BUSINESS

Trump has repeatedly claimed Iran’s military capabilities are crippled as the U.S. has deployed more forces to the Middle East.

Wall Street fears higher prices and slower growth will sink all stocks — but small-caps and housing hold their own.

Large language models are running into limits in domains that require an understanding of the physical world — from robotics to autonomous driving to manufacturing. That constraint is pushing investors toward world models, with AMI Labs raising a $1.03 billion seed round shortly after World Labs secured $1 billion.Large language models (LLMs) excel at processing abstract knowledge through next-token prediction, but they fundamentally lack grounding in physical causality. They cannot reliably predict the physical consequences of real-world actions. AI researchers and thought leaders are increasingly vocal about these limitations as the industry tries to push AI out of web browsers and into physical spaces. In an interview with podcaster Dwarkesh Patel, Turing Award recipient Richard Sutton warned that LLMs just mimic what people say instead of modeling the world, which limits their capacity to learn from experience and adjust themselves to changes in the world.This is why models based on LLMs, including vision-language models (VLMs), can show brittle behavior and break with very small changes to their inputs. Google DeepMind CEO Demis Hassabis echoed this sentiment in another interview, pointing out that today’s AI models suffer from “jagged intelligence.” They can solve complex math olympiads but fail at basic physics because they are missing critical capabilities regarding real-world dynamics. To solve this problem, researchers are shifting focus to building world models that act as internal simulators, allowing AI systems to safely test hypotheses before taking physical action. However, “world models” is an umbrella term that encompasses several distinct architectural approaches. That has produced three distinct architectural approaches, each with different tradeoffs.JEPA: built for real-timeThe first main approach focuses on learning latent representations instead of trying to predict the dynamics of the world at the pixel level. Endorsed by AMI Labs, this method is heavily based on the Joint Embedding Predictive Architecture (JEPA). JEPA models try to mimic how humans understand the world. When we observe the world, we do not memorize every single pixel or irrelevant detail in a scene. For example, if you watch a car driving down a street, you track its trajectory and speed; you do not calculate the exact reflection of light on every single leaf of the trees in the background. JEPA models reproduce this human cognitive shortcut. Instead of forcing the neural network to predict exactly what the next frame of a video will look like, the model learns a smaller set of abstract, or “latent,” features. It discards the irrelevant details and focuses entirely on the core rules of how elements in the scene interact. This makes the model robust against background noise and small changes that break other models.This architecture is highly compute and memory efficient. By ignoring irrelevant details, it requires much fewer training examples and runs with significantly lower latency. These characteristics make it suitable for applications where efficiency and real-time inference are non-negotiable, such as robotics, self-driving cars, and high-stakes enterprise workflows. For example, AMI is partnering with healthcare company Nabla to use this architecture to simulate operational complexity and reduce cognitive load in fast-paced healthcare settings. Yann LeCun, a pioneer of the JEPA architecture and co-founder of AMI, explained that world models based on JEPA are designed to be “controllable in the sense that you can give them goals, and by construction, the only thing they can do is accomplish those goals” in an interview with Newsweek.Gaussian splats: built for spaceA second approach leans on generative models to build complete spatial environments from scratch. Adopted by companies like World Labs, this method takes an initial prompt (it could be an image or a textual description) and uses a generative model to create a 3D Gaussian splat. A Gaussian splat is a technique for representing 3D scenes using millions of tiny, mathematical particles that define geometry and lighting. Unlike flat video generation, these 3D representations can be imported directly into standard physics and 3D engines, such as Unreal Engine, where users and other AI agents can freely navigate and interact with them from any angle.The primary benefit here is a drastic reduction in the time and one-time generation cost required to create complex interactive 3D environments. It addresses the exact problem outlined by World Labs founder Fei-Fei Li, who noted that LLMs are ultimately like “wordsmiths in the dark,” possessing flowery language but lacking spatial intelligence and physical experience. World Labs’ Marble model gives AI that missing spatial awareness. While this approach is not designed for split-second, real-time execution, it has massive potential for spatial computing, interactive entertainment, industrial design, and building static training environments for robotics. The enterprise value is evident in Autodesk’s heavy backing of World Labs to integrate these models into their industrial design applications.End-to-end generation: built for scaleThe third approach uses an end-to-end generative model to process prompts and user actions, continuously generating the scene, physical dynamics, and reactions on the fly. Rather than exporting a static 3D file to an external physics engine, the model itself acts as the engine. It ingests an initial prompt alongside a continuous stream of user actions, and it generates the subsequent frames of the environment in real-time, calculating physics, lighting, and object reactions natively. DeepMind’s Genie 3 and Nvidia’s Cosmos fall into this category. These models provide a highly simple interface for generating infinite interactive experiences and massive volumes of synthetic data. DeepMind demonstrated this natively with Genie 3, showcasing how the model maintains strict object permanence and consistent physics at 24 frames per second without relying on a separate memory module.This approach translates directly into heavy-duty synthetic data factories. Nvidia Cosmos uses this architecture to scale synthetic data and physical AI reasoning, allowing autonomous vehicle and robotics developers to synthesize rare, dangerous edge-case conditions without the cost or risk of physical testing. Waymo (a fellow Alphabet subsidiary) built its world model on top of Genie 3, adapting it for training its self-driving cars. The downside to this end-to-end generative method is the great compute cost required to continuously render physics and pixels simultaneously. Still, the investment is necessary to achieve the vision laid out by Hassabis, who argues that a deep, internal understanding of physical causality is required because current AI is missing critical capabilities to operate safely in the real world.What comes next: hybrid architecturesLLMs will continue to serve as the reasoning and communication interface, but world models are positioning themselves as foundational infrastructure for physical and spatial data pipelines. As the underlying models mature, we are seeing the emergence of hybrid architectures that draw on the strengths of each approach. For example, cybersecurity startup DeepTempo recently developed LogLM, a model that integrates elements from LLMs and JEPA to detect anomalies and cyber threats from security and network logs. 

A mailer from Think Big PAC told voters that the Democratic U.S. House candidate once got $100,000 in support from the former head of failed global exchange FTX.

Jared Drinkwater transformed a snack idea into a fast-growing business with products sold in thousands of stores.

The yearly thresholds for achieving financial security range from roughly $62,700 for one-person households over age 65 to $169,300 for families with two or more adults and three or more kids.

TheStreet aims to feature only the best products and services. If you buy something via one of our links, we may earn a commission.Why we love this dealIf you’re excited for the warm, sunny spring weather and want a super comfortable and relaxing way to lounge, whether poolside or on the back deck, a luxurious patio set is sure to create the atmosphere you want. A lot of the patio sets or lounge chairs are comfortable, but when the sun gets too hot or you want to prevent sunburns, having an overhead canopy to help protect you while you lounge can make you feel like royalty. That’s why the Costway 2–Person Rattan Patio Daybed caught our attention. The lounge chairs and the adjustable overhead canopy make it a solid option for anyone who loves spending time in their backyard. Originally $399, shoppers can now get this daybed for just $280 and save 30% at Walmart.Costway 2–Person Rattan Patio Daybed, $280 (was $399) at Walmart

Courtesy of Walmart

Shop at WalmartWhy do shoppers love it?This daybed not only has a retractable canopy that offers comfortable shade for a nice outdoor nap, but it also features built-in side tables to hold your glass of iced tea, your favorite book, a plate of snacks, or a speaker, which can also be folded up or down as needed. The canopy can easily be adjusted with the pneumatic rod, and can be tied down with fixing ropes when not in use. The seating area has two seat cushions and four back cushions, offering a comfortable, plush area to lie down and relax. The cushions feature removable zipper covers that can be washed. Related: Macy’s is selling a charming $323 rattan porch swing for $102This daybed can hold up to 800 pounds, with the tables holding up to 10 pounds each. It features extra support legs and anti-slip footpads for safety, and the durable polyurethane rattan is designed for outdoor use. The retractable side trays measure 100 inches by 7.5 inches, and the daybed measures 58 inches long by 48.5 inches wide, with a 49.5-inch-wide canopy, offering room for two people to lie down, or three to four people to sit and chat. The pros and cons of a $280 patio daybedProsVersatility: This lounger is great for super sunny, windy, or cloudy weather. The canopy blocks wind and hot sun, or can offer a comfortable sunbathing area, and it can be used poolside, on the deck, or in the backyard grass. Weight capacity: The 800-pound weight capacity offers a sturdy, worry-free lounging experience.The daybed offers lots of room: The size allows you to fully lie down and relax by yourself or with others.Cons Color: The bright red color may not match your current outdoor decor.Care: The cushions should be covered during rainstorms to prevent wear and tear.One reviewer said, “I’m so happy with my purchase! It’s very sturdy and looks beautiful. You will love this lounger.” Another person said, “I love everything about it. The red color is perfect, it’s very nice! The cushions are good, and it’s comfy.”Shop more dealsCostway Outdoor Double Daybed with Canopy, $280 (was $399) at WalmartCostway Patio Convertible Sofa, $210 (was $399) at WalmartCostway 2-Piece Rattan Lounge Chair Set, $176 (was $349) at WalmartWhether you love to spend time lounging outdoors or want a comfortable place to relax with friends, the Costway 2–Person Rattan Patio Daybed is a solid choice. The sturdy metal frame holds up to 800 pounds, and the canopy is easy and convenient to use, whether you want to be out in the sun or prefer a bit of shade. The tables provide room for your phone and a drink, and the set comes with ample cushions for comfort. Shoppers can save $119 on this purchase, paying $280 for the daybed at Walmart.

Oil prices settled higher Friday, with global benchmark Brent crude posting a fifth straight weekly gain, after the Pentagon said it is sending more troops and warships to the Middle East.

One of the major sticking points on the crypto market structure bill may be resolved, at least enough to move toward a Senate hearing to advance the bill.

For decades, it seemed like the sky was the limit for Salesforce (CRM), the cloud-based enterprise software and customer relationship management platform.The company was founded in 1999 by Marc Benioff, Parker Harris, Dave Moellenhoff, and Frank Dominguez, who built their platform in a tiny apartment on San Francisco’s Telegraph Hill (a closet functioned as their server), proving they could innovate from the get-go.Salesforce made its product accessible to businesses of all sizes by offering its CRM as a subscription service over the internet (SaaS) rather than as an expensive, locally installed software, which was more common at the time.As a result, businesses were finally able to integrate their sales, marketing, customer service, and analytics together for a “360-degree” view of their customers — and Salesforce took off.Since its IPO on June 23, 2004, CRM shares have soared by nearly 2,000%.According to Benzinga, if an investor had bought $1,000 of CRM stock 20 years ago, those shares would be worth $20,797.11 in March 2026.But Salesforce’s skies have darkened recently, due to its struggles to prove that its investments in AI are delivering tangible ROI.In fact, reports have surfaced that company insiders don’t even understand how to use its new technology, let alone explain it to customers.What’s going on with Salesforce stock?Rewind to last year, when the turbulence started. On February 26, 2025, Salesforce reported $37.9 billion in fiscal 2025 revenue, which was a 9% increase from the previous fiscal year.Related: Does Apple pay dividends? A history of rewarding shareholdersHowever, its revenue growth estimates slowed to 8.5%, down from 11% the prior year.In addition, Salesforce’s margins dropped to 27.2% from 29.4% in 2024. This suggested that the company’s AI investments had added significant costs to the company — without increasing its bottom line.As a result, CRM shares sold off and have yet to recover: Salesforce ended the year down 20% in 2025.What’s going on with Salesforce’s layoffs?Salesforce’s turmoil stems from its company-wide shift to “Agentic AI.” Through its Agentforce platform, launched on September 12, 2024, AI agents can take on tasks that were once delegated to people. These agents are considered more intuitive than chatbots, and they need less human oversight, as well.On the heels of its rollout, Salesforce posted strong quarterly results. CNBC reported that the company had secured 200 deals for the product and thousands more in the pipeline. CRM shares closed at an all-time high of $365.07 on December 4, 2024.Everyone, it seemed, was aboard the Agentic bandwagon.But the machine’s efficiencies came at a very human cost.In January 2025, Benioff told Bloomberg that Salesforce’s AI agents were already doing 50% of the company’s work.That fall, Benioff went on the Logan Bartlett Show and explained that he was reducing his workforce because “I need less heads.”Related: How many employees does Salesforce have in 2026? CRM’s workforce, locations & layoffsSalesforce cut 4,000 customer service jobs that month and laid off an additional 1,000 marketing employees in early 2026, amounting to around 5% of its total workforce.But while losing a job feels insurmountable on a personal level, for companies like Salesforce, reduced administrative costs increase AI’s return on investment.And it isn’t just Salesforce, either; CEOs across the board have felt the heat in proving their AI investments work, Fortune said.In a January 2026 survey of 3,700 business executives, 61% of CEOs said they were under “increasing pressure” to demonstrate returns on their AI investments.But Salesforce may have already redeemed itself with its latest earnings report, released on Feb 26, 2026, which touted $800 million in annual recurring revenue for its Agentforce system, and 2.4 billion Agentic work units to date — all of which is proof of enterprise adoption.CEO Marc Benioff added that Agentic AI “is a tailwind for our business.”So, is CRM’s long-term growth story likely to continue?Salesforce will invest $15 billion in San Francisco over the next five years, a move the company said will cement the city’s status as a global hub for artificial intelligence https://t.co/HIXASbevqz— The Wall Street Journal (@WSJ) October 13, 2025

Is CRM a good buy now?Salesforce certainly thinks so. The fact remains: Salesforce is still one of the largest enterprise software companies in the world, a niche it basically created, and it boasted $38 billion in fiscal 2025 revenue, a $3 billion increase over the year before.The company has big plans for the future, and dynamic leadership from Benioff could very well make his vision of $60 billion in revenue by 2030 a reality.“We’re leading the next great transformation in business — the era of the Agentic Enterprise — where AI elevates human potential and accelerates growth,” Benioff said, counting Dell, FedEx, Pandora, PepsiCo, and Williams-Sonoma, Inc. among its users.In September 2025, the company launched the newest version of its AI enterprise, Agentforce 360, which promises even deeper platform integration. At the same time, Salesforce pledged to invest $15 billion into AI initiatives around San Francisco.This money will help to establish an AI Incubator Hub and add to workforce development programs.In addition, on March 16, 2026, Salesforce announced a massive $50 billion share repurchase program, half of which is funded by debt — another bullish sign from the company.Company histories:History of Microsoft: Company timeline & factsHistory of Coca-Cola: Timeline, facts & milestonesHistory of Nike: Company timeline and factsThe consensus estimates from analysts compiled by MarketBeat were that this move now makes the stock a “Moderate Buy.”And with shares trading around 14.7 times forward earnings, many even believe the stock isn’t overhyped but rather undervalued.Only time will tell, but Salesforce’s transition into an AI-focused platform may also have changed its business model from a high-flying growth story to a profitable, cash-generating enterprise that appeals to those who can patiently wait for artificial intelligence to prove itself in dollars and cents.Related: Who owns Salesforce in 2026? A look at its largest shareholders & leadership stake