I think there are actually two separate trades being mixed together here: 1. The Marvell Trade The bull case is straightforward. • Nvidia's CEO publicly endorsing Marvell is a huge credibility signal. • AI infrastructure spending is still accelerating. • Hyperscalers like Microsoft, Amazon, and Google continue investing heavily in custom silicon. • Marvell sits in critical areas like networking, optical connectivity, and custom AI chips. The question isn't whether Marvell benefits from AI. The question is whether a $200B company can realistically become a $1T company. That's a much harder leap. Going from $200B to $1T requires another 5x from here. Not impossible, but much of the easy re-rating may already have happened. 2. The AI Token Trade This is where I become more cautious. A lot of AI tokens trade on narrative rather than revenue, cash flow, or adoption. Just because AI hardware companies are winning doesn't automatically mean AI tokens should win. We've seen this repeatedly: • AI stocks rally • AI token narratives follow • Retail piles in late • Tokens correct sharply The market often treats them as related when the fundamentals are completely different. About the ETH Whale A 1.8M USDT ETH transfer sounds large on social media. In reality, it's tiny relative to Ethereum's daily liquidity. One whale moving funds to an exchange is worth monitoring, but I wouldn't build an AI or crypto thesis around it. My View If I had to choose: 🟢 Long-term: I'd be more confident in quality AI infrastructure companies like Nvidia, Marvell, TSMC, ASML, and Broadcom. 🟡 Medium-term: AI tokens can outperform dramatically, but they're far more dependent on sentiment and liquidity. The biggest risk isn't that AI spending collapses tomorrow. The biggest risk is that expectations have run so far ahead of reality that even good news stops moving prices higher. That's usually where corrections begin. So I'm not betting against AI. #AnthropicFilesForIPO #HYPEStakingETFLaunch #USIranOilRisk