🌾 Welcome to StableBread's Newsletter!
I'm sure you've all felt or seen the sell-off over the last couple of weeks. Semis just closed their worst month since October 2008, the Nasdaq 100 traded more than 10% below its June record, and the rare earth and critical mineral names I’ve covered are at multi-month lows.
For perspective, my portfolio went from nearly 200% YTD to just under 100%:
StableBread’s YTD Returns
The best way to stay disciplined is to question whether the companies in your portfolio are in a better or worse position than when you originally bought them. Focus on the business and thesis, not the price movements.
Since early April, I've published 20 individual deep dives. Below, I provide a brief overview of the market backdrop, then revisit five names I'm bullish on from my own portfolio (each from a past write-up).
I recap the original thesis, what has happened to the business since, how the market has reacted, and whether I'm deploying more cash at current prices. All of these names are in a better position than when I wrote them up, and several trade at even wider discounts.
The Market Backdrop
Before I get into the names, it's important to understand what has happened over the last month. Two main stories are worth focusing on:
AI funding reset: The market started questioning how the AI buildout gets funded and repriced everything that depends on outside capital to grow.
Critical minerals escalation: Washington and Beijing spent the final week of July stacking new restrictions on each other's supply chains.
Most of the names I discuss later are directly downstream of one of these two stories.
AI Funding Reset
Here's a recap of last month's key events in relation to the AI buildout:
July 9: S&P downgraded Oracle to BBB-, just above junk-rate bonds. OpenAI makes up ~50% of Oracle's $638B backlog, and S&P projects Oracle's free cash flow deficit widening to $42B a year.
July 14: New York signed the country's first statewide moratorium on new hyperscale data centers, pausing state environmental permits for 50MW+ projects for up to a year.
July 22: Alphabet doubled quarterly capex to $44.9B, raised its 2026 capex guide to $195-205B, and reported negative free cash flow ($5.9B) for the first time in its public history. The stock fell 7.1% the next day, and the Mag 7 had their worst session since April 2025.
July 28: CoreWeave's five-year credit default swaps (cost of insuring its debt) traded near 855 bps, pricing ~50% odds of default within five years. The stock was down 36% in a month.
July 29-31: Microsoft (Azure +43%) and Amazon (AWS +37%) both beat expectations, rising 15.5% and 15.3%, Amazon's best day since 2012. Apple fell 7.4% on supply constraints, and Korea's Kospi followed an 18% three-day drop with a 17.9% single-day gain, its largest on record.
Overall, demand is not the problem. Google Cloud grew 82%, Azure grew 43%, and AWS grew 37% last quarter.
What changed is the market's willingness to fund the buildout with debt. The four hyperscalers are guiding to ~$724B of combined 2026 capex, and even they are starting to strain, with Alphabet's free cash flow turning negative this quarter.
The names one layer down depend on borrowed money entirely, which is why Oracle (BBB-, $167B of debt) and CoreWeave ($24.9B of debt against ~$6.2B of trailing revenue) led the declines.
It doesn't help that futures price ~60% odds of a September Fed hike, the 10-year Treasury ended the week at 4.71% (its highest since January 2025), and oil crossed $100 mid-month during the Iran escalation.
Put simply, higher rates make debt-funded data centers more expensive to build and lower what investors will pay today for revenue that arrives in 2027 and beyond.
And the AI data centers with healthy balance sheets and unfilled capacity are where you should be looking (one such name discussed later).
Critical Minerals Escalation
The critical minerals story is ongoing, but here's what's happened recently:
July 30: Trump signed a Defense Production Act determination authorizing the Commerce Department to restrict exports of recoverable critical minerals, meaning the scrap, black mass, and end-of-life magnets that carry tungsten, lithium, and rare earths out of the country today. Per Reuters, the goal is keeping that feedstock home ahead of January 1, 2027, when defense contractors are barred from buying Chinese tungsten, molybdenum, and rare earth magnets.
July 30: Nikkei reported SpaceX is requiring suppliers to keep Chinese nationals and Chinese-made equipment out of its supply chain, and the WSJ reported Tesla executives were told to prepare a separation of the China business.
July 31: The US added 43 Chinese companies to the Uyghur Forced Labor Prevention Act Entity List (a US import blacklist for goods tied to forced labor in Xinjiang), the largest addition since the law passed in December 2021. Notably, ~20 of the 43 are tied to mining, metals, and materials. The import ban takes effect August 3.
August 1: China's Ministry of Commerce called the listings "completely groundless" and vowed to "take necessary measures" in response.
Clearly, none of this is de-escalation. Entity lists, export controls, and DPA determinations are supply-side restrictions, and the decoupling now shows up in supplier requirements, not just speeches.
Look at the physical markets as well. APT (ammonium paratungstate, the traded form of tungsten) ended July above $3,000 per metric tonne unit in Europe, against ~$340 before China's February 2025 export controls. NdPr set a 2026 high in early July, terbium holds above $1,000/kg, and dysprosium trades near $210/kg.
You'd expect US-based critical mineral stocks to rally on these escalations, but so far they've gone the other way. MP Materials fell 26.1% in July, USA Rare Earth fell 30.7%, and Energy Fuels fell 21.1%, all into multi-month lows, even as the prices of what they produce hold near records and each new restriction makes their assets more strategic.
That's the quick market backdrop. All of this is important to understand since it ties directly to most of the portfolio names I'll discuss below…
Subscribe to StableBread's Research to keep reading!
Full access to this write-up and every future one. Cancel anytime.
View PlansSubscribe to unlock:
- 1-3 write-ups per week
- Periodic industry, market, and thematic research
- Financial models behind most write-ups
- Full archive of all past research
