## When AI Meets the Physical World

What happens when you hand an AI agent the keys to a real robotic arm? The answer, it turns out, is surprisingly impressive. A recent experiment explored exactly that — using an AI coding agent called OpenClaw alongside OpenAI's Codex to configure, operate, and even train a physical robot arm from scratch.

The AI successfully set up the hardware, used an onboard camera to perceive its surroundings, and gradually learned to grasp objects. It even helped train a secondary AI model to pick up and place specific items. If that sounds like science fiction, it increasingly isn't.

## The Robot: A LeRobot 101

The hardware at the center of this experiment is the **LeRobot 101**, a prebuilt robotic arm that's part of an open-source initiative from HuggingFace. The project is designed to lower the barrier to entry for robotics experimentation, making it more accessible and affordable than traditional industrial systems.

The LeRobot system includes two arms working in tandem:

- A **controller arm**, operated manually by a human using a handle and trigger
- A **follower arm**, fitted with a camera, that mirrors the controller's movements

This setup enables teleoperation-based AI training. As a person moves the controller arm, the AI model observes the follower arm's camera feed and learns to replicate those movements autonomously over time.

## Building With AI: Triumphs and Growing Pains

Getting started wasn't entirely smooth. Before leveraging OpenClaw, the experimenter spent several hours attempting to connect and calibrate the robot manually — at one point nearly burning out the motors by applying incorrect settings that caused them to overheat.

Once OpenClaw and Codex entered the picture, progress accelerated significantly. The AI agent navigated the complex process of establishing connections to the robot hardware, assisted with joint calibration, and generated a Python script that combined multiple libraries to detect and grip a red ball on command.

This approach — sometimes called **vibe coding** — isn't flawless. AI-generated code can produce hallucinations and introduce bugs, particularly when dealing with specialized hardware. But the overall results were striking enough to turn heads.

### From Simple Gripping to Model Training

Grabbing a red ball was just the beginning. The next phase involved using OpenClaw to guide the training of a dedicated control model for the arm. The AI proved adept at walking through each stage of the training process and evaluating the model's error rate after every training run — a task that would typically require significant robotics expertise.

## The 'Code as Policy' Revolution

The concept powering this experiment has an academic name: **code as policy**. First introduced in a landmark 2022 research paper, the idea proposes that AI-generated code can serve as a versatile and powerful method for programming and controlling robots — one that sidesteps the need for highly specialized robotics knowledge.

Since that paper was published, AI coding capabilities have advanced at a remarkable pace, and the code-as-policy approach has gained serious momentum across research institutions worldwide.

### A New Benchmark: CaP-X

A collaborative research team from UC Berkeley, Nvidia, Carnegie Mellon University, and Stanford has developed a new evaluation framework called **CaP-X** to measure how well AI coding models can handle robotic programming tasks.

The findings are revealing. According to CaP-X results, **Google's Gemini** outperforms both Claude and ChatGPT when it comes to robot programming — likely because Google DeepMind has heavily prioritized multimodal training, equipping its models with a stronger grasp of physical-world reasoning.

Alongside the benchmark, the team released:

- **CaP-Gym** — a controlled environment for testing coding agents on both simulated and real robots
- **CaP-Agent0** — an agentic framework that significantly boosts coding model performance, enabling them to outperform some models trained specifically for direct robotic motion control

## Industry Leaders Are Taking Notice

Ken Goldberg, a prominent roboticist at UC Berkeley and one of the researchers behind this work, frames the significance clearly:

> *"AI-powered coding is super exciting because it has the potential to bridge the gap between conventional engineering methods, which are reliable but don't generalize, and contemporary vision-language-action models, which generalize but are not yet reliable."*

Goldberg's team is also collaborating with Nvidia to expand the practical reach of the code-as-policy method, working to make it compatible with a broader ecosystem of robot software tools.

Spencer Huang, who has been organizing internal robotics hackathons at Nvidia and is currently involved in a joint research project with Goldberg, believes the implications extend far beyond the lab.

> *"Nearly anyone can get into robotics, which is the true holy grail,"* Huang explains. Enabling people to direct robots through spoken commands, typed instructions, or simple physical demonstrations is, in his words, the **"critical unlock for robots in society."**

## What This Means for the Future of Robotics

The barriers that once made robotics the exclusive domain of specialists — deep programming knowledge, costly hardware, and years of training — are beginning to erode. AI-powered coding tools are emerging as a genuine equalizer, enabling enthusiasts, researchers, and developers to experiment with physical robots in ways that were unimaginable just a few years ago.

Whether this signals an imminent robotics breakthrough or simply an exciting step in a longer journey, one thing is clear: the relationship between artificial intelligence and the physical world is growing closer, faster, and more capable with every passing year.

An AI coding agent configured, operated, and trained a physical robot arm. Here's what the experiment revealed about the future of robotics and AI.

AI Agent Controls a Real Robot Arm: What Happened

large_AI-Lab-Rock-Em-Sock-Em-Vibe-Coded-Robots-Business.jpg

medium_AI-Lab-Rock-Em-Sock-Em-Vibe-Coded-Robots-Business.jpg

small_AI-Lab-Rock-Em-Sock-Em-Vibe-Coded-Robots-Business.jpg

thumbnail_AI-Lab-Rock-Em-Sock-Em-Vibe-Coded-Robots-Business.jpg

AI-Lab-Rock-Em-Sock-Em-Vibe-Coded-Robots-Business.jpg

How I Gave an AI Agent Full Control of a Robotic Arm — And What Happened Next

## Hackers Target Open Source Ecosystem in Sweeping Supply Chain Attack

A sophisticated and ongoing cyberattack has successfully compromised numerous widely-used open source software packages, putting developers and organizations around the globe at serious risk. Cybersecurity researchers are raising urgent alarms as the attack continues to evolve.

### What Is Happening?

On Tuesday, leading cybersecurity firms StepSecurity and SafeDep issued warnings about a fresh wave of so-called supply chain attacks. These attacks work by targeting the developers behind popular open source projects, then exploiting that access to inject malicious code into software updates that eventually reach end users downstream.

According to findings from SafeDep, attackers seized control of at least one developer's account and proceeded to release more than 630 compromised versions across 317 separate packages — all within a staggering 20-minute window. The primary objective of this campaign is to harvest credentials from various services, including password managers, enabling the attackers to steal sensitive data and further propagate the malware.

### High-Profile Packages Affected

Among the compromised packages is Antv, a well-known library developed by Chinese tech giant Alibaba. Security researchers at JFrog Security also confirmed that in several instances, the hackers published malicious updates directly through GitHub, one of the world's most widely used software development platforms.

### The Mini Shai-Hulud Campaign

This latest wave of intrusions is part of a broader, coordinated hacking campaign that researchers have named **Mini Shai-Hulud** — a name derived from its connection to a preceding, larger-scale attack operation. The campaign specifically targets open source projects and the developers who integrate that code into their own applications and services.

In a particularly alarming development last week, hackers operating under this same campaign managed to compromise the workstations of two OpenAI employees. The breach was made possible after attackers successfully infiltrated the popular open source library **TanStack**. OpenAI, however, was just one of multiple organizations affected in that wave of attacks.

### Why Supply Chain Attacks Are So Dangerous

Supply chain attacks are especially concerning because they exploit the inherent trust developers place in open source libraries and tools. Rather than attacking a single target directly, cybercriminals compromise a shared resource used by thousands of developers simultaneously, dramatically amplifying the potential damage.

Organizations that rely on open source components — which today includes virtually every major software company — are urged to audit their dependencies immediately, monitor for suspicious package updates, and implement strict verification protocols for any third-party code integrated into their projects.

### What You Should Do

- **Audit your dependencies** regularly and look for unexpected version changes.
- **Enable multi-factor authentication** on all developer accounts and package registries.
- **Use verified security tools** to scan for malicious code within third-party libraries.
- **Stay informed** by following advisories from cybersecurity firms tracking this campaign.

As this attack remains active and ongoing, developers and security teams are strongly advised to remain vigilant and prioritize software supply chain security as a critical component of their overall cybersecurity strategy.

Hackers have compromised 317 open source packages in an ongoing supply chain campaign called Mini Shai-Hulud, targeting developers and organizations worldwide.

Supply Chain Attack Hits Dozens of Open Source Packages

large_code-on-a-screen.jpg

medium_code-on-a-screen.jpg

small_code-on-a-screen.jpg

thumbnail_code-on-a-screen.jpg

code-on-a-screen.jpg

Ongoing Supply Chain Attack Compromises Dozens of Popular Open Source Packages

## What Makes Tempo Meals Worth Trying?

At first glance, Tempo meals might remind you of old-school TV dinners — but don't let appearances fool you. Unlike traditional frozen meals, Tempo's offerings have never been frozen at any point. Thanks to cutting-edge packaging technology, each meal arrives fresh, flavorful, and surprisingly close to home-cooked quality.

Dishes like beef barbacoa and sous-vide chicken genuinely impressed with their moisture retention and tenderness — qualities that rarely survive the heat-and-eat process. Even the green beans held up well, staying vibrant in color with a satisfying snap. Meals are prepared in a professional commissary kitchen and delivered directly to your door in specialized packaging designed to maintain freshness without relying on artificial preservatives.

While no reheated meal will fully replicate something freshly made, Tempo comes surprisingly close — especially when warmed in an air fryer.

---

## Why Tempo Stands Out for High-Protein and GLP-1 Plans

Tempo has built a strong reputation in the health-conscious meal delivery space, particularly for its **high-protein** and **GLP-1 support plans**, both developed in collaboration with registered dietitians. These plans are specifically designed to help individuals preserve muscle mass during calorie-restricted diets or the natural metabolic shifts that come with midlife.

Regular pricing falls between **$11 and $13 per serving**, but a range of Tempo Meals discount codes makes the service far more accessible — especially for first-time customers.

---

## Current Tempo Promo Codes and Discounts

### Save 30% Off Your First Month of Four Boxes

Like many subscription meal services, Tempo uses introductory offers to attract new customers and give them a genuine taste of the experience. New subscribers can currently access a **Tempo Meals coupon code** that delivers meaningful savings across the first month.

Depending on the active promotion, this could mean:

- A flat **30% discount** across the first four weeks of deliveries
- **50% to 60% off** the first box, followed by reduced pricing on subsequent boxes in the first month
- A **lifetime free breakfast item** added to your subscription

For returning customers who previously paused their subscription, Tempo also sends **exclusive re-engagement promo codes** directly to your inbox. In some cases, these discounts are applied automatically when you revisit the Tempo website.

---

### Heroes Discount: 50% Off Now and 10% Off for Life

Most Tempo promotional codes are limited to introductory periods, but there's one exception worth highlighting. Tempo's **Heroes Discount** is a permanent benefit available to:

- Active and veteran military personnel
- First responders
- Teachers
- Doctors, nurses, and other healthcare workers

To unlock this ongoing discount, eligible individuals must verify their status through the **ID.me verification system**. Once confirmed, they receive **50% off initially and a lasting 10% discount** on every future order — making it one of the best long-term value options for anyone committed to a structured high-protein meal plan.

---

## How to Apply Your Tempo Meals Coupon at Checkout

Tempo frequently runs multiple promotional offers simultaneously, and activating them is straightforward. Here's how the process typically works:

1. **Visit the Tempo website** and look for any available pop-up promotions to activate your coupon.
2. **Enter your zip code and email address** to confirm that Tempo delivers to your area.
3. **Answer a few preference questions** about your dietary goals and meal tastes.
4. **Choose your weekly meal quantity** — plans range from **6 to 20 meals per week**.
5. **Review your checkout summary** to ensure the correct discount code has been applied, which may include free first-week shipping or a complimentary bonus item like a free breakfast.

Double-checking the applied discount before completing your purchase ensures you're getting the maximum savings available to you.

---

## Is Tempo Worth the Investment?

For anyone looking for a convenient, nutritionist-backed meal solution that doesn't compromise on quality, Tempo presents a compelling option. With rotating promo codes offering up to **60% off your first box**, the barrier to entry is low enough to make trying it a no-brainer — especially if you're focused on protein intake, weight management, or simply reclaiming time in the kitchen.

Use a Tempo Meals promo code to save up to 60% on fresh, high-protein ready-to-heat meals. Explore all current discounts, hero deals, and how to redeem them.

Tempo Meals Promo Code: Get Up to 60% Off in 2026

large_WIRED-Coupons-R2_3.png

medium_WIRED-Coupons-R2_3.png

small_WIRED-Coupons-R2_3.png

thumbnail_WIRED-Coupons-R2_3.png

WIRED-Coupons-R2_3.png

Tempo Meals Promo Code: Save Up to 60% on Ready-to-Heat High-Protein Meals in 2026

## OSHA Opens Probe Following Fatal Incident at SpaceX's Starbase

Federal safety regulators have launched a formal investigation after a worker lost their life at SpaceX's Starbase launch facility located in South Texas. The incident, which occurred in the early morning hours of May 15, adds yet another chapter to an already concerning record of workplace safety issues at the high-profile site.

According to reporting by the San Antonio Express-News, the unidentified worker died at approximately 4:17 a.m. local time. The fatality was later confirmed by the local county sheriff to The Wall Street Journal. The Occupational Safety and Health Administration (OSHA) has since confirmed to media outlets that it is actively investigating the apparent accident.

Requests for comment directed to the Brownsville police and fire departments went unanswered. Neither SpaceX nor the newly incorporated city of Starbase issued any statement in response to media inquiries.

## Details Remain Scarce as Investigation Gets Underway

The exact circumstances surrounding the worker's death have not yet been made public. OSHA has indicated it will withhold further details until its investigation concludes — a process that may take several months to complete.

The timing of the tragedy is notable. The fatality occurred just days before SpaceX had planned the inaugural launch of its upgraded Starship rocket. The company is also reportedly preparing to release a detailed prospectus for what is expected to be the largest initial public offering in history, anticipated to take place the following month.

## A Pattern of Safety Concerns at Starbase

The Starbase facility, which serves as both an active construction zone and the primary testing and launch site for SpaceX's Starship program, has long been associated with elevated safety risks.

An analysis of OSHA data conducted in 2025 revealed that the Texas facility's injury rate significantly exceeded those of industry competitors and ranked as the most hazardous among all SpaceX operational sites. A separate investigative report by Reuters in 2023 brought to light dozens of previously unreported injuries at SpaceX's McGregor, Texas, test site, including a worker fatality that had occurred back in 2014.

### Recent Violations and Penalties

Earlier this year, OSHA issued SpaceX seven citations classified as "serious" safety violations. These stemmed from, among other failures, the company's failure to conduct proper inspections of a crane that subsequently collapsed at Starbase in June of the prior year. Regulators imposed the maximum allowable financial penalty on six of the seven violations, bringing the total fine to $115,850. SpaceX has since filed to contest these penalties, according to federal records.

### Worker Lawsuits and Ongoing Legal Battles

SpaceX has also faced a wave of civil litigation connected to injuries sustained at the Starbase site. In December, a subcontractor employee named Eduardo Cavazos filed suit after being struck by a large metal support that fell from a crane. Cavazos sustained serious injuries, including a broken hip, knee, and tibia. OSHA subsequently opened a rapid response investigation into the incident.

However, that investigation was later closed without any punitive measures being taken, according to public records. The lawsuit itself was also dropped, as Cavazos' attorney noted that his employer — the subcontractor — carries workers' compensation insurance that shields the company from direct legal action.

## Safety Scrutiny Intensifies Around Starbase

With SpaceX continuing to expand its ambitions in commercial spaceflight and Starship development, the recurring safety incidents at Starbase are drawing increased attention from regulators, workers, and the public alike. The latest fatality is likely to intensify calls for more rigorous oversight at a facility that sits at the center of some of the most ambitious aerospace projects in history.

OSHA is investigating a worker's death at SpaceX's Starbase site in Texas, a facility with a history of safety violations and the highest injury rate among Spac

OSHA Investigates Worker Death at SpaceX Starbase Texas

large_starship-super-heavy-spacex.jpg

medium_starship-super-heavy-spacex.jpg

small_starship-super-heavy-spacex.jpg

thumbnail_starship-super-heavy-spacex.jpg

starship-super-heavy-spacex.jpg

OSHA Launches Investigation After Worker Dies at SpaceX Starbase Facility in Texas

## Graduation Speakers Are Getting Booed for Praising AI — and Students Aren't Sorry

For anyone planning to deliver a commencement address in the coming months, here's a piece of advice worth taking seriously: think twice before singing the praises of artificial intelligence. A growing number of graduating students aren't buying the hype — and they're making that crystal clear in real time.

## When AI Enthusiasm Met a Very Loud Crowd

During this year's commencement season, at least two high-profile speakers discovered just how unwelcome AI cheerleading has become among the cap-and-gown crowd.

Gloria Caulfield, a senior executive at real estate development firm Tavistock Development Company, took the stage at the University of Central Florida and opened with broad strokes about living through a period of "profound change" — both thrilling and intimidating, she acknowledged. Then came the fateful line: "The rise of artificial intelligence is the next industrial revolution."

The audience responded with a wave of boos that grew louder by the second. A visibly caught-off-guard Caulfield turned to her fellow speakers with a laugh and asked, "What happened?" She quickly collected herself, noting, "Okay, I struck a chord," before attempting to continue. Her next sentence — "Only a few years ago, AI was not a factor in our lives" — was swiftly drowned out, this time by sarcastic cheering and applause from the crowd.

## Eric Schmidt Faced Criticism Before He Even Spoke

Former Google CEO Eric Schmidt encountered a similarly frosty reception at the University of Arizona's commencement ceremony. His troubles, however, started well before he stepped up to the podium.

Several student groups had already organized efforts to have Schmidt removed as a speaker following a lawsuit in which a former girlfriend and business partner accused him of sexual assault — allegations he has firmly denied. By the time Schmidt actually took the stage, the booing had reportedly already begun.

Things didn't improve once he started speaking. When Schmidt told graduates, "You will help shape artificial intelligence," the crowd erupted in sustained disapproval. Rather than pausing, Schmidt pushed through, telling students that AI agents could help them accomplish things they never could on their own, and adding, "When someone offers you a seat on the rocket ship, you do not ask which seat, you just get on." The audience wasn't moved.

## Not Every AI Mention Has Bombed

To be clear, AI hasn't become toxic at every graduation stage across the country. Nvidia CEO Jensen Huang recently addressed Carnegie Mellon University's graduating class and spoke freely about how AI has "reinvented computing" — without drawing any notable pushback.

So the backlash isn't universal. But it is increasingly visible, and it speaks to something deeper than just tech fatigue.

## What's Really Behind the Frustration

The tension on these campuses reflects a broader anxiety gripping young Americans. According to a recent Gallup poll, only 43% of Americans between the ages of 15 and 34 believe it's currently a good time to find a job in their local area. That's a dramatic fall from 75% who felt that way back in 2022.

While economic pessimism has multiple causes, technology journalist and industry critic Brian Merchant argued that for many in this generation, AI has come to represent something more troubling than just automation. In his words, it has become "the cruel new face of hyper-scaling capitalism."

"I too would loudly boo at the prospect of this next industrial revolution if I was in my early twenties, unemployed, and had aspirations for my future greater than entering prompts into an LLM," Merchant wrote.

Even speeches that avoided the topic of AI directly tended to lean heavily on themes of "resilience" this graduation season — an implicit acknowledgment of the difficult landscape awaiting new graduates. Schmidt himself recognized this in his remarks, saying that many young people today feel "a fear that the future has already been written, that the machines are coming, that the jobs are evaporating, that the climate is breaking, that politics are fractured, and that you are inheriting a mess that you did not create."

## Reading the Room Matters

In Caulfield's case, observers noted that her speech may have been poorly matched to her audience even before AI entered the picture. She was addressing graduates from arts and humanities programs — a crowd unlikely to respond warmly to corporate cheerleading. One student noted that Caulfield had already begun losing the room with what they described as "generic" praise of business figures like Jeff Bezos.

Graduate Alexander Rose Tyson summed up the mood plainly when speaking to The New York Times: "It wasn't one person that really started the booing. It was just sort of like a collective, 'This sucks.'"

The message from the class of 2025 seems clear enough: if you want to inspire the next generation, lead with empathy — not algorithms.

Graduation speakers are getting loudly booed for praising AI. Here's why students are pushing back and what it says about the generation entering the workforce.

Why AI Talk at 2026 Commencement Speeches Gets Booed

large_university.jpg

medium_university.jpg

small_university.jpg

thumbnail_university.jpg

university.jpg

Why Mentioning AI at 2026 Graduation Speeches Is a Recipe for Boos

## From Lonely Thesis to Billion-Dollar Vindication

Back in 2015, betting on the physical world was not exactly a crowd-pleaser in Silicon Valley. When Lior Susan launched Eclipse Ventures and began championing the digitization of real-world industries, most investors were chasing enterprise software and SaaS plays.

"It felt fairly lonely the first couple of years," Susan admitted during a recent StrictlyVC event held in San Francisco.

Fast forward to today, and Eclipse Ventures is no longer on the fringe — it is squarely at the center of one of technology's most exciting investment moments.

## The Cerebras Payday That Validated Everything

Eclipse's conviction was put to the test in 2016, when the firm led a $6.5 million Series A round in Cerebras Systems, a semiconductor company building specialized AI chips. Over the years, Eclipse continued backing Cerebras, deploying a total of $147 million into the company.

When Cerebras went public this week at an IPO price of $185 per share, that long-term commitment paid off in a dramatic fashion — generating a total return of $2.5 billion, representing a roughly 17-fold multiple on invested capital.

For Susan, however, the Cerebras exit is not the main event. It is proof of concept for a thesis he has held for over a decade: that 85% of global GDP is rooted in the physical world, and that companies operating at that intersection of hardware, software, and real-world infrastructure represent one of the greatest untapped opportunities in venture capital.

## Why Software Moats Are Eroding

Susan's optimism about physical-world technology is not purely about semiconductors. It is also a pointed commentary on what is happening to pure-play software businesses.

Public market investors have grown increasingly skeptical of traditional SaaS models, with many software stocks experiencing notable selloffs earlier this year amid concerns that AI tools — such as Anthropic's Claude Code and OpenAI's latest models — could enable enterprises to build their own custom software solutions in-house.

Susan put it bluntly: "I think people understand that the real moat in software is gone. You can vibe code pretty much whatever you want."

But there are clear limits to what AI-assisted coding can replace.

"What you cannot do with 'vibe code' is manufacture wafers, because you need machines and silicon, and they need clean rooms, and a bunch of other things," Susan said.

This is the fundamental argument at the heart of Eclipse's investment philosophy — physical complexity creates durable competitive advantages that no amount of AI-generated code can replicate overnight.

## A Portfolio That Is Catching Fire

The numbers backing Susan's thesis are hard to ignore. Eclipse's portfolio companies — spanning robotics, energy, defense, and semiconductors — raised nearly $15 billion from outside investors in the past year alone. In just the first quarter of 2026, that figure reached $4.5 billion, a striking contrast to the firm's first eight years, during which its portfolio companies raised less than $4 billion in total.

Some of the most significant recent funding rounds include:

- **$1.2 billion** raised by Wayve, an autonomous driving AI company
- **$650 million** secured by True Anomaly, a space technology firm
- **$270 million** closed by Bedrock Robotics
- **$200 million** raised by Oxide Computer

Notably, Eclipse was the Series A investor in all four of these companies — a track record that any venture firm would be proud to claim.

## Five Forces Aligning for a Historic Moment

Susan argues that the current surge of interest in physical-world technology is driven by more than just AI hype. He identifies five key forces that must align for this market to truly flourish: technology, capital, customer demand, talent, and policy.

For the first time, Susan believes all five are working in concert. Investors and engineers are pivoting away from SaaS toward sectors like robotics, semiconductors, space, and critical minerals. Simultaneously, the U.S. government is offering subsidies and shaping regulatory frameworks to actively encourage domestic growth in these industries.

"This is the first time I believe in America ever, from Henry Ford and Carnegie, those five forces are aligned," Susan said. "For builders like us, this is the best time to build those companies."

Public market signals support this optimism as well. Shares of semiconductor giants TSMC and Micron recently reached all-time highs, and a new generation of elite founders is increasingly drawn to startups that blend hardware innovation with software intelligence.

## The Bigger Picture

What Eclipse Ventures represents is a broader recalibration of where value is truly created in the technology economy. As AI commoditizes software development and erodes the barriers that once protected SaaS businesses, the companies that build, manufacture, and operate in the physical world are emerging as the next frontier for venture-scale returns.

For Lior Susan and Eclipse, the $2.5 billion Cerebras outcome is a landmark moment — but if their thesis holds, it may ultimately be remembered as just the opening chapter.

Eclipse Ventures earned a $2.5B return on Cerebras Systems' IPO. Here's why founder Lior Susan believes physical-world tech is the next great investment frontie

Eclipse Ventures' $2.5B Cerebras Win and the Physical-World 

medium_55242803156_4c25fab693_c.jpg

small_55242803156_4c25fab693_c.jpg

thumbnail_55242803156_4c25fab693_c.jpg

55242803156_4c25fab693_c.jpg

How Eclipse Ventures Turned a $147M Bet Into a $2.5B Cerebras Windfall — and Why the Physical World Is Just Getting Started

## A Desk Gadget That Actually Holds You Accountable

Working from home sounds ideal until the reality sets in — your back aches, your water bottle sits untouched, and hours pass without a single stretch. Plenty of apps promise to fix these habits, but they're easy to swipe away and even easier to ignore. Hardware, on the other hand, has a way of demanding attention.

After nearly a decade of refining a home office setup — ergonomic chairs, lumbar cushions, adjustable monitors — one thing remained elusive: consistently good posture. That changed when I started using **Isa**, a desk device built by Deep Care, a startup founded by three former Bosch engineers based in Germany.

## What Makes Isa Different

Isa isn't just another wellness gadget. It monitors posture, hydration habits, ambient light, sound levels, and physical movement — all without a camera and without an internet connection. In an age where smart devices routinely harvest user data, that offline-first design philosophy is a genuine selling point.

The device resembles a compact table clock, featuring a **5.5-inch IPS HD display** and powered through a standard USB-C connection at roughly 2.45W — low enough that any existing charger will do the job.

## The Technology Inside

### Sensors That Do the Heavy Lifting

At the heart of Isa is a **Time-of-Flight (ToF) 3D depth sensor** — the same underlying technology found in facial recognition systems and modern smartphone cameras. This sensor handles posture analysis and movement detection across a range of 0.15 to 1.8 meters, meaning it can register whether you're seated, slouched, or standing at your desk.

Beyond the primary sensor, Isa is packed with additional monitoring hardware:

- **ToF 1D sensor**
- **Gyroscope**
- **Barometer**
- **Ambient light sensor**
- **Sound level sensor**
- **CO₂ and VOC sensor**
- **Temperature and humidity sensor**

All of this processing is handled by a quad-core 2 GHz processor built into the device. Wi-Fi connectivity is available exclusively for software updates and can be disabled at any time.

### How It Displays Information

Isa communicates your posture status through a **squircle ring** on its screen — a rounded square that fills or depletes based on how well you're sitting. When your posture deteriorates, the ring shifts from green to yellow or red. It's a subtle visual cue that proved surprisingly effective in practice; seeing that color change triggered an almost instinctive urge to sit up straight.

A separate widget styled like a water tank tracks your hydration, using the ToF sensor in beta mode to count drinking events. Another widget monitors whether you've been stationary too long, prompting you to stand and guiding you through brief on-device exercises. Once you return to your desk after a break, the movement tracker resets automatically.

The device also vibrates when you've been slouching past a set threshold — a gentle but firm reminder that works better than any phone notification.

## Setup and Usability

Getting Isa up and running is relatively painless. The device walks you through entering basic details about yourself and your daily work schedule. One notable gap: at launch, Isa only supports EU and US time zones, leaving users in Asia and other regions without proper local time support. The company has acknowledged this limitation, and broader time zone coverage would be a straightforward improvement worth prioritizing.

## Where It Falls Short

Because Isa relies entirely on depth-sensing rather than a camera, it's occasionally fooled by its environment. Objects placed between you and the sensor — a water bottle, a stack of books — can be misread as a person, logging false stationary periods. Pets or household members walking through the sensor's range can trigger similar errors.

Isa typically recognizes when you've left your workspace and switches to a standard clock display, but there's no manual override button to tell the device you've stepped away. Adding a simple physical or on-screen shortcut for this would significantly improve accuracy.

Occasionally, the device flagged prolonged inactivity after less than 30 minutes of sitting — a minor frustration but not a dealbreaker. Overall, these are manageable quirks rather than fundamental flaws.

## Pricing and Subscription Model

Isa is priced at **€299 (approximately $354)** and comes with two subscription tiers:

- **Core Plan — €4.99/month:** Posture tracking, healthy sitting habits, hydration detection, and access to the exercise library
- **Pro Plan — €7.99/month:** Everything in Core, plus monitoring of light levels, noise, and CO₂ for a healthier work environment

Deep Care initially sold Isa exclusively to businesses before opening sales to individual consumers — a move that reflects growing confidence in the retail wellness hardware market.

## What's Coming Next

Deep Care has signaled ambitions beyond posture correction. The company plans to leverage Isa's sensor suite for **mental health and stress tracking**, using signals like posture shifts, head movement, and chest movement patterns to estimate breathing rhythms. Combined with environmental data — noise levels, lighting, CO₂ concentration — the goal is to introduce a stress-related wellness score. Whether that feature lands well will depend heavily on accuracy and user trust.

## Final Verdict

Isa is not the cheapest solution for improving your desk habits. Between the upfront hardware cost and the ongoing subscription, it's an investment that demands genuine commitment to workplace wellness. But for remote workers who have tried and abandoned apps, stretching reminders, and ergonomic accessories without lasting results, Isa offers something more persistent and harder to ignore.

It's one of the more thoughtfully designed desk wellness devices available today — and if you're serious about protecting your posture and building healthier movement habits, it's worth serious consideration.

Deep Care's Isa tracks posture, hydration, and movement offline without a camera. Here's what makes this $350 desk gadget stand out for remote workers.

Isa Desk Device Review: Offline Posture Tracker Worth $350?

large_Deepcare-feat.jpg

medium_Deepcare-feat.jpg

small_Deepcare-feat.jpg

thumbnail_Deepcare-feat.jpg

Deepcare-feat.jpg

Meet Isa: The Privacy-First Desk Device That Finally Fixed My Posture

## Amazon Cuts Support for Older Kindles — and Users Are Pushing Back

When Amazon announced it would be ending technical support for a range of older Kindle devices effective May 20, many owners assumed their trusty e-readers would soon become expensive paperweights. Instead, a growing number of them are taking matters into their own hands — by jailbreaking their devices.

The decision by Amazon means that affected Kindle models will no longer receive software updates or be able to sync new purchases. Users will be limited to reading content already stored on the device. For avid readers who rely on their Kindles daily, that's simply not an acceptable outcome.

## What Does Jailbreaking a Kindle Actually Mean?

Jailbreaking refers to the process of removing the software restrictions that Amazon builds into its Kindle ecosystem. Once those restrictions are bypassed, users gain the ability to:

- Install custom fonts and personalized screensavers
- Access alternative reading applications
- Use third-party tools that significantly expand the device's capabilities
- Continue downloading and managing new reading content

For many Kindle owners, the motivation is straightforward: they paid for a device that still works perfectly well, and they shouldn't be forced to buy a replacement simply because Amazon has decided to move on.

## The Risks You Need to Understand Before You Start

Jailbreaking is not without its dangers, and anyone considering it should go in with eyes wide open.

### Device Compatibility Is Not Guaranteed

Not every Kindle model or firmware version supports the jailbreaking process. Before attempting anything, owners must research whether their specific device and software version are compatible. Skipping this step could spell disaster.

### Things Can Go Wrong — Seriously Wrong

The most significant risk is what the tech community calls "bricking" — essentially rendering your device completely non-functional. If something fails during the jailbreaking process, you may end up with a device that does nothing at all, which is arguably worse than a device with limited functionality.

### Performance and Battery Life May Suffer

Running unofficial or third-party software on a jailbroken Kindle can introduce instability. App crashes, sluggish performance, and noticeably reduced battery life are all real possibilities that users should factor into their decision.

> **Disclaimer:** This article is intended for informational purposes only. Always ensure your actions comply with applicable laws and terms of service.

## A Safer Alternative: Sideloading Books via USB

For those who want to continue adding books to their older Kindle without taking on the risks associated with jailbreaking, there is a simpler option. You can sideload content directly onto your device using a standard USB cable connected to a computer — no system modifications required. This method allows you to transfer compatible e-book files without touching the device's core software.

## Considering a New Device? These Are Worth a Look

If jailbreaking feels too risky and sideloading doesn't fully meet your needs, the e-reader market has expanded well beyond Amazon's ecosystem. A few notable alternatives include:

- **Boox Palma** — A versatile Android-based e-reader with broad app support
- **Vivlio E-Reader** — A solid option popular in European markets
- **Xteink X3** — A compact, innovative device designed to attach directly to the back of your smartphone

Each of these offers a fresh start without the dependency on Amazon's support cycles.

## The Bigger Picture

The wave of Kindle jailbreaking activity reflects a broader conversation in the tech world about device ownership and the right to repair. When a company ends support for a product, users who invested money in that product are left to decide: upgrade, adapt, or find creative workarounds. For a significant portion of Kindle owners, jailbreaking represents a form of digital self-reliance — a refusal to let a perfectly functional piece of hardware go to waste.

Amazon is ending support for older Kindle models. Learn why users are jailbreaking their devices, the risks involved, and safer alternatives to consider.

Kindle Jailbreaking Guide After Amazon Ends Device Support

large_2007-kindle-1st-gen-angle-gettyimages-93559279.jpg

medium_2007-kindle-1st-gen-angle-gettyimages-93559279.jpg

small_2007-kindle-1st-gen-angle-gettyimages-93559279.jpg

thumbnail_2007-kindle-1st-gen-angle-gettyimages-93559279.jpg

2007-kindle-1st-gen-angle-gettyimages-93559279.jpg

Kindle Owners Fight Back: Why Thousands Are Jailbreaking Their Devices After Amazon Drops Support

## GoPro Joins the Growing List of Companies Pivoting to Defense

If you want to capture investor attention in today's market, the playbook seems increasingly clear: build data centers, develop energy storage solutions, or pivot to defense. GoPro, the iconic action camera brand, appears to be following that third path — and doing so while simultaneously exploring whether it should sell itself entirely.

## A Market Trend That's Hard to Ignore

Across the technology and manufacturing sectors, companies are racing toward opportunities that align with government spending and infrastructure demand. Ford's fledgling energy-storage division sent its stock surging to multi-year highs despite being a small fraction of Tesla's operation. Redwood Materials secured $425 million from powerhouse investors including Google and Nvidia after repositioning toward data center energy storage. Cerebras delivered one of the most talked-about IPOs of 2026. And defense-focused startup Anduril recently closed a staggering $5 billion funding round.

The message from investors and public markets alike is unmistakable: companies with even a plausible shot at landing government contracts should be pursuing them aggressively.

## GoPro's Long Road From Darling to Distress

GoPro was once the undisputed king of the action camera market. Throughout the 2010s, competitors and tech commentators repeatedly proclaimed the arrival of so-called "GoPro killers" — from TomTom's action cameras to Google's short-lived Clips device — yet none managed to unseat the California-based pioneer. The company not only survived but helped define an entirely new product category.

However, survival and prosperity are two very different things. In recent years, GoPro's financial picture has grown increasingly grim. Revenue has fallen steadily, losses have widened, and the company's stock price stagnated at around $1 for the better part of two years.

## The Defense Pivot and Its Mixed Results

Last month, GoPro announced it would formally "explore defense and aerospace market opportunities." On the surface, the strategy isn't entirely far-fetched. The company's cameras are known for exceptional image quality and rugged durability — hardware that has survived motorcycle crashes and even a fall from the edge of space. Those attributes translate reasonably well to military and surveillance applications.

The announcement delivered a short-term boost, nearly doubling GoPro's stock price for a brief window. But that rally quickly faded, suggesting investors remain skeptical about whether the defense pivot represents a genuine transformation or simply a headline-driven bump.

## Hiring a Bank and Weighing a Sale

The situation took a more serious turn when GoPro announced it had engaged investment bank Houlihan Lokey to evaluate a "potential sale and other strategic alternatives." The company's board of directors disclosed that it had received "several unsolicited inbound strategic inquiries from parties across various sectors including defense, consumer and financial" — language that signals the company is at a genuine crossroads.

This isn't the first time a sale has been discussed internally. Founder and CEO Nick Woodman acknowledged in 2018 that the idea had briefly come up. But the circumstances today are considerably more urgent.

## Layoffs and a Shrinking Workforce Add to the Pressure

Beyond the financial deterioration, GoPro recently announced it is cutting approximately 25 percent of its workforce. The layoffs are particularly striking given how much the company has already contracted — its headcount has fallen to fewer than 600 employees after once reaching as high as 1,500. For a brand that was considered a tech darling just 15 years ago, the decline is stark.

## Navigating an Uncertain Future

GoPro's situation reflects a broader reality facing many consumer technology companies that thrived in the smartphone era but now find themselves squeezed between slowing demand and intensifying competition. With the Pentagon's budget continuing to expand, the allure of defense contracts is understandable — even for a company best known for capturing surfing videos and ski runs.

Whether GoPro can successfully reposition itself for defense customers, find a suitable buyer, or chart some other path forward remains to be seen. What's clear is that the company that once inspired a generation of adventure filmmakers is now fighting a very different kind of battle.

GoPro is exploring defense and aerospace opportunities while hiring a bank to evaluate a potential sale amid falling sales and significant layoffs.

GoPro Explores Defense Pivot and Potential Sale

large_gopro-hero-11.png

medium_gopro-hero-11.png

small_gopro-hero-11.png

thumbnail_gopro-hero-11.png

gopro-hero-11.png

GoPro Eyes Defense Contracts and a Potential Sale as Its Business Struggles to Stay Afloat

## From Film School to the AI Frontier

Runway doesn't fit the standard Silicon Valley mold. There are no Stanford dropouts among its founders, no former Google engineers, and no massive seed funding that allowed the company to sidestep revenue pressure for years. Instead, the three co-founders — two from Chile and one from Greece — crossed paths at NYU's Tisch School of the Arts and quietly built something remarkable from New York City.

Yet despite its unconventional origins, Runway has emerged as one of the most closely watched artificial intelligence companies in the world — not simply for what it has already created, but for the far more ambitious vision it is now pursuing.

## Why Video, Not Language, May Be the Real Path to AI Intelligence

For years, the dominant assumption in AI development has been that intelligence is rooted in language. Flagship models like OpenAI's ChatGPT and Anthropic's Claude were built on that foundation, trained on vast oceans of human-written text.

Runway is placing a fundamentally different wager. Its founders believe the next leap in artificial intelligence won't emerge from text at all — it will come from video and so-called world models, systems capable of learning how physical reality actually operates rather than simply how people talk about it.

That distinction may sound theoretical. Its consequences are anything but.

### The Case Against Language-Only AI

Runway co-founder and co-CEO Anastasis Germanidis argues that training AI models on observational data from the real world represents the next true frontier of the field. The companies that achieve it first, he contends, won't be those that mastered language.

"We're basically bound by our own understanding of reality," Germanidis explained during a conversation at Runway's sunlit headquarters near New York's Union Square. "Language models are trained on the entire internet — message boards, social media, textbooks — distilling existing human knowledge. But to get beyond that, we need to leverage less biased data."

## Runway's Journey: From Creative Tool to Scientific Infrastructure

Founded in 2018, Runway initially made its name developing AI-powered video generation tools, including its current flagship model, Gen-4.5. These tools allow users to transform simple text prompts into polished, editable cinematic content.

Today, Runway's technology is embedded in professional production pipelines at film studios and advertising agencies alike. The company has secured partnerships with major media organizations including Lionsgate and AMC Networks, and its tools have contributed to acclaimed productions such as *Everything Everywhere All At Once*.

The company is now valued at $5.3 billion, and according to one of its founders, generated $40 million in new annual recurring revenue during the second quarter of 2026 alone.

### Expanding Into World Models

Within the past six months, Runway has moved decisively beyond video generation. The company launched its first world model in December — AI systems sophisticated enough to simulate environments and predict how they will behave — and has additional releases planned for this year.

Runway is not the only player in this space. Startups Luma AI and World Labs are pursuing comparable goals, while Google has been developing its own Genie world model. Former Meta chief scientist Yann LeCun and AI pioneer Fei-Fei Li are also competing in this arena.

But Germanidis frames world models not merely as a product category — he sees them as foundational scientific infrastructure. The more diverse sensory data a single model is trained on, the closer it gets to functioning as a digital twin of the physical universe — a simulation you could run experiments on far faster than any physical laboratory allows.

"If we can build a better scientist than human scientists, we can accelerate progress in how we understand the universe and how we solve problems," Germanidis said.

## The People Behind the Vision

Germanidis first discovered programming at age 11 in Athens. He arrived in the United States at 18 to study neuroscience and film, before returning to computer science and working at several Silicon Valley companies — an experience that left him disillusioned with the culture.

Co-CEO Cristóbal Valenzuela grew up in Santiago, Chile, studying economics before transitioning into film and eventually software. Chief Innovation Officer Alejandro Matamala Ortiz, also from Santiago, studied advertising and ran his own design firm.

The three met in 2016 at NYU's Interactive Telecommunications Program — a graduate course Valenzuela described as "an art school for engineers." All three had harbored filmmaking ambitions at various stages of their lives.

### A Mission That Kept Evolving

Runway's original question was straightforward: could AI make everyone a filmmaker? After releasing their first video-generation model in early 2023, that mission sharpened into something more demanding — could AI make everyone a *great* filmmaker?

Pursuing that goal required expanding the team to 155 employees spread across offices in New York, London, San Francisco, Seattle, Tel Aviv, and Tokyo. Along the way, the founders realized the technology was revealing something larger about itself.

"Throughout this process, we learned that these models can understand how the world works," Matamala Ortiz noted. "And if you scale them, they can be useful for many other different things."

Those things now include robotics, drug discovery, and climate modeling. Runway launched a dedicated robotics division last year that Germanidis says has already produced real-world testing and deployments. His own long-term ambition for the technology extends to biological world models and anti-aging research.

## The Competition and the Challenges Ahead

Whether Runway can successfully carry its leadership in AI video generation into the world model era remains an open question. The competitive landscape has grown formidable.

Kian Katanforoosh, CEO of AI benchmarking firm Workera and a lecturer at Stanford, notes that no one has yet demonstrated a proven leap from video intelligence to generalized reasoning through world models — but he doesn't dismiss the possibility. He does, however, flag a critical practical concern: compute access.

"How are you going to build a foundational model without a cluster?" Katanforoosh asked pointedly. "I don't think anybody can do that."

Runway has partnerships with CoreWeave and Nvidia and has raised $860 million in total funding, including a $315 million round in February backed by strategic investors including AMD Ventures and Nvidia. That funding puts it broadly in line with direct competitors — Luma AI has raised approximately $900 million and World Labs $1.29 billion — but far behind the industry's deepest pockets.

### Google Looms Large

The most formidable threat on the horizon is Google. The tech giant's Veo model competes directly with Runway's core video generation business, while its Genie world model targets the same long-term territory Runway is racing toward. Alphabet, Google's parent company, carries a market capitalization of $4.86 trillion. OpenAI, another indirect competitor, has raised roughly $175 billion according to CEO Sam Altman.

For a company born out of an arts graduate program with no traditional Silicon Valley pedigree, the scale of that competition is stark. But Runway's founders have consistently framed their outsider status as a source of clarity rather than weakness — a perspective that has already taken them further than most expected.

Runway AI is betting that video generation leads to world models — challenging Google and redefining what artificial intelligence can become. Here's how.

Runway AI: From Filmmaking Tool to World Model Challenger

large_Runway-founders.jpg

medium_Runway-founders.jpg

small_Runway-founders.jpg

thumbnail_Runway-founders.jpg

Runway-founders.jpg

How Runway Is Challenging Google and Redefining the Future of AI

## Gantri's Bold New Vision for Wireless Lighting

California-based lighting company Gantri has taken a significant step forward in home design with the launch of a new wireless lamp collection developed in partnership with renowned design firm Ammunition. Built using corn-derived polylactic acid (PLA) plastic and manufactured in Gantri's Bay Area production facilities, these lamps combine sustainable materials with premium aesthetics — all at a price point that sits between $200 and $500.

## A Partnership Built on Design Excellence

Gantri and Ammunition first joined forces back in 2020, with the shared goal of elevating what thoughtfully designed lighting could look like. Ammunition is no stranger to iconic product design — the firm is responsible for the look of Beats by Dre headphones, Square's widely recognized point-of-sale tablets, and a range of other high-profile projects spanning everything from robotic coffee machines to Jay-Z's short-lived cannabis vape line.

Their latest collaboration expands on that original vision with a broader range of fixtures, including floor lamps, table lamps, and compact handheld models. The smaller designs draw inspiration from the piers along the San Francisco waterfront, giving them a distinctly regional character.

## Wireless Freedom With Long Battery Life

One of the most notable features of this new collection is that every lamp in the lineup is completely wireless. Each unit detaches from its charging base and offers what Gantri claims is a battery life of ten or more hours — enough to carry a lamp from one space to another without being tethered to a wall outlet.

Gantri CEO Ian Yang draws an interesting parallel to justify this design direction. For the vast majority of human history, portable light sources — torches, candles, lanterns — were carried from place to place. The idea of lights being permanently fixed to walls and ceilings is, historically speaking, a relatively recent convention. Yang believes this collection challenges that assumption.

"I really think this product is going to change the way that people think about lighting, but also think about the power of digital manufacturing, about this new material that's plant-based," Yang said.

## Smart Home Integration on the Horizon

Gantri is currently developing a companion app to give users greater control over their lamps. The company also plans to integrate Matter compatibility — the cross-platform smart home connectivity standard designed to allow devices from different brands to communicate seamlessly — though that feature is not expected to roll out until next year.

## A Custom Charging System by Design

Perhaps one of the more debated aspects of the new collection is its proprietary charging system. The lamps use a custom charging port that allows them to be positioned upright and rotated in any direction while still receiving power. However, this means they are not compatible with USB-C or any standard charging cable.

For some users, this could limit the freedom that a wireless lamp is supposed to offer — if you move a lamp to a different room, you need to bring the proprietary charger along with it. Yang acknowledges this trade-off but insists the decision was intentional.

"What we want is something that we can use every single day," Yang explained. "To do that, you have to build a habit into charging. We want to make this as simple and seamless as possible."

Beyond the home, Gantri also sees commercial potential in this design choice. Yang notes that the lamps could work well in restaurant settings, where proprietary chargers act as a mild deterrent against theft — making USB-C-powered alternatives far easier targets by comparison.

## Rethinking an Entire Category

For Yang, this product launch represents the fulfillment of a long-held ambition. When he began exploring the portable lighting market, he found most existing options lacking in substance — too small, too dim, and more decorative than functional.

"They're really tiny, more like a candle, an accent, a mood thing — it's not like a real light," Yang said, framing Gantri's entry into this space as "an opportunity for us to not just replace that one form factor, but actually to rethink the lighting category."

With a focus on sustainable manufacturing, considered design, and genuine portability, Gantri appears to be making a compelling case that the future of lighting doesn't have to be plugged in.

Gantri teams up with Ammunition to release wireless, plant-based 3D-printed lamps with 10+ hour battery life and future Matter smart home support.

Gantri Launches Wireless 3D-Printed Lamps With Ammunition

large_26-04-16-MOVEMENT-521_HR.jpg

medium_26-04-16-MOVEMENT-521_HR.jpg

small_26-04-16-MOVEMENT-521_HR.jpg

thumbnail_26-04-16-MOVEMENT-521_HR.jpg

26-04-16-MOVEMENT-521_HR.jpg

Gantri and Ammunition Team Up to Launch a New Line of Wireless 3D-Printed Lamps

## Cerebras Systems Makes a Stunning Market Debut in 2026's First Major Tech IPO

AI chip designer Cerebras Systems kicked off 2026 with one of the most dramatic stock market entrances in recent memory, raising $5.5 billion through its initial public offering and watching its share price more than double on the first day of trading.

## IPO Pricing Blows Past Expectations

Shares were priced at $185 on Wednesday evening — significantly above the company's originally projected range of $115 to $125, which had already been revised upward to $150 to $160. The company also expanded the offering to 30 million shares, signaling strong institutional confidence heading into the debut.

When the stock opened to public trading on Thursday, retail investors piled in aggressively, pushing the price to $385 — a staggering 108% premium over the IPO price. While the stock pulled back somewhat as the session progressed, trading above $330 at midday, it ultimately closed at $311. That closing price translated into a fully diluted market capitalization of approximately $66 billion, according to Yahoo Finance. After-hours activity suggested the stock had not lost its upward momentum.

## Founders Emerge as Billionaires Overnight

Even at the initial offering price of $185, Cerebras entered public markets with a fully diluted valuation of $56.4 billion. Co-founder and CEO Andrew Feldman saw his personal stake valued at nearly $1.9 billion, while co-founder and CTO Sean Lie held shares worth approximately $1 billion. With the stock trading well above $300, both executives — and early investors — are sitting on considerably larger gains.

## A Rocky Road to Wall Street

The path to this IPO was anything but smooth. Just a year ago, the prospects of Cerebras going public looked genuinely uncertain. The company, which designs oversized AI-specific chips from the ground up as an alternative to Nvidia's dominant hardware, had originally filed for a public listing in 2024. However, the process stalled due to scrutiny over a significant investment from Abu Dhabi-based Group 42, which triggered an extended national security review by the Committee on Foreign Investment in the United States (CFIUS). Adding to investor hesitation was the fact that Group 42 accounted for nearly all of Cerebras' revenue at the time — a concerning level of customer concentration.

Those IPO plans were ultimately shelved as the company worked to address both regulatory and financial concerns.

## Improved Financials Revived the IPO Push

Momentum shifted in April when Cerebras returned with a dramatically improved financial profile. The company reported revenues of $510 million for 2025, representing a 76% year-over-year increase, with that revenue now spread across a broader customer base. Even more compelling was the swing to profitability: Cerebras posted net income of $237.8 million, a remarkable turnaround from a loss of nearly half a billion dollars the previous year.

## A Growing Force in AI Inference

Cerebras has positioned itself as a serious competitor in the AI inference market — the continuous computing work required to process user queries and generate responses from AI models. Its current customer roster includes some of the most recognizable names in the industry: OpenAI, G42, Saudi Arabia's Mohamed bin Zayed University of Artificial Intelligence, and Amazon Web Services. The OpenAI relationship involves a notable circular arrangement that has drawn attention from analysts watching the company's revenue dynamics.

With a landmark IPO behind it and a growing list of enterprise clients, Cerebras Systems has firmly established itself as one of the most closely watched names in the global AI hardware race.

Cerebras Systems raised $5.5 billion in its 2026 IPO, with shares surging over 108% on debut day to reach a $66 billion market valuation.