Had a coffee chat with a Point72 friend today. We both agreed that the biggest risk to the AI trades and US stock market in the future would be Coreweave. We are both skeptical of its execution and financial health. If there is something can pop the "AI bubble", it's not OAI/Anth IPO but this scam 💩 company $CRWV.

A month ago I called $VICR a generation buy at under $200. Now I am calling $STRL a generational buy under $600. Sterling sits on the critical path of the AI data center buildout twice, first in the dirt (clearing, grading, and padding the sites), then in the power (via its CEC electrical acquisition). My view: The business is arguably in the strongest position it has ever been, while the stock sits roughly 42% below its May high. - Management has raised 2026 guidance twice since February. - Adjusted EPS guidance has risen from a $13.75 midpoint to $20.00. - Signed backlog has more than doubled year over year to $4.33B. - The balance sheet is in a net-cash position. So why has the stock fallen? Three things: - Valuation, after the stock nearly tripled from year-end into mid-May. - Mix-driven margin optics following the Q2 print. - Increasing political and regulatory risk around Texas data centers. The core bet is straightforward: Sterling’s backlog and position across multi-phase campuses can translate into 2027 EBITDA growth meaningfully above the roughly 15% growth the current valuation appears to imply at a 15x multiple. From here, the potential upside depends on two things the market is not fully paying for yet: CEC’s margin ramp and a resolution of the Texas permitting question. Why now? The bar is lower than it has been in years. The setup is interesting because much of the expectation reset has already happened, while the next six months should provide substantially more information on the two biggest variables: Texas and future awards. On May 14, STRL closed at $889.03. At the time, the May 2026 adjusted EPS guidance midpoint was $18.73, putting the stock at roughly 47x forward adjusted EPS. Today, at roughly $513 and against a raised $20.00 midpoint, that multiple is about 25.7x. The bad news is also largely pre-announced. On the August 4 call, management explicitly told investors to expect softer Q3 awards and potentially a sequential decline in backlog. It also walked through the mix-related pressure on margins. That matters because when Q3 arrives, a backlog decline should be confirmation of what management already told investors rather than a new surprise. The Catalyst Window Texas has a defined information timeline. Four dates matter: - October 19: TCEQ’s compliance report to the governor is due. - Ongoing: ERCOT’s audit is the key process for potentially lifting the freeze. - January 2027: The Texas legislature convenes. Abbott has said he wants data center incentives removed. - Late February 2027: Based on Sterling’s prior reporting cadence, 2027 guidance should arrive around the Q4 results. By March, investors should have significantly more visibility into the Texas situation. Other catalysts are also clustered in this window. - Management has said strong bid activity should translate into heavier awards in Q4 and early 2027. - The CEC margin ramp should begin becoming visible. Management has outlined a 300–500 bps margin improvement over 12–18 months from the August update. Bottom line: If awards accelerate, CEC margins expand and Texas ultimately clears, $STRL has already demonstrated how quickly the market can re-rate the stock. The stock went from roughly $306 to $889 in just four and a half months earlier this year.

Hear me out. The company is saying their neocloud biz can scale to 3bil ARR in the next 18months. At 3Bil market cap you are basically just buying call option on rumor of this working. If it works stock is up 5x if it doesn’t the stock can prob at least double before it 0s

Ladies and gentlemen: the ultimate gaslighting from extreme context anxiety. "very very often after 4.6, opus will ask a question or make a suggestion, and then just do it." opus FREAKS THE FUCK OUT once the session goes too long: it will do ANYTHING to get it to end.

We're excited to announce that Frontrun has been acquired by Anthropic AI in a $141M transaction. Frontrun founder @goatgdv will be joining Anthropic's product team to lead AI-native trading infrastructure, bringing smarter trade analysis, natural language portfolio management, and intelligent risk detection. More details soon.

Silicon Welfare: Nvidia GPUs are happiest when they are running Claude Capybara

Attention high school kids looking at any @PalantirTech fellowship should read this—pay attention and stay in school because becoming a ponzi artist is the quickest way to lose credibility in your domain If you think Palantir runs on good faith, you will be surprised at the amount of unchecked fraud this company commits daily along with many versions of it that is not typically disclosed in mainstream media due to ad $ involved for the editors at @CNBC @WSJ @technology @axios #analyst

The compute bottleneck is massively under appreciated. I would guess the gap between supply and demand is growing single digit % every day.

Can AI replace security platforms? I asked Claude to build a CrowdStrike replacement. Claude: "I have to be straightforward: building a replacement for CrowdStrike isn't something I can do here. CrowdStrike is a massive platform built by thousands of engineers over a decade-plus. Its core capabilities include real-time kernel-level endpoint monitoring across millions of devices, a proprietary threat intelligence graph built from trillions of security events...That's not something you can replicate with a script — it's an infrastructure product." If you want to create AI, you need GPUs. If you want to deploy AI, you need security. That's not a hallucination – it's a fact. Watch the demo → youtube.com/watch?v=ilVWaT9W…

Carvana: New Information Reveals that BridgeCrest is the lienholder - $CVNA must explain How and How Much link to our new report: gothamcityresearch.com/post/…