Syz the moment

The AI boom is redistributing cash flow on an extraordinary scale. According to JPMorgan, Nvidia’s projected cash flow gains exceed the entire expected decline in hyperscaler free cash flow. Think about that. The tech giants funding the AI infrastructure build-out are seeing their free cash flow squeezed. Their chip supplier is capturing extraordinary profits. Same boom. Very different economics. For Nvidia, the sale generates cash now. For hyperscalers, the spending comes first. The returns depend on future demand, monetisation and utilisation. JPMorgan still sees attractive returns on certain data-centre investments. But attractive returns and immediate cash flow are two different things. The timing gap matters. And it raises a crucial question for investors: Who captures the cash—and who carries the risk? The biggest winners from the AI boom may be the suppliers getting paid upfront, while their customers wait for the payoff. Source: TME, JP Morgan

If you invested $10,000 in $NVDA at the bottom of the 2002 market crisis, you'd have about $43 million today. Source: Bull Theory

What’s even more wild is capex was under 4% of FCF last quarter making this one of most capital efficient infrastructure businesses ever built. Source: Shay Boloor

ByteDance, valued at $550 BILLION, and Alibaba, worth $268 BILLION, are among the firms that could soon get APPROVAL to buy Nvidia’s new chip. Beijing’s industry ministry has reportedly signaled that "approvals are coming". Source: Coin bureau

The credit market is asking a harder question: who pays for the AI buildout? The cost of insuring Nvidia’s debt against default has roughly doubled this year. SpaceX credit default swaps have approached 180 basis points. Both companies sold $25 billion of bonds in June, while hyperscalers are on track to spend more than $750 billion on data centers in 2026. Equity investors see growth, revenue and momentum. Credit investors are looking at the debt needed to fund it—and the returns required to repay it. Rising insurance costs do not mean a default is coming. But they do show that lenders are reassessing the risk. The stock price tells you how excited investors are about AI. Credit spreads tell you what financing that excitement costs. Watch both. $NVDA Source: Kurt S. Altrichter, CRPS

Huang is pushing for faster U.S. AI development and says safety concerns are a matter for “good old-fashioned engineering,” putting him at odds with other industry leaders calling for a slowdown and greater oversight. Trump is increasingly mirroring Huang’s policy stances on AI topics, telling Huang that “the robots are not going to be taking over the world.” Source: CNBC

That's how OpenAI's Greg Brockman closed the GPT-6 Astra launch. The headlines: → Trained on 100,000+ GPUs at Stargate, Texas → Claims the lead in coding, science, reasoning and cybersecurity → Works inside software: spreadsheets, 3D worlds, circuit boards, tax forms → First model to hit OpenAI's "Critical" cyber threshold — access restricted via Daybreak → Priced at 2.5x GPT-5.6 Why this matters for AI hardware & memory chips: 1) 100k GPUs for one model = the compute race is accelerating, not plateauing. 2) Agentic models that build, reason and act burn far more inference tokens. Inference = sustained demand for HBM and DRAM, not one-off training spend. 3) 2.5x pricing shows customers will pay for capability → hyperscaler capex stays funded. The picks-and-shovels trade remains alive and well.

Chart below is the relative performance of Nvidia relative to the broader semiconductors SOX index. Source: Bloomberg, GS, Negligible capital on X