print-icon
print-icon
premium-contentPremium

"Here Come Humanoids": Morgan Stanley Braces For The Looming Phase Shift in AI

Tyler Durden's Photo
by Tyler Durden
Monday, May 05, 2025 - 01:35 AM

This article is so good
it's for premium members only.

Does that sound like you?

Already a member? Sign in.

PREMIUM


ONLY $30/MONTH

BILLED ANNUALLY OR $35 MONTHLY

All BASIC features, plus:

  • Premium Articles: Dive into subscriber-only content, market analysis, and insights that keep you ahead of the game.
  • Access to our Private X Account, The Market Ear analysis, and Newsquawk
  • Ad-Free Experience: Enjoy an uninterrupted browsing experience.

PROFESSIONAL


ONLY $125/MONTH

BILLED ANNUALLY OR $150 MONTHLY

All PREMIUM features, plus:

  • Research Catalog: Access to our constantly updated research database, via a private Dropbox account (including hedge fund letters, research reports and analyses from all the top Wall Street banks)

By Katy Huberty and Adam Jonas of Morgan Stanley

A core strategy of Morgan Stanley Research is combining deep sector and macro expertise with a global footprint and collaborative culture to identify investment themes that help our clients generate alpha. Over the past 30 years, fewer than 3% of stocks drove the entire $80 trillion of global equity market expansion. These stocks tend to sit at the center of the most important investment themes. We expect our four key themes for 2025 AI Diffusion, Longevity, Future of Energy, and Multipolar World – to regain their leadership positions and even accelerate as markets calm.

Clients often ask if the AI infrastructure investment that led the US equity market over the past two years will be sustainable when markets recover. We think it will be and that we’re still in the early innings of the AI infrastructure buildout. As AI-driven productivity takes hold, we expect the beneficiaries to broaden out. Our research has shown that every 10 years a new computing cycle emerges, with access to compute and thus the TAM increasing tenfold. Based on this pattern, $1 trillion of CPU-based compute for traditional software applications (the last cycle) implies $10 trillion for GPU-based compute deploying AI applications. This scenario appears to be playing out, with early-stage AI capex tracking at 10x early cloud capex.

Want more of the news you won't get anywhere else?

Sign up now and get a curated daily recap of the most popular and important stories delivered right to your inbox.