Strings

Interesting take on how wealthy New York-style investing is different from VC-like San Francisco-style investing. To me it seems like New York-style investing is still "traditional finance-like." Here's an interesting comment on the whole Situational Awareness matter. I can see why one would do heavy, leveraged thematic bets when you see where money would flow, but mitigating risk still should be a priority, especially when you use leverage. Source: New York Times Not financial advice; do your own research!

$META reports its Q2 2026 earnings today. What am I looking for? 👇 $META’s largest and fastest growing division is the Family of Apps (FoA). This division houses Facebook, Instagram, WhatsApp, Messenger, and Threads and generates about 98-99% of $META’s total annual revenue. The way this division mainly makes revenue is through businesses paying $META to host their advertisements. $META has been spending a lot of capex on AI infrastructure, i.e., their data centers and necessary hardware (GPUs, SSDs, fiber optic cables, etc.) needed to run their specialized AI models (like Llama). The reason for this is so that their AI models can be trained effectively and understand $META’s customers’ behavior on their platforms (like what reels do users like, what ad types work best, etc.). $META’s platforms then adapt their algorithms for users and continue showing them content they like, leading to longer user engagement and customer retention, as well as a higher chance of buying the product displayed in the ads. As the number of customers who do this increases, businesses will see $META’s customers buying the services/products in their ads, and hence continue hosting their ads on $META’s platforms. When more and more business compete for an ad spot on $META’s platforms, this allows $META to charge higher and higher prices, thus increasing their revenue. Hence, what I’d like to see in $META’s earnings today is that metrics that track the number of customers use their platforms (like Daily Active People) continue staying strong and even increase, if possible. I’d also like to see that the volume of ads served and the average price per ad increased. This would signal to me that businesses continue paying $META’s prices for ad placements at an increasing rate, which justifies the spending they are doing on capex.

$GOOGL reported beautiful Q2 2026 earnings yesterday but their stock is currently getting punished due to the huge increase in capex spending of $195-205B from earlier estimates of $180-190B. Their Google Cloud revenue grew 82% YoY to $24.8B; Cloud backlog grew to $514B. This to me screams that enterprise AI adoption is accelerating, which justifies the premium that hardware suppliers are currently charging. Compute capacity is extremely important; see Kimi K3's tweet about their GPUs being constrained due to an unprecedented amount of users. This whole selloff doesn't make sense to me; sure, $GOOGL's free cash flow is negative for the first time, but that's because they are increasing their capex to further their fastest growing business, which is going to be Google Cloud IMO. (In their earnings, they specifically said that enterprise demand for cloud compute capacity is outstripping the physical hardware and data center infrastructure they currently have up and running.) Their CFO literally said this in the earnings call: "Given the supply-constrained environment, we plan to expand the use of third-party capacity in Q3 as a bridging strategy while we build out more internal capacity. This strategy allows us to keep growing our customer base and capture greater overall value. However, it will create modest margin pressure in the near term as we utilize this capacity." Enterprises who download open-weight models like Kimi K3 and want to build their own specialized LLMs for company uses would still not be able to use the full 2.8 trillion parameters on their own computing hardware unless they want to pay for thousands of dollars in GPU, HBM, NAND flash storage costs. Hence, they'll look for digital cloud providers (like $GOOGL) to provide the computing infrastructure and pay as they go (for compute power and storage) when they need to heavily run their specialized LLMs.

$GOOGL reports its much anticipated Q2 2026 earnings today. What am I looking for? 👇 Due to the large buildout last year in AI infrastructure and data centers, essential components such as GPUs, HBM, and flash storage have become very expensive, so companies like $MU, $SNDK, etc, have enjoyed extremely pricing power due to hyperscaler ($GOOGL, $MSFT, $META, $AMZN, etc.) capex spending and contracts. Now, investors are worried about the spending hyperscalers are doing (which combines up to around $700B), as it has only been increasing and their free cash flow is getting more and more compressed as a result of this heavy spending. Hence, the whole semis selloff recently (from customer concentration risk). Everyone is waiting with bated breath that $GOOGL (and $META and $AMZN next week) announces that they are keeping / increasing the spending on hardware; otherwise, I think companies like $MU, $SNDK, etc., might suffer further drawdown. Hence, here’s what I’m looking to hear in $GOOGL’s earnings today. I want to see that they have either reaffirmed or increased their spending on hardware, while still showing growth in their Google Cloud division. More importantly, I would like to see this division contribute a majority of their revenue growth this quarter, even more so than their TPU business. This would show to me that enterprise AI adoption is growing only faster and faster while keeping the hardware market afloat and justify their current premiums.