Another Round of Layoffs Points to the Growing Tension Within the Tech Industry

🔦 Spotlight

Happy Friday, LA.

The largest checks in tech are increasingly going toward companies trying to build their way out of America’s biggest physical constraints.

This week, two Los Angeles startups raised a combined $2.37 billion in equity to tackle two particularly urgent ones: how the country manufactures critical hardware and where it will find enough electricity to power the AI era.

Torrance-based Hadrian is building highly automated factories for defense and aerospace. El Segundo’s Valar Atomics wants to manufacture nuclear reactors at scale. Different industries, same underlying bet: the next generation of technology will depend on our ability to produce physical infrastructure much faster than we do today.

Hadrian raised $1.37 billion in Series D funding, bringing its valuation to $7.87 billion. The company plans to use the capital to open new factories, expand research and development, and increase its capacity to produce critical defense, aerospace and industrial systems.

Hadrian’s pitch is straightforward, if wildly ambitious: America needs to relearn how to build things and build them quickly.

Its factories combine skilled workers with AI, robotics and proprietary software to manufacture precision components and, increasingly, complete mission-critical systems. Its customers include defense giants such as Lockheed Martin and RTX, along with newer players like Anduril.

The company has come a long way from simply making aerospace parts. Hadrian is positioning itself as a piece of America’s industrial infrastructure, offering manufacturers a way to rapidly scale domestic production at a time when wars abroad, strained supply chains and growing defense demands have made the country’s manufacturing gaps increasingly difficult to ignore.

Investors are clearly buying the argument. The new round comes just over a year after Hadrian raised $260 million, suggesting that “reindustrialization” has officially graduated from venture capital buzzword to billion-dollar investment thesis.

Meanwhile, roughly 15 miles away in El Segundo, Valar Atomics is moving even faster than its enormous ambitions suggested.

When we last wrote about Valar, the company was reportedly raising $450 million at a $2 billion valuation and racing to prove that nuclear energy could move on AI’s timetable. Now, it has closed a $1 billion Series B led by Sequoia Capital, secured an additional $200 million credit facility and reportedly reached a $6 billion valuation.

Valar is developing standardized, factory-built nuclear power plants designed to avoid the enormous costs and decades-long construction timelines associated with traditional nuclear projects. Its goal is not merely to build a working reactor, but to eventually manufacture fleets of them.

That ambition also sounds considerably less theoretical than it did when we first covered the company. In June, Valar’s Ward 250 reactor achieved a self-sustaining nuclear reaction. Just one week later, the company demonstrated the reactor generating electricity to power an Nvidia Blackwell system. Valar now says the new funding will help it move from proving its technology works to producing reactors at scale.

The timing is no coincidence. AI’s enormous appetite for electricity is forcing the tech industry to confront a basic reality: the cloud still has to plug into something. Training models and operating massive data centers will require far more reliable power, and nuclear energy is rapidly becoming one of Silicon Valley’s favorite answers.

Hadrian and Valar may be solving different problems, but their unusually large rounds point to the same shift. AI can design, predict and automate, but it cannot manufacture a missile component or generate a megawatt of electricity on its own. That requires factories, energy systems, supply chains and a great deal of capital.

For years, venture-backed companies competed to build the software layer. Now, some of the biggest bets are being placed on the infrastructure underneath it.

The future may run on AI. But first, someone has to build what keeps it running.

More from this week’s LA startup and venture scene below.

🤝 Venture Deals

Download the dot.LA App

🔦 Spotlight

Hello LA,

This week’s startup story began three years ago with an AI assistant built for almost any small business. It ended, or perhaps graduated, with an acquisition by one of the fitness industry’s largest technology providers.

ABC Fitness has acquired Replify, an AI platform that manages customer communication for gyms and wellness businesses across phone, text, email and chat. Its virtual agents can answer questions, qualify leads, schedule tours and classes, follow up on missed calls and run outbound campaigns. Financial terms were not disclosed.

Before Replify found its footing in fitness, it was HeyLibby, a general-purpose AI assistant founded in 2023 by former Zillow colleagues Spencer Rascoff, Tony Small and Anna Rodriguez. The company was incubated inside Rascoff’s 75 & Sunny Labs and initially set out to help small businesses turn incoming messages into qualified leads.

That broad vision gave HeyLibby a large potential customer base, from real estate agents and contractors to hairstylists and event planners. But as the team searched for product-market fit, one industry’s problem stood out. Gym and wellness employees were often too busy helping customers in person to answer every call, text or email, leaving prospective members waiting and potential revenue on the table.

That insight reshaped the company. HeyLibby narrowed its focus to fitness and wellness, raised a $4.5M seed round in 2025 and later rebranded as Replify. It went on to work with brands including Gold’s Gym and UFC Gym, proving that its AI agents could do more than answer routine questions. According to the company, customers have captured up to 10 times more leads and shortened sales cycles from roughly 30 days to as little as three to five days.

ABC Fitness became a natural next step. The company provides software to more than 30,000 fitness businesses serving over 40 million members worldwide. By adding Replify to its platform, ABC can offer gyms an always-available AI front desk while bringing Replify’s technology to a much larger global customer base.

Replify’s journey offers a useful lesson amid the rush to build AI products for everyone. The company began with a broad promise, identified a customer with a specific and expensive problem, and built deeply around that need. Gym owners did not need another flashy chatbot. They needed someone to answer the phone when the front desk could not.

Sometimes the smartest AI strategy is simply picking up the call.

More from this week’s LA startup and venture scene below.

🤝 Venture Deals

Download the dot.LA App

🔦 Spotlight

Hello LA,

Nine years after his turbulent exit from Uber, Travis Kalanick is back with a new company, an enormous war chest and, apparently, some unfinished business.

Los Angeles-based Atoms announced this week that it has secured a $1.7 billion equity investment led by Andreessen Horowitz, with a16z cofounder Ben Horowitz joining its board. Bain Capital, Fifth Wall, Uber and several other investors participated, while a roster of major banks, including Goldman Sachs, JPMorgan and Bank of America, are listed as debt partners.

Yes, Uber itself is now backing the comeback of its famously ousted cofounder. Silicon Valley may preach disruption, but it has always appreciated a good redemption arc.

Atoms is the culmination of the company Kalanick has spent the past eight years building largely out of public view. Formerly known as City Storage Systems, the parent company behind CloudKitchens, it is now bringing its businesses together under one ambitious umbrella: Atoms Food, Atoms Mining and Atoms Transport.

The premise is that AI’s next major frontier will not be confined to screens, chatbots or software. Atoms wants to build what Kalanick calls a “computer for the physical world,” using software, sensors, robotics and AI to automate how physical goods are produced, stored and moved.

That means tackling decidedly unglamorous but enormous industries such as mining, construction, food production and heavy transportation. Rather than betting on humanoid robots that can theoretically do everything, Atoms is focused on specialized machines designed to perform specific, economically useful jobs.

In other words, the robot does not need a face. It needs a business model.

For a16z, the investment is as much a bet on Kalanick as it is on industrial AI. In an essay bluntly titled “Travis Is Back,” Horowitz argues that Kalanick possesses the rare mix of technical range, endurance and sheer force of will required to drag old-line industries into a new technological era. The firm’s broader thesis is that robotics will eventually handle much of the repetitive work involved in making, moving and storing physical goods, creating a market potentially as consequential as computing itself.

There is also some history being settled. Kalanick, Horowitz and Marc Andreessen nearly partnered during Uber’s early days but never completed the deal. In a new conversation about Atoms, Kalanick and Horowitz revisit that missed opportunity and the long road that brought them back together. Sixteen years later, the check is considerably larger.

The scale of the investment is remarkable, but so is its location. Atoms is headquartered in Los Angeles, giving the city a front-row seat to one of tech’s boldest industrial AI bets. It also reinforces something increasingly evident across LA’s startup ecosystem: the next era of AI will not only be written in code. It will be built in kitchens, warehouses, mines, vehicles and factories.

Whether Atoms becomes the operating system for the physical world or simply proves that even $1.7 billion cannot make atoms behave like bits remains to be seen. But Kalanick is taking another enormous swing, and this time, Los Angeles is where the comeback story begins.

More from this week’s LA startup and venture scene below.

🤝 Venture Deals

Download the dot.LA App