🧠 Original Views Summary

📈 AI Cycle & Earnings Quality

• AI stocks do not look like a full-blown bubble as long as EPS keeps compounding. Last earnings season showed EPS growth outpacing the S&P 500’s gain, meaning stocks rose while valuations actually got cheaper. Earnings-backed rallies are healthy. Macro-driven dips are long-term add points, not automatic sell signals. 1

• The AI capex cycle is still closer to the early innings than the cycle top. Yes, AI is cyclical. Oversupply and valuation compression will eventually show up. But current bottlenecks in chips, advanced nodes, memory, power, and data centers are slowing disorderly expansion, which may stretch the investment cycle longer. 2

• The “AI needs less capex” bear case looks weak when supposedly low-capex models later become compute constrained. The market panic in January 2025 and the usage limits by July 2026 point to the same conclusion: compute demand is still real, and underbuilt infrastructure can quickly become the choke point. 3

🧭 Fundamentals vs. Technicals

• For long-term investing in quality companies, fundamentals matter more than chart-reading. Ignoring business quality and outsourcing conviction to crypto-style technical calls is how investors lose money outside their circle of competence. 4

• The best setup combines fundamentals and technicals, but fundamentals remain the anchor. Technical indicators are just probability tools: if A happens, B may happen. Treating crypto, stocks, and gold as the same just because the chart looks similar misses their very different underlying drivers. 5

🎯 Small-Cap Speculation

• $WYFI reclaiming $30 would be a constructive technical signal, but this is still a small-cap speculative name. Position sizing and risk control matter. The bullish case is still tied to confidence in Leopold. 6