Daave “3 and 30” Tepper

See Through the Noise Given market volatility, which is, of course, uncertainty, I want to articulate a view that is strongly held and informs positioning. This is not the time to make heroic bets, but volatility is opportunity to put cash to work and reset portfolios, because the market is telling us something about the future. If you believe stocks discount the future somewhat rationally, you have to listen to the market and incorporate that into your thesis. What I see happening is this: post-election, the market expects data center construction to slow. Post the memory/ bottlenecks fallout that began June 2026, the market says bottleneck supply will come, model efficiency will reduce DRAM demand, and capital markets put compute supply, and therefore bottleneck demand, at risk. Put those two together, the election result and the bottleneck crunch, view them in context, and the market is being extremely rational right now. A Rational Market First, hyperscalers are holding up fine. They have installed compute. They can raise token prices. All else equal, token supply will be less than it would have been without the election results. That means they have a little more pricing power on tokens and the optionality to deploy capital rationally. In other words, the quantity of compute supply will be less, all else equal, so there's less competitive threat of others taking share. The net result is that hyperscalers should be much better free cash flow vehicles over the next 2 to 3 years than they would have been otherwise. That's why the hyperscalers, Amazon and Microsoft being the purest, are doing quite well. Meta is also holding in, with high optionality to monetize its compute into this market. Google is a little hard to read because of last quarter's search numbers, but aside from that earnings move, Google's been fine as well. Second, the market is rational about OpenAI and Anthropic concerns. This is reflected in OpenAI first reducing its capital raise and then falling short even at a lower valuation, after it had already pushed its IPO to next year. And Anthropic is tenuously coming to market for its IPO now, but all indications suggest it will be a weak IPO, with pre-money valuations being talked down. Third are the raw GPU clouds and time to power plays. These companies provide supply and take construction and financing risk for the benefit of the most capital-sensitive buyers of tokens, meaning the frontier labs and the largest venture-backed companies not requiring value-add software. There's a lot of uncertainty in their financing, their pipelines, and now their offtake because of item two above. Forward I outlined three material and rationale aspects of current market but what does this mean going forward? First, the bottleneck names are clearly uncertain. It's hard to step into memory, storage, photonics etc. Even hard to step into GPUs and ASICs, because in truth, it's hard to forecast earnings even 2 to 3 years out. There are just too many variables in the equation, including data center build potential but also capital availability. For me, the bottlenecks are a no-fly zone. On the hyperscalers, this may be the most attractive opportunity from a risk-reward perspective, while also potentially providing mid to high double-digit returns. You're not going to double your money on a short horizon, but against a backdrop of massive uncertainty in the technology space, the hyperscalers feel like sleep-at-night holdings and there will be volatility so underwrite your capacity.

texas going blue gov means buy $amzn , no question installed compute just got more valuable