Morgan Stanley's Andrew Sheets: Why AI Infrastructure Spending Remains Steadfast
Andrew Sheets, Global Head of Fixed Income Research at Morgan Stanley, elaborated on the economic theory behind the sustained massive spending on AI infrastructure in the latest "Thoughts on the Market" podcast.
Despite market concerns about an AI bubble, tech giants are accelerating investments in infrastructure such as data centers, chips, power, and cooling systems. Morgan Stanley estimates that capital expenditures by major hyperscale cloud providers will reach approximately $470 billion this year, rising to $620 billion next year, and potentially reaching $1 trillion annually by 2028.
Sheets believes this spending is highly price-insensitive, driven by the perception of AI as the most important technology for the next decade, with leading companies willing to invest continuously for long-term strategic advantages.
Source: Public Information
ABAB AI Insight
Andrew Sheets points out that AI capital expenditures have transitioned from the experimental phase to large-scale infrastructure development, primarily driven by a few cash-rich tech giants. These companies view AI as a core competitive advantage and will maintain high levels of investment even if short-term returns are uncertain. Morgan Stanley emphasizes that the credit markets (corporate bonds, asset-backed securities, etc.) will play an important financing role in this trillion-dollar expenditure.
In terms of capital pathways, AI infrastructure spending is becoming a major engine driving global capital expenditure growth, significantly benefiting related supply chains (chips, power, data center REITs), while also putting immense pressure on traditional infrastructure such as power grids and energy.
Structural judgment: Essentially, this represents capital concentration. AI is viewed as a strategic technology, leading to massive capital being highly concentrated among a few tech giants and their infrastructure supply chains, with the mechanism being that the expectation of long-term competitive advantages far exceeds short-term profit pressures, pushing global fixed income markets and real investments towards AI-related fields, forming a new round of technological capital cycles.
ABAB News · Cognitive Law
True strategic investments are never about short-term returns, but about winning or losing ten years down the line. Price-insensitive spending is the correct approach to building a moat. Those who treat AI infrastructure as a necessary battle will hold the capital pricing power in the next technology cycle.