What's up: OpenAI, Broadcom forge multibillion-dollar chip-development deal; Salesforce plans $15 billion investment in hometown San Francisco; Google to invest $24 billion in AI in U.S., India

Good morning. Here’s an interesting stat that captures rising corporate demand for AI. The number of forward-deployed engineers, who help companies customize software to their specific needs, is soaring.

Startups embed staff with customers to teach enterprises how to use the AI and to teach AI how to act in specific customer environments, WSJ Pro VC reports. AI requires more hand-holding, especially by the standards of software-as-a-service.

The median number of job vacancies for forward-deployed AI engineers rose 1,066% to 1,050 vacancies through August, according to Job.zip. The median number of companies hiring for the role rose 650% to 450.

“The proliferation of startups using FDEs comes at a pivotal moment in AI adoption,” given that many companies still struggle to get value out of AI investments. In theory, the forward-deployed engineer should help AI companies and their customers get more value from one another.

Marc Vartabedian and Yuliya Chernova of WSJ Pro VC provide an example of how it works:

Jordan Chin changed his role at New York-based startup Regal to a forward deployed engineer earlier this year. Regal, which raised $83 million in venture capital from Emergence, Homebrew, Founder Collective and others, helps businesses build voice AI agents for customer support and sales.

Chin works with Regal’s customers—from roadside assistance services to Medicare Advantage plans—to tailor AI voice agents’ intonations, conversation content, and even names. He takes what he learns back to Regal’s product team to suggest new features.

“There’s this misconception that AI either works or doesn’t work,” Chin said. In reality, “it involves a lot of iteration and testing.”

The use of such engineers makes sense now, but over time, they could add too much cost to AI deployments, according to Renata Quintini, co-founder and managing director of venture firm Renegade Partners. “If every sale needs engineers in the field, you’re scaling people and not software,” Quintini said.

For now, the use of forward-deployed engineers provides a useful measure of demand for AI, and how customers are working through the very real challenges of making sure their investments are actually useful. 

OpenAI, Broadcom forge multibillion-dollar chip-development deal.

The two companies are working to develop and deploy 10 gigawatts of custom AI chips and computing systems over the next four years, the WSJ reports. As part of the agreement announced Monday, the chips will be deployed by Broadcom starting in the second half of next year.

The Information reports that OpenAI also is talking with Arm about designing a CPU that would work with the server chip it is making with Broadcom. The chip designer is a subsidiary of SoftBank which has invested billions in OpenAI. 

OpenAI’s hunger for compute. The deal brings the total scale of computing capacity OpenAI has agreed to buy from chip giants Broadcom, Nvidia  and Advanced Micro Devices to 26 gigawatts, enough to meet the summer electricity needs of New York City more than two times over.

Salesforce plans $15 billion investment in hometown San Francisco over five years

The funds will launch an AI Incubator Hub, expand workforce-development programs, and help companies become “agentic enterprises,” the WSJ reports.

CEO Marc Benioff framed the plan as a commitment to jobs, innovation, and community impact. Salesforce also on Monday pledged $39 million to local education and healthcare initiatives.

The announcement coincides with the company’s annual Dreamforce conference. Shares have fallen more than 25% this year.

Google to invest $24 billion in AI in U.S., India

Google said it plans to establish its first AI hub in India as part of a $15 billion investment aimed at adding gigawatt-scale compute capacity to help meet growing demand for digital services both in India and around the world. The company also this week revealed plans to invest $9 billion by 2027 to expand its data center campus in Berkeley County, South Carolina, and support the ongoing construction of two new sites in Dorchester County.

California Gov. Gavin Newsom signed a handful of bills aimed at the tech industry including AB 56, which requires health warning labels for social media and SB 243, regulating AI chatbots.

Nvidia plans to start selling its desktop-sized DGX Spark this week, the Verge reports. The “personal AI supercomputer” comes with Nvidia’s GB10 Grace Blackwell Superchip.

French advertising group Publicis boosted its annual outlook for the second time this year and said artificial intelligence powered its top line. “We are demonstrating that artificial intelligence at Publicis is not a future promise, it is a reality today that is driving our growth,” Publicis Chairman and Chief Executive Arthur Sadoun said.

Qantas Airways said customer data from a July cyber incident that involved some 5.7 million records has been found online, Bloomberg reports that the airline said it is working with authorities.

President Trump’s surprise announcement of 100% tariffs against China on Friday triggered a cryptocurrency selloff that wiped out more than $19 billion in leveraged positions. Two accounts that placed bets against the market minutes before the news broke scored a $160 million windfall, WSJ reports.

President Trump declared an end to the Gaza war Monday after more than two years and vowed to extend his peacemaking to the wider Middle East, taking a victory lap in a region still facing deep divisions and numerous unresolved conflicts. (WSJ)

Ford Motor is temporarily cutting production of at least five models including popular SUVs after a devastating fire at a crucial aluminum supplier’s plant. (WSJ)

JPMorgan Chase reported a 12% jump in third-quarter profit that beat expectations, reaping the benefits of a surge in trading and dealmaking on Wall Street and steady consumer spending on Main Street. (WSJ)

The White House’s move to lay off roughly 4,000 government workers during the federal shutdown is expected to affect a range of services, including offices that focus on financing community development, reducing chemical pollution and supporting special education. (WSJ)

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.