Despite NVIDIA raking in nearly $148 billion in revenue from its AI chip business in the nine months ending October, a massive increase from the $27.5 billion in the same period of 2023, a crucial attempt by the company to transition into an integrated software-hardware platform has suffered a major setback. As CEO Jensen Huang's core bet to enter the $50 trillion manufacturing and logistics market, its Omniverse software business has yielded minimal returns after four years of heavy investment, with its commercialization process severely lagging. According to insiders, due to demand being "almost non-existent" since its 2022 launch, NVIDIA shut down the Omniverse Cloud service in August 2025. The company had previously spent hundreds of millions of dollars leasing thousands of GPUs from Oracle, Google, and Microsoft to support the project, but for most of that time, lacking external customers, it was forced to scramble to find internal uses to avoid idle chips. Although Jensen Huang will continue to heavily promote the business at this week's CES, portraying it as a multi-trillion dollar "physical AI" opportunity, internally, he is frustrated with the department's slow progress. Media reports citing sources state that Huang has repeatedly criticized the team for wasting engineering resources on "demos" rather than "products," and the business has not been widely adopted by large enterprises as expected. Although the explosive growth of the AI chip business has so far prevented the market from reacting violently to Omniverse's revenue shortfall, this exposes the difficult challenge the hardware giant faces in trying to open a second growth curve. If issues with software usability and industry adaptation cannot be resolved, NVIDIA's ambitions in robotics and industrial digitalization may face long-term delays. Demand evaporated and cloud services were shut down. NVIDIA officially launched Omniverse in 2021, positioning it as a platform for designers to collaborate on 3D designs in a virtual space, and attempted to expand it into a core tool for building "digital twins" of cars, factories, and robots. Huang's vision was to push companies to simulate before physical manufacturing, thereby driving tens of billions in chip purchases and up to $4,500 per chip annually in software licensing revenue. However, reality has fallen far short of the vision. According to former NVIDIA employees, although the company listed clients including BMW, Siemens, Foxconn, and Boston Dynamics, very few actually signed up to use Omniverse Cloud servers for large-scale simulations. Software developers widely reported the platform as "difficult to use," incomplete, and prone to crashing. Software developer Valentin Forager stated that when trying to simulate human activity in a virtual environment, the system simply crashed; "as soon as you try to do anything slightly outside its predefined scope, this thing breaks." Furthermore, the platform's scene creation tools were complex to operate and had outdated documentation, making many technical issues difficult to fix. At an event last November, even an NVIDIA representative admitted the software was not yet ready to meet specific needs, suggesting customers instead use competitor Unity's software. This lack of product maturity directly led to the termination of the cloud service project. Internal anxiety and management pressure. Although Omniverse's revenue is minuscule within NVIDIA's vast financial statements, Jensen Huang has shown clear anxiety over its performance. People familiar with his thinking reveal he has long worried that competitors might seize the initiative, thus pushing the company hard to find new revenue streams. This frustration has erupted repeatedly in internal meetings. According to attendees, at an all-hands meeting in 2023, Huang publicly questioned Omniverse Vice President Rev Lebaredian on whether the team had achieved profitability. In another conference call, faced with a request to add staff for new product development, Huang angrily berated the team for nearly an hour, accusing them of wasting engineer time on demos and emphasizing that the existing headcount was sufficient. The actual implementation of partnership projects has also sparked management dissatisfaction. Media reports cite sources stating Huang was furious upon learning that the scale of the announced partnership with BMW was far smaller than he initially expected. Although NVIDIA executives had claimed over 20,000 BMW planners globally used the software, they did not detail how much of this translated into actual sales. Long-term bet faces industry barriers. In the face of skepticism, NVIDIA executives compare Omniverse to CUDA—which took over a decade of investment to revolutionize deep learning. Rev Lebaredian stated that Omniverse is the foundational software for unlocking the huge market of "physical AI," and the long-term vision is now paying off through adoption by robotics and automotive companies. To foster ecosystem growth, NVIDIA not only had Huang's children join the department but also invested in companies like Synopsys (which recently acquired Ansys) and MetAI, attempting to integrate semiconductor design and industrial simulation tools. However, the field faces intense competition and structural barriers. In robotics simulation, Unity Technologies and the open-source simulator Gazebo are strong competitors. More critically, many large enterprises like Tesla prefer developing in-house simulation software rather than relying on NVIDIA's general-purpose platform. Industry-specific technical hurdles are another major challenge. Co-founders of two robotics startups pointed out that NVIDIA's Isaac Sim tool is impractical for handling complex objects with changing shapes, like clothing. Additionally, cost-effectiveness hinders adoption; Dan Cole, COO of industrial automation design firm Loupe, bluntly stated that leasing cloud servers is not cost-effective. As MetAI CEO Daniel Yu remarked, Omniverse is currently just a horizontal, open platform for developers to build on, not a complete application, meaning NVIDIA's attempt to create a market from scratch still requires a long cultivation period.
Comments