Original Viewpoint Summary

🤖 AI Capex & Meta Compute Cycle

• Meta selling compute does not automatically mean it is quitting frontier model development. The real reversal signals are whether AI demand stays strong and whether monetization can catch up with compute supply. The internal message is mixed: Zuck admitted the past four months of AI agent development were slower than expected, and the restructuring has not delivered yet, but Meta still sounds optimistic on AI returns over the next 3-6 months and plans to keep investing and adjusting. 1

• Hyperscaler capex is still biased upward. Keep building AI. Keep selling compute. Entering cloud means even more capex, not less. The only real brake is demand rollover: commercialization speed falling too far behind compute supply. The core variable is still demand depth and monetization growth. 2

• Meta is both the semiconductor selloff trigger and the potential rescue bid. If Watermelon has really caught up with OpenAI GPT-5.5, and it uses an order of magnitude more compute than Avocado, the read-through is bullish for the AI semiconductor chain because compute intensity is still ramping, not peaking. 3

🧠 Semiconductors & Stock-Specific Views

• $AVGO still looks underrated. The view is straightforward: valuation has not fully priced in the quality and upside. 4

• For a beaten-down quality company, position sizing decides the playbook. Small position: average down at low levels to reduce cost basis. Heavy position: sell calls to harvest premium and wait for the bounce. 5

📈 Market Seasonality & Event Risk

• $SPX seasonality favors staying constructive in July, but the setup gets choppier afterward. Over the last 14 years, July was 13 green / 1 red, August was 8 green / 6 red, September was 7 green / 7 red, and October was 8 green / 6 red. The trade is to enjoy July strength, then respect a potentially more volatile next 3 months. 6

• Trump ringing the NYSE bell from the Oval Office next week is framed as a possible sentiment/event-risk marker. The only prior time he rang the NYSE bell was Dec 12th 2024; that day $SPY dropped .52% and $QQQ dropped .65%. Not a causal model, but worth noting as a tape-risk reference. 7