CNBC·2026-02-23 06:06Core Insights - Hyperscalers are significantly increasing their AI capital expenditure (capex) and are increasingly utilizing credit markets for funding, which is challenging their previously held 'fortress balance sheet' status [1][2][5] - Investors are concerned that this shift disrupts the "unspoken contract" that kept speculative AI spending separate from debt markets, raising questions about creditworthiness [4][5] Capital Expenditure Trends - Major tech companies like Amazon, Meta, and Alphabet have announced substantial increases in their full-year capex plans, with UBS projecting that aggregated capex among AI hyperscalers could exceed $770 billion by 2026, a 23% increase from prior expectations [2] - This increase in capex is expected to lead to an additional $40 billion to $50 billion in borrowing, pushing public market debt issuance to between $230 billion and $240 billion for the year [2] Market Dynamics - The shift towards bond markets is altering the relationship between hyperscalers and investors, as these companies are now seen as taking on more debt rather than relying solely on cash flow for AI investments [3][4] - Investors are now scrutinizing the debt levels of these companies, which were previously viewed as low-risk due to their strong credit ratings [5][10] Investor Sentiment - BlackRock has indicated that mega-cap tech companies are using the current credit issuance boom to bridge the gap between current investments and future revenues, raising concerns about rising corporate borrowing adding supply to bond markets [5][6] - The focus of the market has shifted to how AI adoption will translate into revenues and profits, creating a need for active investing strategies [9] Financial Health and Risks - While AI hyperscalers maintain strong balance sheets and cash flow generation, they are taking on more leverage, which raises concerns about potential hidden risks in the system [12][13] - There are fears that rapid technological advancements could render large data centers obsolete, impacting the long-term viability of investments in these assets [10][11]