Is this email difficult to read? View it in a web browser. ›

The Wall Street Journal. The Wall Street Journal.

Sponsored by
Deloitte logo.

The Morning Download: CPUs Stage Comeback in the Age of AI

By Tom Loftus | WSJ Leadership Institute

 

Good morning. The CPU, the workhorse of the Information Age, is having another moment in this Age of AI.

CPU champion Intel, written off just months ago, on Thursday reported first quarter sales of $13.6 billion, beating analyst expectations by 11%, the Journal reports. 

The WSJ Leadership Institute's Belle Lin today reports on a multibillion-dollar chip deal between Amazon and Meta Platforms that involves, you guessed it, CPUs. The social-media company will use tens of millions of Amazon Web Services’ Graviton chip cores to support its AI agents and other AI initiatives.

 
Content from our sponsor: Deloitte
When Quantum Meets Orbit: Why It’s Not Too Early to Start Planning

Space is now a $600 billion commercial market, and quantum capabilities are maturing fast. Here’s why the convergence matters—and what leaders should consider today. Read More

More articles for CIOs from Deloitte
 
Share this email with a friend.
Forward ›
Forwarded this email by a friend?
Sign Up Here ›
 

A Graviton chip. Amazon Web Services

Why are CPUs a hot commodity? As Brendan Burke, a research director at Futurum Group, tells Belle, CPUs are useful for running certain applications and feeding that activity back to GPUs, the chips that made the AI boom possible. In other words, the two chips work well together for a variety of tasks that AI agents can accomplish, said 

CPUs are also a key part of the “post-training” step for large language models, or the part of the AI model-building process where pretrained models are more actively trained toward certain goals.

Infrastructure Week at the Download. Together, today's news (and news earlier this week on Google's inference chips) reflect the diversified approach to AI infrastructure, and shows that no single chip architecture can efficiently serve every computational task.

But not everyone benefits from the AI boom. Read on.

 

The Week in AI-Powered Layoffs

Meta plans to spend up to $135 billion on AI infrastructure this year Tony Avelar/AP

Meta to lay off 10% of employees in May. The social-media company has been unusually aggressive pushing to incorporate AI into its employees’ workflows, the WSJ's Meghan Bobrowsky reports.

It has started grading employees on their AI use in performance reviews, created ultra-flat teams with almost no managers and begun to develop a “CEO agent.” 

And this week it kicked into high gear

On Thursday, the company said it planned to lay off 10% of its workforce, or about 8,000 people, on May 20. In a memo sent to current employees that was viewed by The Wall Street Journal, Chief People Officer Janelle Gale said the cuts were necessary to allow the company to operate more efficiently and to offset its investments.

The company also said it would also cancel plans to hire for 6,000 open roles, the memo said.

Earlier in the week, an internal memo notified employees of a new software tool that would record their keystrokes, mouse movements and click locations to teach “the next generation of our AI models to use computers,” Reuters reported.

And in a separate missive in a memo-filled week, Meta’s technology chief, Andrew Bosworth, said the company is building towards a vision where AI agents primarily do the work. 

The new layoffs come on top of 1,500 jobs Meta already cut in January from its Reality Labs division. As of December 2025, according to Meta’s latest reported data, head count had reached 78,865.

Meanwhile, Microsoft announced buyouts to long-timers.

The program is part of a broader shift by Microsoft to alter its performance system and how it awards bonuses and stock options, according to an internal memo viewed by The Wall Street Journal.

To be eligible, employees must be at a senior director level or below, and their years of employment and age must add up to at least 70. Roughly 7% of the company’s U.S. employees are eligible. 

The not-so-invisible hand of AI. The memo, which also includes changes to the way it awards stock, comes amid a cascade of people changes, as struggles to build its own AI models and products and increased scrutiny over the costs and payoffs of building out AI data centers have helped send shares down nearly 20% in the past six months.

 

The SaaSpocalypse: Still a Thing?

SAP has reported a strong first quarter. Tatyana Makeyeva/Reuters

Shares in SAP took off in Europe after it reported a strong first quarter that showed the resilience of its cloud business. Sales from SAP’s core cloud business grew 27% to 5.96 billion euros.

The German tech company, by nature of the core software-driven services it provides, has been recognized as one of the many firms in danger of being replaced by new artificial-intelligence tools

But as we reported earlier, companies are not exactly ripping out their SAP ERP systems.

Neither are they dropping ServiceNow or IBM, but sentiment can change quickly. 

 

Can software companies survive an AI apocalypse? Tech reporter Sebastian Herrera tells us why Salesforce's CEO is pushing back against investor skepticism. Imani Moise hosts.

 

On Thursday, ServiceNow plunged 18% after the cloud-based software company cut its projection for operating margin. IBM stock slumped 8.3% after the company disappointed investors by keeping its revenue guidance the same.

Falling shares of megacap tech names also weighed on major benchmarks. Microsoft stock closed 4% lower after announcing voluntary buyouts while Meta shares slid 2.3% on news of those job cuts.

 

What We're Following

New models from OpenAI, DeepSeek. OpenAI billed GPT‑5.5, as a "new class of intelligence" and its strongest "agentic coding model to date" and boasting cybersecurity capabilities "a step up" from GPT‑5.4, released just months earlier. DeepSeek said its long-awaited V4 model V4 matches some top-tier U.S. products released late last year. However, its performance in certain areas still lagged behind leading closed-source models such as Anthropic’s Claude Opus 4.6. Despite being pricier than DeepSeek's earlier models, V4 remains much cheaper than its Western competitors, the Journal reports.

  • ICYMI: OpenAI Is Working With Consultants to Sell Codex

Oracle’s $300 billion megadeal with OpenAI tests Wall Street’s appetite for debt. While Wall Street has largely giving a blank check for the AI ambitions of the most creditworthy tech companies, Oracle is in a comparatively weaker position, says the WSJ's Peter Rudegeair and Berber Jin.

It has a lower investment-grade credit rating, more debt and is burning cash. Much of its future revenue is tied to a money-losing startup [OpenAI] that is facing growing competitive pressure. The cost of protecting Oracle’s bonds against a potential default via credit-default swaps roughly quadrupled between late September and late March, though it has fallen slightly since then.

The Journal notes that  Oracle's challenges highlight a risk for the multitrillion-dollar data center boom where any slowdown in data center construction would stymie a build-out that AI players desperately need.

Weyerhaeuser is using AI to digitize the forest. The country’s top logger outlined plans to boost annual profits by $1 billion—roughly double 2025’s, the WSJ reports. Executives expect a big chunk of those gains to come through AI, currently used on processes ranging from predictive maintenance of its mills to optimizing the routes taken by its trucks. Initiatives under exploration include an AI model to calculate seedling survival rates and semiautonomous logging equipment.

America’s first commercial nuclear-power projects in a decade just broke ground. A project by TerraPower, a company founded by Bill Gates almost 20 years ago, started construction Wednesday in Wyoming, while Kairos Power broke ground last week in Tennessee on a plant that intends to sell power to Google.

 

A New Arms Race in Trading

Remember when electronic trading firms fought to build the fastest microwave networks so they could execute trades in milliseconds? Now it's increasingly about raw compute. 

Meet Alex Gerko, CEO of algorithmic trading firm XTX Markets, and one of Britain’s wealthiest people thanks to his early bet on tapping AI to make money. He is set to make more, the WSJ reports.

Alex Gerko founded XTX in 2015, naming the trading firm after a mathematical formula. XTX Markets

XTX this year begins perations at the first of five planned data centers in Finland, expanding the firm’s ability to crunch vast quantities of financial data and train models to forecast price moves. And the firm is building the $1 billion-plus complex itself, not renting it. XTX also has amassed 25,000 AI chips, mostly from Nvidia.

Rivals are pursuing their own AI-based strategies. Earlier this month, Jane Street invested $1 billion in CoreWeave and signed a $6 billion deal to use the company’s AI infrastructure.

 

Everything Else You Need to Know

The U.S. has burned through so many munitions in Iran that some administration officials increasingly assess that America couldn’t fully execute contingency plans to defend Taiwan from a Chinese invasion if it occurred in the near term, U.S. officials said. (WSJ)

Federal authorities on Thursday charged a U.S. Army soldier who participated in the mission to capture Nicolás Maduro with using classified information about the operation to reap more than $400,000 in profit from bets on the former Venezuelan leader’s ouster. (WSJ)

 

The WSJ Technology Council

The WSJ Tech Council brings together CIOs, CTOs and CISOs advancing innovation and shaping the future. Join this trusted community where tech executives connect with peers to explore emerging trends and gain the perspective they need to stay ahead of disruption.

Request Information


Deloitte Logo.
 

About Us

The WSJ CIO Journal Team is Steven Rosenbush, Isabelle Bousquette and Belle Lin.

The editor, Tom Loftus, can be reached at thomas.loftus@wsj.com.

 
Desktop, tablet and mobile. Desktop, tablet and mobile.
Access WSJ‌.com and our mobile apps. Subscribe
Apple app store icon. Google app store icon.
Unsubscribe   |    Newsletters & Alerts   |    Contact Us   |    Privacy Policy   |    Cookie Policy
Dow Jones & Company, Inc. 4300 U.S. Ro‌ute 1 No‌rth Monm‌outh Junc‌tion, N‌J 088‌52
You are currently subscribed as [email address suppressed]. For further assistance, please contact Customer Service at sup‌port@wsj.com or 1-80‌0-JOURNAL.
Copyright 2026 Dow Jones & Company, Inc.   |   All Rights Reserved.
Unsubscribe