---
title: "NVIDIA's ambitious layout in the digitalization of the manufacturing industry has yielded minimal returns so far"
type: "News"
locale: "zh-HK"
url: "https://longbridge.com/zh-HK/news/271634136.md"
description: "NVIDIA has achieved significant growth in its artificial intelligence chip business, with revenue reaching $148 billion in 2023. However, CEO Jensen Huang remains focused on expanding new businesses, particularly the Omniverse software, which aims to address issues in the manufacturing industry and enter a $50 trillion market. Despite high expectations, the commercialization of the software has progressed slowly, market returns are meager, and cloud services have been shut down due to insufficient demand. Jensen Huang will reiterate this vision at the upcoming Consumer Electronics Show"
datetime: "2026-01-06T10:00:48.000Z"
locales:
  - [zh-CN](https://longbridge.com/zh-CN/news/271634136.md)
  - [en](https://longbridge.com/en/news/271634136.md)
  - [zh-HK](https://longbridge.com/zh-HK/news/271634136.md)
---

> 支持的語言: [简体中文](https://longbridge.com/zh-CN/news/271634136.md) | [English](https://longbridge.com/en/news/271634136.md)


# NVIDIA's ambitious layout in the digitalization of the manufacturing industry has yielded minimal returns so far

In the past two years, NVIDIA's artificial intelligence chip business has soared, driving the company's revenue to nearly $148 billion in the nine months ending in October, far exceeding $27.5 billion in the same period of 2023. However, CEO Jensen Huang is not satisfied. Due to ongoing concerns about potential risks to the business, Huang has been urging the team to expand into new areas—selling chips and software to assist in robotics research and solving manufacturing challenges. This new business is expected to become a growth engine for the company in the future.

The aforementioned software **only supports running on NVIDIA chips**, assisting users in creating digital twin models and simulation tests for real-world objects such as cars and machinery. Huang has stated that this tool, named "Omniverse," is expected to help NVIDIA tap into the $50 trillion manufacturing and logistics market. However, according to four current and former employees familiar with the business, after four years of effort, NVIDIA's attempts to commercialize the Omniverse software have yielded little success.

**Key Points**

1.  The Omniverse software in the robotics field, which NVIDIA has high hopes for, has seen slow market returns to date.
2.  Due to almost no demand, NVIDIA has shut down the Omniverse cloud service.
3.  The Omniverse software is difficult to operate and prone to failures, causing dissatisfaction among developers.

Despite the slow progress, Huang continues to describe this business in public as a **trillion-dollar opportunity** for NVIDIA. He is expected to reiterate this vision at the annual Consumer Electronics Show in Las Vegas this week. At that time, Huang will deliver a keynote speech, and NVIDIA will hold multiple events to showcase its chips and software applications in the manufacturing and robotics sectors. Attendees will include the CEO of industrial engineering giant Siemens, the CTO of leading humanoid robotics company Agility Robotics, and executives from the two major design and simulation software providers, PTC and Ansys.

However, four individuals familiar with Huang's thoughts have stated that he is frustrated by the slow growth of the Omniverse business. Nevertheless, thanks to NVIDIA's explosive overall revenue growth, the shortcomings of the Omniverse business have not attracted much attention.

According to two individuals with direct knowledge, NVIDIA has spent hundreds of millions of dollars renting thousands of graphics processing units (GPUs) from Oracle, Google, and Microsoft to launch the Omniverse cloud service for large enterprises, helping companies conduct large-scale simulation testing in the cloud. However, former NVIDIA employees have stated that despite NVIDIA's claims that many companies, including BMW, Siemens, Foxconn, and Boston Dynamics, are using the Omniverse software, very few companies have signed contracts to use the Omniverse cloud servers for large-scale simulations According to informed sources, since the launch of the Omniverse cloud service in 2022, demand has been nearly zero until at least early 2025. NVIDIA has had to urgently find internal uses for these chips to avoid equipment idleness. A source with direct knowledge revealed that in August 2025, NVIDIA shut down this cloud service due to a lack of demand.

Some tools under Omniverse support customers in building three-dimensional scenes and object models, and simulating their operational status in a virtual environment, such as testing the operation of factory conveyor belts; other tools like DRIVE Sim support large-scale simulation calculations to aid in the development of autonomous vehicles; one tool named Isaac Sim allows developers to complete robot training in a virtual environment.

Many developers using Omniverse tools for scene and object modeling and simulation have reported that the software is not only difficult to operate and prone to crashes, but also has significant functional deficiencies.

Three former NVIDIA employees pointed out that one reason for this issue is that engineers in the Omniverse department are often busy developing new features and demonstration versions of new products—these versions run smoothly in controlled environments but cannot adapt to the needs of real application scenarios.

Two informed sources stated that in meetings with subordinates, Jensen Huang has criticized the Omniverse team for wasting engineers' time on demonstration projects that cannot be converted into actual products.

Jensen Huang's vision is that all companies can first conduct virtual simulation tests on robots, vehicles, and factories before physical production, thereby driving huge orders for NVIDIA chips—just as companies currently purchase NVIDIA chips to train and run artificial intelligence models. However, five informed sources revealed that Huang's patience with the slow implementation of Omniverse in large enterprises is gradually wearing thin.

**Long-term Bet**

The Omniverse business holds significant weight in Jensen Huang's mind, as both of his children work in the department: his daughter Madison serves as a senior manager in the Omniverse marketing team, while his son Spencer is the manager of the robotics product line.

NVIDIA is also increasing its investment in simulation tool companies, hoping to further expand Omniverse's market penetration by promoting the compatibility of these companies' software with Omniverse tools. A recent example is that in December 2025, NVIDIA announced a $2 billion investment in Synopsys—a global leader in semiconductor design software. Synopsys had acquired ANSYS in July 2025 for $35 billion, whose software is widely used for simulating real physical behaviors.

Theoretically, the market demand potential for simulation technology is enormous. Among them, the robotics field is one of the largest potential opportunities. Currently, developers lack spatial and motion data on human action patterns and object interaction methods, which are crucial for developing humanoid robots. Some robots, represented by Tesla's "Optimus," still rely on workers wearing motion capture devices to collect such data However, compared to simulation technology, this method is costly and time-consuming.

Nevertheless, NVIDIA is facing fierce competition in this field. Unity Technologies, which helped develop popular games like Pokémon GO, has also launched scene and object modeling tools that support clients in designing, testing, and training robotic systems in virtual environments. Another competitor is the open-source robotic simulation platform Gazebo, developed by the nonprofit Open Robotics. Additionally, as there is currently no standardized software that meets the needs of all companies, many choose to develop their own internal simulation software.

Indeed, the field of simulation technology is still in its early stages of development. NVIDIA executives liken Omniverse to the company's early bet on CUDA technology—CUDA is a software suite launched by NVIDIA that supports the training and running of artificial intelligence models on its GPUs. Although analysts and activist investors raised objections at the time, arguing that it was unwise for NVIDIA to invest billions of dollars in a business with little short-term return, Jensen Huang still fully supported the development of CUDA.

"As we invested in CUDA over more than a decade to ultimately disrupt the field of deep learning, NVIDIA has been continuously building Omniverse for years, positioning it as foundational software to empower the next massive market of accelerated computing—physical artificial intelligence," said Rev Lebaredian, NVIDIA's vice president responsible for the Omniverse business, in a written statement.

He added, "This long-term vision is now gradually coming to fruition, with major cloud service providers building dedicated simulation infrastructure, and robotics and automotive companies leveraging our framework to develop and train the next generation of autonomous intelligent devices."

**Immature Software Products**

NVIDIA officially launched Omniverse in 2021, positioning it as a software and simulation tool that supports designers in collaborative 3D virtual design, similar to a shared Google Document for 3D models and environments.

BMW is one of the earliest well-known customers of the software, using NVIDIA's Isaac robotics software to optimize material flow processes in its automotive factories.

In recent years, NVIDIA has made high-profile announcements of partnerships with numerous companies for Omniverse, including industrial manufacturers like Dassault and Foxconn, autonomous vehicle developers like Mercedes-Benz and Jaguar Land Rover, and robotics companies like Boston Dynamics and Figgler AI.

NVIDIA's intention is that the Omniverse software will not only drive demand growth for its GPUs but also generate profits through enterprise software licensing and technical support services—annual licensing fees for a single chip can reach up to $4,500. Additionally, customers can choose to subscribe to Omniverse cloud services, which bundle GPU server access with software licensing.

However, just as the Omniverse team was pushing forward with collaborations, NVIDIA's artificial intelligence chip business experienced explosive growth. At the end of 2022, OpenAI launched ChatGPT, triggering a race among major tech companies in artificial intelligence technology, with the vast majority of companies' R&D efforts based on NVIDIA's AI chips Despite the tremendous success of the artificial intelligence chip business, Jensen Huang is known for his concern about being surpassed by competitors and his push for the company to explore new revenue sources.

He is particularly dissatisfied with the lackluster performance of the Omniverse business. According to a participant at the meeting, Huang questioned the team at Omniverse during the 2023 all-hands meeting about whether they had achieved profitability.

Additionally, according to another insider, Huang learned that the collaboration project announced between NVIDIA and BMW was much smaller in scale than he had initially been informed, which made him very frustrated with the Omniverse team. A reporter sent an email and called BMW for comments on this matter but received no response. An NVIDIA executive stated in 2023 that BMW has over 20,000 factory planners worldwide using Omniverse software, but did not clarify the actual impact of this number on sales.

Another example: during a conference call, a team from the Omniverse department requested additional personnel to develop new products, which infuriated Huang. According to a participant, Huang severely reprimanded the team for an hour afterward, accusing them of wasting engineers' time and advising them on the true way NVIDIA develops products. He made it clear that the current staffing of the team was more than sufficient.

Between 2023 and 2024, the Omniverse team presented Huang with demo versions of several software products, including a 3D modeling tool that supports multi-user collaboration, a concept product for streaming software, and a vehicle simulation program. However, none of these demo products ultimately transitioned into commercial versions. Insiders noted that Huang pointed out that the team was pursuing too many projects simultaneously and should focus on one product that could truly be realized.

This criticism prompted the team to shift their R&D focus more towards digital twin modeling software. Meanwhile, as more companies began using Isaac Sim to generate synthetic data, robotics also became an important component of the Omniverse business.

However, insiders indicated that the robotics industry is still in a developmental stage, with many companies, including Tesla, continuing to develop simulation software independently rather than adopting NVIDIA's solutions.

Moreover, robotics developers have expressed grievances about the shortcomings of NVIDIA's software. Co-founders of two robotics startups stated that, for example, Isaac Sim cannot be used to train robots to manipulate complex objects, especially items like clothing that have constantly changing shapes.

They pointed out that Omniverse tools can indeed be used to train robots to walk or move independently, but enabling robots to interact with other objects is the technical key.

Driving user subscriptions to Omniverse cloud services is equally challenging. Dawn Cole, COO of Loop Company, stated that he had purchased several computers equipped with NVIDIA chips for researchers to conduct factory equipment simulation tests. Although he learned that companies could use the cloud-based Omniverse servers through licensing, he believed that this was not cost-effective considering the expenses involved. Loop Company primarily designs automated production equipment for clients in fields such as aerospace The problems exist not only in robot-related software. Several software developers have reported that the scene creation tools in Omniverse are difficult to operate and can only run stably in limited scenarios; once they exceed the preset range, they are prone to failure. They also pointed out that the usage instructions for the tools are not only rudimentary but also outdated, making it difficult for users to troubleshoot issues when building virtual objects and scenes.

Software developer Valentin Follage stated that the Omniverse software performs reasonably well in creating virtual retail shelves—his company is using this feature to develop sensors for tracking consumer product browsing in physical stores. However, when he attempted to add human models in the same virtual environment, the computer crashed directly.

"As long as you try to perform some operations that are slightly beyond the norm, this software will fail and become almost completely unusable," Follage said.

Follage revealed that at an official NVIDIA event held in Seoul in November 2025, he reported this issue to an NVIDIA representative, who admitted that Omniverse currently cannot meet his specific needs and suggested he switch to software from Unity.

**Current Industry Demand**

NVIDIA's layout in the simulation technology field may be somewhat hasty. Company executives stated that they are trying to create a market that has not yet formed, one that users themselves are not even aware of their needs. Former NVIDIA employees and industry insiders pointed out that many industrial manufacturers lack the relevant technical reserves and are unwilling to invest funds in this area.

Moreover, NVIDIA also finds it difficult to develop a simulation software that can widely adapt to the needs of different industry users. There are significant differences in the simulation technology demands across the robotics, automotive, and industrial equipment sectors.

To address this dilemma, NVIDIA's proposed solution is to increase investment in digital twin and simulation service startups to gain deeper insights into customer needs while promoting these companies to develop industry-specific services based on Omniverse's software tools.

One investment case is the stake in the Taiwanese startup MetAI, which focuses on creating digital twin models for semiconductor wafer factories, covering services such as construction planning and layout simulation for TSMC factories.

"I believe people underestimate the resources required to develop simulation systems with real domain expertise," said Daniel Yu, CEO of MetAI. "This work requires a significant amount of time. Omniverse is not a fully functional application; it is a horizontally open platform for developers."

### 相關股票

- [NVIDIA (NVDA.US)](https://longbridge.com/zh-HK/quote/NVDA.US.md)
- [GraniteShares 2x Long NVDA Daily ETF (NVDL.US)](https://longbridge.com/zh-HK/quote/NVDL.US.md)
- [XL2CSOPNVDA (07788.HK)](https://longbridge.com/zh-HK/quote/07788.HK.md)
- [XI2CSOPNVDA (07388.HK)](https://longbridge.com/zh-HK/quote/07388.HK.md)
- [YieldMax NVDA Option Income Strategy ETF (NVDY.US)](https://longbridge.com/zh-HK/quote/NVDY.US.md)
- [Direxion Daily NVDA Bear 1X ETF (NVDD.US)](https://longbridge.com/zh-HK/quote/NVDD.US.md)
- [T-Rex 2X Long NVIDIA Daily Target ETF (NVDX.US)](https://longbridge.com/zh-HK/quote/NVDX.US.md)
- [T-Rex 2X Inverse NVIDIA Daily Target ETF (NVDQ.US)](https://longbridge.com/zh-HK/quote/NVDQ.US.md)
- [Direxion Semicon Bull 3X (SOXL.US)](https://longbridge.com/zh-HK/quote/SOXL.US.md)

## 相關資訊與研究

- [Fractile Seeks $200 Million as Nvidia Faces Inference Pressure](https://longbridge.com/zh-HK/news/281215985.md)
- [Nvidia Stock (NVDA) Braces for a Quick Snapback after a Rare Two-Quarter Losing Streak](https://longbridge.com/zh-HK/news/281360100.md)
- [Nvidia Stumbled in the First Quarter: Here are My Top 3 Nvidia Predictions for the Second Quarter](https://longbridge.com/zh-HK/news/281620093.md)
- [Nvidia (NVDA) H100 Prices Surge 40% as New GPUs Fail to Meet Insatiable Demand](https://longbridge.com/zh-HK/news/281588202.md)
- [Benchmark Co. Keeps Their Buy Rating on Nvidia (NVDA)](https://longbridge.com/zh-HK/news/281238653.md)