This document is intended for institutional investors and is not subject to all of the
independence and disclosure standards applicable to debt research reports prepared for retail
investors under U.S. FINRA Rule 2242. Barclays trades the securities covered in this report for its
own account and on a discretionary basis on behalf of certain clients. Such trading interests
may be contrary to the recommendations oered in this report.
Barclays Capital Inc. and/or one of its ailiates does and seeks to do business with companies
covered in its research reports. As a result, investors should be aware that the firm may have a
conflict of interest that could aect the objectivity of this report. Investors should consider this
report as only a single factor in making their investment decision.
* This individual is a member of the Product Management Group and is not a Research Analyst
All research referenced herein has been previously published. You can view the full reports,
including analyst certifications and other required disclosures, by clicking the hyperlinks in this
publication or by going to our Research portal on Barclays Live.
FOR ANALYST CERTIFICATION(S) PLEASE SEE PAGE 31.
FOR IMPORTANT EQUITY RESEARCH DISCLOSURES, PLEASE SEE PAGE 31.
FOR IMPORTANT FIXED INCOME RESEARCH DISCLOSURES, PLEASE SEE PAGE 32.
Global Portfolio Manager's Digest
Striking a Balance
We provide context and perspective on research across regions
and asset classes, this week highlighting our thoughts on what
could turn sentiment positive on the "AI trade"; adjusting our
call following the Fed's decision to cut rates by 50bp; and what
the Fed's move means for Asia markets.
• What's Next for the "AI Trade": The bull case around AI boils down to a simple concept: if
scaling laws hold and frontier model performance continues to improve at great leaps over
coming generations, then end-user products in consumer and enterprise will similarly
improve greatly in terms of functionality and accuracy and adoption will increase. To that
end, we see GPT-5’s upcoming release as a potential catalyst to get things going again. If
successful, sentiment could shi greatly, AI hyperscaler capex would need to ramp further,
and all this investment will likely end up being another expensive moat for mega-cap tech.
Further, we think that another recent breakthrough around test-time-compute could usher in
a new sub-category of frontier AI lab innovation around finding a more optimal mix of training
and inference compute, which could dramatically reduce the capex needs of the overall
industry, but it's too early to tell. An alternative view, and much more bearish take, is that
scaling laws don’t hold up, the foundation models stop improving, and the capex wave stops.
In this scenario, GPT-5 would likely be introduced in late 2024 or early 2025 to great fanfare,
only to see the same kind of experimental enterprise AI workloads running on top of it but no
killer apps or broad-based consumer/enterprise adoption.
• The Fed's Recalibration: The FOMC initiated its rate-cutting cycle with a 50bp cut, against
our expectation of a 25bp cut, in an eort to recalibrate the policy stance and to maintain the
labor market at full employment. We think that the FOMC statement suggested that future
rate cuts are likely to be in 25bp steps, by showing no alarm about the cooling labor market,
and mentioning that risks to achieving "employment and inflation goals are roughly in
balance." The new Summary of Economic Projections (SEP) also suggested a slower pace of
Cross Asset Research
22 September 2024
FOCUS
Equity Product Management Group
Terence Malone
*
+ 1 212 526 7578
terence.malone@barclays.com
BCI, US
Rob Bate
*
+44 (0)20 7773 3576
rob.bate@barclays.com
Barclays, UK
FICC Product Management Group
Ben McLannahan
*
+44 (0)20 3134 9586
ben.mclannahan@barclays.com
Barclays, UK
Jennifer Cardilli
*
+1 212 526 8351
jennifer.cardilli@barclays.com
BCI, US
Completed: 20-Sep-24, 23:05 GMT Released: 22-Sep-24, 13:00 GMT Restricted - External
剩余37页未读，继续阅读
资源评论
soso1968
- 粉丝: 2669
- 资源: 1万+
最新资源
- ResNet-50 是一个深度卷积神经网络架构，常用于图像识别任务 它是 ResNet 系列网络中的一个变种，具有 50 层深度 ResNet-50以其深度残差连接结构而闻名，这种结构允许网络训练更深
- java前后端分离vue个人博客系统源码数据库 MySQL源码类型 WebForm
- 基于ESP8266和继电器模块实现远程控制电脑电源（网页界面远程开关机）.zip
- 卷积神经网络研究综述-周飞燕
- Jesse的个人博客，以梦为马，不负韶华 技术栈：HTML、CSS、JavaScript、TypeScipt、Vue、React、Angular、Node、Hybrid App、数据结构与算法等
- Mars 是微信官方的终端基础组件，是一个使用 C++ 编写的业务性无关、平台性无关的基础组件 目前已接入微信 Android、iOS、Mac、Windows、WP 等客户端
- 零基础入门计算机图形学必不可少的在线网络公开课，手把手教您现代 OpenGL 的点点滴滴，构建爆款游戏引擎 基于 OpenGL 的 3D 游戏引擎，开发出爆款开源游戏
- 基于JavaFX的UI组件库-含常用的UI组件-快速构建JavaFX应用程序界面+使用说明.zip
- 基于Java和Kotlin的炉石传说自动化脚本项目源码+说明文档.zip
- 基于Laravel开源免费的自媒体商城博客CMS企业建站系统
资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！
点击此处反馈
