AI Spending Sparks $680B Depreciation Bomb for Tech Giants

Tech giants are experiencing significant financial strain due to their heavy investments in AI infrastructure. Companies like Microsoft, Oracle, Meta Platforms, and Alphabet are projected to incur over $680 billion in depreciation charges in the coming four years. This financial reality contrasts sharply with their historical status as asset-light businesses, for which current financial disclosures are not adequately prepared.

The increased capital expenditure is primarily driven by the construction of data centers and the acquisition of specialized hardware like GPUs. These assets have long construction and useful life assumptions, which are challenging to forecast accurately. For instance, Alphabet has extended the assumed useful life of its data center equipment, boosting its reported earnings. However, this practice is under scrutiny, with some, like Michael Burry, labeling such accounting as 'earnings inflation.'

Morgan Stanley's analysis reveals that depreciation expenses could quadruple for Alphabet by its 2028 year-end, and Oracle's depreciation charge might balloon to $56 billion by 2029. These figures challenge the prevailing market assumption that hyperscaler operating margins will improve. To meet current expectations, these companies would need to drastically reduce other operating expenses, which is contrary to their recent spending trends.