Hyperscaler earnings could accelerate dramatically after 2028.
Here's why: Their combined backlog has now reached roughly $2 trillion. Nearly half of that backlog is expected to convert into revenue over the next 12–24 months. As existing contracts expire, much of that capacity will be renewed at significantly higher prices, since current market rental rates are well above legacy contractual rates and continue to climb. Now add continued AI-driven demand and volume growth. At the same time, capital expenditure growth is expected to slow to around 10% annually by 2028, reducing the pressure on cash generation. The result? A powerful combination of higher pricing, rising volumes, moderating capex, and expanding margins could drive an inflection point in free cash flow, making the post-2028 earnings outlook far stronger than many investors currently expect. Source: Oguz Erkan
Star AI investor Leopold Aschenbrenner is unwinding trades after steep losses
Leopold Aschenbrenner's $20 billion AI hedge fund may be facing its first real stress test. After leaving OpenAI, Aschenbrenner became one of the most influential voices in AI with his viral Situational Awareness essay predicting AGI by 2027. He launched Situational Awareness LP in late 2024 with $225 million. Within two years, assets reportedly surged to ~$20 billion, fueled by massive gains in AI stocks. Now the trade is reversing. Oracle and AMD have each fallen around 20% in July, while smaller AI names such as Nebius, Bloom Energy, and SanDisk have dropped even more. The biggest signal? The fund is reportedly offering some investors the opportunity to purchase assets directly from its portfolio. That's unusual. Rather than selling into a weak market and pushing prices even lower, it may be seeking liquidity through private transfers. When even the hottest AI funds start looking for liquidity, investors should pay attention. Source: FT, Bull Theory
Updated look at the increasingly "circular AI financing" complex, mapped out by Bloomberg. All roads lead to Nvidia
Source: Hedgeye, Bloomberg
Narrative violation. WSJ yesterday.
Source: David Sacks
Fed will deliver surprise rate hike this week, says Citadel
Source: Barchart
Michael Burry is warning that private equity may have found a way to shift losses onto the public.
Apollo, KKR and Blackstone have acquired life insurers and loaded them with $849 billion of hard-to-value private credit, more than double the 2014 level. If an insurer fails, state guaranty systems protect policyholders. The cost is initially borne by other insurers, but they can offset those payments against state taxes, ultimately reducing public tax revenue. The risk is no longer theoretical. Bankruptcies such as First Brands and Tricolor exposed valuation problems in private credit. Burry argues AI could be the next test, as data centre financing increasingly relies on similar opaque debt structures. If those investments disappoint, the losses may extend well beyond Big Tech and into the insurance system that ultimately protects millions of policyholders. Source: Bull Theory
A little over one year ago, Coatue released a 100-page keynote presentation on the evolving technology landscape across public & private companies.
They predicted these 40 companies will be the largest by market cap in 2030. Note the absence of Alphabet $GOOG $GOOGL Source: Koyfin @KoyfinCharts
Global electricity consumption by data centers
Source: zerohedge
