The Morning Download: Staying Small Is Silicon Valley's New Flex

Good morning. Startups building AI products and services aren’t getting bigger—at least by head count. That’s because the latest trend to emerge from Silicon Valley is the “lean AI startup,” a company that is small, but supercharged by AI-based coding software and other tools.

Some of those startups told me they’re tracking “revenue per headcount,” a relatively new metric that measures the amount of revenue generated by each employee.

In true Silicon Valley fashion, there’s even a leaderboard that tracks the most efficient AI startups—and showing up on it has become a bragging right for some founders.

“I did some research and found a bunch of companies who all had more than five million in [annual recurring revenue], were less than five years old, under 50 employees and growing rapidly,” said Henry Shi, the creator of the leaderboard. Shi is co-founder of Super.com—itself a lean AI startup—and a member of Anthropic’s technical staff.

The Silicon Valley fad toward small teams mirrors that of corporate America, where some CEOs have been singing the praises of AI tools and their ability to make employees more efficient. That’s accompanied by a leaner new normal for employment in the U.S., plus deep cuts and stagnant growth in white-collar positions.

For business leaders, the lean startup movement may signal that the new management flex isn't scaling your team, but proving you don't need to.

The lean ethic is helping flight-tracking app Flighty keep its five employees productive, founder and CEO Ryan Jones told me. The company uses AI tools in nearly every aspect of its operations, from coding to customer support, design, copywriting and market research, Jones said.

And for Linear, which builds a project management tool for software development, the focus on staying small helps the 120-person startup stay focused, co-founder and CEO Karri Saarinen told me. “There has been an idea that companies subscribe to, that we need to have more people to go faster. Now, more people are questioning it,” Saarinen said.

But some entrepreneurs say staying lean has led to risks like hampering growth, especially when selling to enterprises who still demand a human touch. It’s not possible—yet—to sell to large companies without dedicated human salespeople, some founders say.

So unless the enterprise sales cycle dramatically changes, these lean AI startups can’t be fully lean and mean if they want to close big deals.

Have you adopted the “lean” mentality at your company, and is it aided by AI tools? Drop us a line and let us know.

Pure Storage is relaunching with a new name, a broader focus and an acquisition. 

Pure Storage said today that it has rebranded itself as Everpure, capping a yearslong expansion beyond memory to "AI-ready" data management.

The company also announced that it has entered into a definitive agreement to acquire data intelligence and orchestration company 1touch.io, which provides a comprehensive and unified view of an enterprise’s information.

Everpure, which on Friday had a market cap of $24.4 billion and a PE ratio of 195, is scheduled to release its fourth quarter earnings on Wednesday.

It reported in December that it beat third-quarter expectations, but Barron’s said the stock fell amid “plans to reinvest AI hyperscaler revenue into research and development, and sales and marketing, potentially impacting 2027 profit margins.”

Pure Storage, launched in 2009, with a focus on high-speed, all-flash memory, which is based on chips instead of spinning disks. The company went public in 2015, trading on the NYSE. It’s based in Santa Clara, Calif.

In 2017, CEO Scott Dietzen was succeeded by Charles Giancarlo, a Silicon Valley veteran who held senior roles at Cisco Systems spanning deal-making and technology. He later became a managing director and operating partner at private equity firm Silver Lake.

The relaunch occurs as the entire data infrastructure world is retooling for AI.

“Everpure reflects the company we have become as we help enterprises unleash the full power of their data,” Giancarlo said in a statement. “With 1touch, we are taking the next step in helping organizations not only gain control of their most valuable asset—data—but also understand, enhance, and contextualize that data for actionable intelligence.”

Everpure said its cloud architecture transforms storage into “a unified, virtualized cloud of data, governed by an intelligent control plane.” It seeks to eliminate the “friction of manual configurations.”

Everpure said New York-based 1touch will extend its capabilities by adding data discovery and classification and semantic context, making enterprise data “AI ready at the source” and capable of generating insights faster.

— Steven Rosenbush

Call it the AI immunity trade. After a three-year love affair with anything related to artificial intelligence, U.S. investors are flocking to companies that have seemingly strong odds of surviving the technological revolution intact, WSJ reports. In the past month, the S&P 500 sectors for industrials, materials, utilities and consumer staples have surged ahead of the overall index.

AI comes for the real-estate broker. Commercial real-estate firms are facing investor skepticism over whether AI will erode their brokerage fees and commissions, the WSJ reports. Shares of major players like CBRE and JLL have nosedived despite strong earnings. Industry executives argue that complex deals built on relationships and proprietary intelligence can't be replicated by technology, but analysts broadly agree that smaller, standardized deals are vulnerable. 

AI startups are juicing valuations. AI startups increasingly are using a tiered funding tactic where a lead investor buys shares at a lower valuation, while subsequent investors pay a much higher price. The result: the startup gets to publicize an inflated "headline valuation."

Critics say the practice artificially inflates valuations. The growing prevalence of tiered funding deals signals a subtle shift in AI investment dynamics, suggests the Journal. While founders once held most of the negotiating power, investors can now structure more favorable terms for themselves, suggesting the market has cooled slightly.

Nvidia is entering the consumer laptop market with its chips set to appear in Dell and Lenovo PCs as early as this year, the WSJ reports. Through partnerships with Intel and Taiwanese chip designer MediaTek, Nvidia aims to bring its GPU and AI capabilities to everyday PCs in a system-on-a-chip design, standard in smartphones but not yet widespread in PCs — promising thinner, lighter laptops with better battery life. 

The hollowing out of the CISA. The Cybersecurity and Infrastructure Security Agency, the agency Trump created in his first term to defend against cyberattacks, has lost more than a thousand employees since January 2025, the New York Times reports.  A government shutdown has furloughed most of those who remained. 

Mexico’s military killed the country’s most powerful drug kingpin, Nemesio “Mencho” Oseguera, escalating the government’s war against cartels and sparking a widespread, violent gang response. (WSJ)

From Berlin to Tokyo, the world’s biggest exporting countries have reacted to U.S. tariffs by further committing to economic policies that support exports, subsidizing manufacturers to help them leap over the tariff wall. (WSJ)

A massive winter storm bearing down on the Northeast and mid-Atlantic put states on emergency measures Sunday and upended air travel, with airlines canceling thousands of flights ahead of the arrival of blizzard conditions. (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.