Meta and Google enter AI agent race as ‘agent wars’ intensify

This report is from this week’s The Tech Download newsletter. Is it what you see? You can subscribe here.

Earlier this year, the agent-based AI tool OpenClaw went viral, with everyone and their grandma lining up to download the digital assistant onto their devices.

Nvidia’s Jensen Huang raved about the tool, calling it “the next ChatGPT,” and OpenAI criticized OpenClaw’s creator Peter Steinberger (which is a sign of the AI ​​Lab’s own intentions in this space).

Several months have passed, and the race to develop agent technology (AI tools that can perform tasks on your behalf) has begun to intensify among big tech companies.

In the past week, both Meta and google Reportedly working on an AI agent. Meta is building “highly personalized AI assistants to perform everyday tasks” for users, the Financial Times reported. According to Business Insider, Google is developing a “24/7 personal agent for work, school, and daily life powered by Gemini.”

Meta did not respond to a request for comment, and Google declined to comment.

“The immediate catalyst was OpenClaw,” Nick Patience, head of AI at Futurum Group, told CNBC. “This open source agent demonstrated a real desire for AI to act, not just give answers.”

Competitive pressures are a visible driver, but there is a “deeper logic”, he added. “Agents represent the point at which an AI platform transitions from a cost center to a revenue infrastructure, whether through commerce, advertising, or enterprise productivity.”

Malik Ahmed Khan, senior analyst at Morningstar, said trading agencies could be a “huge value driver” for tech companies like Google and Meta, both of which have large advertising and e-commerce businesses.

Gartner analyst Arun Chandrasekaran said Big Tech companies ultimately see AI agents as a way to shore up user subscriptions and maintain control of their platforms.

“Agents can generate more engagement, utility, and locked-in customers on the platform because they can provide more tangible value,” he said. “Also, the agent continues to learn and acquire user context over a period of time, making it more sticky.”

However, security and governance around AI agents is still being developed. In February, a Meta employee made headlines when he posted that OpenClaw had deleted a large number of emails of its own volition.

Then there’s also the issue of trust and how companies can manage the risk of an AI agent doing the wrong thing.

“Moving from an AI system that says the wrong thing to an AI system that does the wrong thing is a qualitatively different risk management challenge,” Patience said. “Most companies, and probably most vendors, don’t yet have the capacity to handle this at scale.”

Either way, AI agents will continue to dominate analyst chats. AMD CEO Lisa Su told CNBC earlier this week that agents are driving huge demand in the AI ​​cycle.

“Agent development is not a side project; it is a theme of the company’s 2026 roadmap and represents a turning point from exploration to action,” said Craig Le Clair, principal analyst at Forrester.

Competition between Big Tech, frontier model companies, established software vendors and emerging startups will only intensify as companies race to develop money-making AI tools, said Arjun Bhatia, co-head of technology equity research at William Blair & Co.

“The agent war is very much underway,” he added.

Softbank Stocks posted their best day since 2020 earlier this week, with stocks up 18%, amid a broad tech-driven rally that saw Japan’s Nikkei stock average soar to an all-time high. One commentator said the company’s gains were amplified by Arm’s close relationship with AI research institute OpenAI.