising competitive risk from China’s CXMT (bearish for DRAM incumbents over time). Mentions CXMT nearing IPO (not directly tradable in US public markets as of text) and names SK hynix, Micron, Samsung as key incumbents potentially affected.
Ray Wang @rwang07 May 20, 2025 Report: Malaysia becomes the first country outside China to deploy Huawei chips (likel...
Post cites a report that Malaysia is deploying Huawei AI chips (likely Ascend GPUs), servers, and DeepSeek’s LLM as part of a national AI infrastructure launch—implying incremental adoption of China-based AI compute + model stacks outside China and potential substitution vs US/Nvidia-centric stacks in some emerging-market sovereign/regulated deployments.
Proof-backed call history
These are recent thesis calls tied to original source content where available.
Post is a question plus a quoted technical comparison: AMD HBM4 data rate per pin (7.5 Gbps) vs Nvidia HBM4 (10.7 Gbps). No explicit supplier named; no clear catalyst date; implies potential performance/throughput disadvantage for AMD vs NVDA if accurate.
Post is a question plus a quoted technical comparison: AMD HBM4 data rate per pin (7.5 Gbps) vs Nvidia HBM4 (10.7 Gbps). No explicit supplier named; no clear catalyst date; implies potential performance/throughput disadvantage for AMD vs NVDA if accurate.
Post highlights a purported technical disadvantage for AMD in HBM4 (lower per-pin data rate vs Nvidia). If accurate and market-relevant, it supports a near-term narrative tailwind for NVDA and a relative headwind for AMD in AI accelerator competitiveness, but the post lacks context (config, total bandwidth, stack size, validation status), making actionability moderate-to-low.
Post highlights a purported technical disadvantage for AMD in HBM4 (lower per-pin data rate vs Nvidia). If accurate and market-relevant, it supports a near-term narrative tailwind for NVDA and a relative headwind for AMD in AI accelerator competitiveness, but the post lacks context (config, total bandwidth, stack size, validation status), making actionability moderate-to-low.
Post highlights a potential “biggest memory supercycle” (bullish for memory pricing/earnings) but flags rising competitive risk from China’s CXMT (bearish for DRAM incumbents over time). Mentions CXMT nearing IPO (not directly tradable in US public markets as of text) and names SK hynix, Micron, Samsung as key incumbents potentially affected.
Post highlights a potential “biggest memory supercycle” (bullish for memory pricing/earnings) but flags rising competitive risk from China’s CXMT (bearish for DRAM incumbents over time). Mentions CXMT nearing IPO (not directly tradable in US public markets as of text) and names SK hynix, Micron, Samsung as key incumbents potentially affected.
Post highlights a potential “biggest memory supercycle” (bullish for memory pricing/earnings) but flags rising competitive risk from China’s CXMT (bearish for DRAM incumbents over time). Mentions CXMT nearing IPO (not directly tradable in US public markets as of text) and names SK hynix, Micron, Samsung as key incumbents potentially affected.
Post quotes Nvidia CEO Jensen Huang testifying to the U.S. House Foreign Affairs Committee, framing U.S. policy on AI leadership vs “retreat and retrench” as an “inflection point.” This is directional context for AI policy/export-control/regulatory outcomes, with the most direct public-market linkage to Nvidia (NVDA) and broadly to U.S. AI infrastructure beneficiaries/risks.
Post cites a report that Malaysia is deploying Huawei AI chips (likely Ascend GPUs), servers, and DeepSeek’s LLM as part of a national AI infrastructure launch—implying incremental adoption of China-based AI compute + model stacks outside China and potential substitution vs US/Nvidia-centric stacks in some emerging-market sovereign/regulated deployments.
Post cites a report that Malaysia is deploying Huawei AI chips (likely Ascend GPUs), servers, and DeepSeek’s LLM as part of a national AI infrastructure launch—implying incremental adoption of China-based AI compute + model stacks outside China and potential substitution vs US/Nvidia-centric stacks in some emerging-market sovereign/regulated deployments.
Post contains only the phrase “Drones and China.” with no specific claim, catalyst, company, ticker, or tradeable implication. Treated as low-actionability context.
Post highlights a reported Chinese policy initiative: an “AI Industry Development Action Plan” backed by China Bank support, providing ~1 trillion yuan (~$137B) over five years to support China’s AI industry chain. This is framed as a major 2025 Chinese AI policy catalyst. Actionable mainly as a medium/long-horizon pro-China AI/tech sector tailwind rather than a single-name catalyst.
About this channel
Channel bio, source link, and public-market context from YouTube.
@rwang07
Most recognized assets
Unlock the full track record
Create an account to inspect the complete author history, trust-weighted rankings, and persisted research across authors, theses, and assets.
4 more thesis calls are available after sign-up.