Amazon's Mega Bond Sale Reignites the Hyperscaler Capex Debate
Amazon is reportedly considering issuing at least $25 billion in dollar-denominated bonds, with most of the proceeds earmarked for AI infrastructure investment. Deutsche Bank charts show hyperscalers' capital expenditure now running ahead of their operating cash flow, feeding a persistent market question: how long can this keep up?
Some read the borrowing as a sign that Big Tech is running low on cash, but the hosts pushed back: lenders wouldn't extend credit unless the borrowing made sense, and companies wouldn't borrow unless it were worth it. The key point is that this spending is investment, not cost — just as nobody asks 'why did you spend so much' when someone buys a home, AI capex should be judged the same way.
A U.S. Treasury internal report reportedly flagged AI-bubble warning signs, noting that financial markets are now far more tightly interwoven with the real economy than during the dot-com bubble, meaning one major AI failure could cascade broadly. That systemic risk, the hosts suggested, is exactly why the U.S. government has taken direct stakes in OpenAI and Intel.
Morgan Stanley issued a report arguing that the narrow, semiconductor-led rally is ending and market leadership is broadening, recommending trimming semiconductor exposure in favor of hyperscalers — framing it as 'cyclical rotation' rather than the end of the AI capex cycle. The report coincided with a sharp selloff in U.S.-listed memory names that day.
The hosts called this logic a 'circular reference error' — if the AI ecosystem keeps expanding, hyperscaler capex growth ultimately flows back into semiconductor demand, so the two can't be cleanly separated. They also recalled Morgan Stanley's shaky track record, including its 'Winter Is Coming' report that rattled markets before the firm apologized roughly six months later, and noted that SK Hynix reportedly excluded Morgan Stanley from its U.S. ADR listing's institutional allocation — widely read as payback for past bearish calls that cast doubt on the messenger's credibility.