NVIDIA’s dual-faced investing tentacles: NVentures locks star startups into a compute-power empire

Author: Jae, PANews

As global tech insiders focus their attention on semiconductor giant Nvidia’s GPU deals worth billions—up to thousands of billions of dollars—an updated disclosure statement from the UK Companies House reveals what is arguably the largest bet Nvidia has made in the digital finance space to date.

The filing shows that NVentures, Nvidia’s venture capital arm, holds 141,834 shares of the European digital bank Revolut, with a position value of nearly $196 million.

Nvidia’s bet on a digital bank not only demonstrates how it is branching out across multiple domains beyond semiconductors, but also brings NVentures—previously hidden behind Nvidia’s shadow—into the public eye. Since its establishment in 2021, NVentures has been weaving a capital net spanning cutting-edge technology fields using a playbook different from traditional corporate venture capital (CVC).

A “two-pillar” capital structure: industrial investments entrench the compute power hegemony, NVentures locks onto the tech frontier

In Silicon Valley’s capital narrative, outsiders often lump all of Nvidia’s external investments together as the will of the parent company. But few know that Nvidia internally has long built a “two-pillar” investment system with clearly defined roles and highly complementary functions: Corporate Development and the venture capital arm NVentures operate in parallel, jointly forming Nvidia’s capital map.

Corporate Development focuses on large-scale equity investments and industrial M&A in the tens of billions to hundreds of billions of dollars. Its head is Vishal Bhagwati. As reported by CNBC, in just the first four months of this year, the investments led by this division have already exceeded $40 billion. Representative cases include a $30 billion framework investment in OpenAI, a $10 billion capital injection into Anthropic, and a $2 billion equity investment in xAI, Musk’s company—each deal is a major move capable of shaking up the industry.

Corporate Development’s primary goal is to tightly bind foundation model companies and cloud service giants to lock in future GPU procurement agreements and consolidate Nvidia’s long-term core base. At its essence, it is “building walls”—using billions of dollars in real capital to stack Nvidia’s compute power hegemony and cement its leading position.

NVentures, by contrast, stands out sharply. Its positioning is strategic venture capital, focusing on uncovering early-/growth-stage bets in hard-tech tracks. NVentures is led by Nvidia Vice President and venture capital head Mohamed Siddeek, whose career also mirrors Nvidia’s rise: as early as the late 1990s, when he worked at Morgan Stanley, he personally accompanied Jensen Huang to complete Nvidia’s IPO roadshow. After that, he held long-term responsibility for venture capital in TMT (technology, media, and telecom) and digital healthcare at top investment institutions such as Mubadala, the sovereign wealth fund of the UAE, and SoftBank Vision Fund.

Under Siddeek’s management, NVentures’ active performance has been keeping pace with the top Silicon Valley venture capital firms.

By mid-2026, NVentures has cumulatively invested in 96 startups, incubating 20 unicorns from them. In just the past 12 months alone, it completed 43 new investments, with the number of deals up more than 900% year over year. Most of its single-investment sizes range from several million dollars to hundreds of millions of dollars, and more than 85% of deals involve participation through follow-on investments.

NVentures does not pursue control. Instead, it seeks to “connect”—using each venture capital check to extend Nvidia’s reach into frontier technology territories that the semiconductor industry has not yet arrived at, positioning itself lightly as the “rule maker” of vertically integrated industries.

NVentures invests across five major tracks, widening the boundaries of the compute-power ecosystem

Siddeek has publicly said that NVentures’ investment boundaries “basically cover anywhere Nvidia’s software and hardware can reach.” Translated, it means: wherever compute power will be needed in the future, Nvidia’s flag should be planted in advance.

AI infrastructure: reinforcing the invisible fortress of CUDA

Building technical barriers that entrench the CUDA ecosystem is NVentures’ most basic and most important mission. AI automation code review tool CodeRabbit not only received investment from NVentures, but was also rolled out and implemented by Nvidia across the entire company. It deeply integrates Nvidia’s Nemotron open model, significantly improving enterprise-level code review efficiency, becoming a typical example of CUDA ecosystem deployment.

In addition, Contextual AI’s retrieval-augmented generation (RAG) middleware helps enterprises localize RAG deployments of proprietary data while ensuring data security. From video generation platform Synthesia, to AI music platform Suno, to multimodal world model developer Runway, and to Tensormesh, an AI inference infrastructure startup that NVentures invested in this May, NVentures is reshaping the compute-power ecosystem for creativity and inference from the ground up.

Robotics and embodied intelligence: connecting the physical-world closed loop of AI

As embodied intelligence companies such as Skild AI, Bedrock Robotics, and Flexion Robotics were successively included in the investment portfolio, Nvidia’s ambitions in robotics have long surpassed mere hardware sales.

Take Skild AI as an example: its universal robot foundation model, trained through physical simulation, enables robots to control different hardware and adapt to complex environments without targeted adaptation. Nvidia provides compute power and underlying simulation development platforms such as Isaac and Omniverse. This means Nvidia starts defining the robot’s “brain” from the R&D stage. This is an “implantation”-style investment. By the time robots enter mass production, they have already been “trained” by Nvidia.

By combining autonomous driving trucks like Waabi and logistics automation company Outrider, Nvidia is trying to penetrate the software ecosystem early, before the hardware standards for robotics have been established.

Digital healthcare and digital biology: the most heavily weighted track

Digital biology is the track with the highest number of NVentures investments, accounting for more than one-fifth. Nvidia views “AI + biology” as the next “revolution at the transistor level of computing,” following semiconductors, and also expects it to become one of the most important high-value compute-power scenarios for GPUs beyond large model training.

For example, Abridge focuses on clinical speech AI transcription, freeing doctors’ productivity from cumbersome clinical documentation and archiving. Basecamp Research is using Nvidia’s BioNeMo platform to develop EDEN biological foundation models for gene engineering. By investing in drug R&D companies such as Generate Biomedicines, Nvidia is gradually “coding” life science, opening up a new high-value computing market independent of traditional large model training.

If future new drug R&D no longer relies on test tubes, but instead on GPU-simulated computations, Nvidia will take the largest slice of global pharma R&D budgets.

Cutting-edge computing and quantum technology: laying compute hubs for the post-silicon era

A hybrid architecture of classical computing and quantum computing is becoming the direction of evolution for next-generation supercomputing. NVentures’ moves in quantum come with strong defensive attributes: French quantum computing startup Alice & Bob and Silicon Valley quantum computing pioneer PsiQuantum are both on its investment list.

In the near term, quantum computing cannot replace traditional silicon-based chips, but it has already shown potential in specific ultra-large-scale scientific computing scenarios. By investing in quantum computing companies, Nvidia is promoting the development of hybrid architectures combining classical GPU compute power and quantum control systems. The purpose is straightforward: even in a future non-silicon computing era, Nvidia will still occupy the central position in compute scheduling.

Green energy and power grid dispatch: breaking through the physical ceiling on compute expansion

Today, the constraints on global data centers have fundamentally changed: power supply has replaced hardware production capacity and land resources, becoming the final bottleneck for compute-power expansion.

Against this backdrop, NVentures has invested in energy coordination and management platforms such as Emerald AI. Emerald AI can help Nvidia’s “AI Factories” connect to existing power grids more safely and quickly, and during peak electricity use, flexibly dispatch local energy storage and backup generators—turning high-energy-consumption compute clusters into “flexible assets” that are power-grid friendly. In addition, with bets on nuclear fusion R&D company Commonwealth Fusion, and on battery recycling and energy storage giant Redwood Materials, Nvidia is personally building an energy barrier for its unfinished compute-power empire.

The most recent investment in Revolut for $196 million means NVentures has formally extended its reach into the fintech sector. As a digital bank with tens of millions of users and massive data, Revolut is not only a potential customer for Nvidia; its large user base and data resources will also become a foothold for Nvidia’s exploration of the financial space. By backing Revolut through NVentures, Nvidia is effectively making a strategic deployment of “AI + digital finance” application scenarios.

Nvidia’s “full household bundle” encirclement: not just giving money, but “locking it down”

What attracts the most to NVentures for investee companies is not its money, but the “empowerment premium” it provides.

While ordinary VCs just wait to exit after writing the checks, NVentures brings “Nvidia’s full household bundle” afterward: priority integration of top engineering teams and development platforms, and even access to global distribution channels. CodeRabbit’s CEO has publicly said that Nvidia engineers helped them embed Nemotron models into their product, and then promoted it across the company, directly earning endorsement at the level of a Fortune 500 company.

All underlying empowerment is deeply tied to Nvidia’s CUDA software ecosystem. Whether using Nemotron or BioNeMo, whether doing robotics simulation on Isaac or building digital twins based on Omniverse, the underlying computing architecture is locked into Nvidia’s technical system without exception.

As NVentures extends its reach across the upstream and downstream of the industry chain, many companies have been accustomed since their founding to build their own technology stacks within Nvidia’s ecosystem. If competitors such as AMD or Google offer better cost performance in the future, they will face extremely expensive migration costs and the risk of an ecosystem rupture.

For investees, NVentures’ capital injection is a ticket into the world’s largest AI ecosystem. But the other side of the coin is that tying into Nvidia also means a high reliance on a single vendor. In effect, NVentures uses a single venture capital check to weld them firmly onto Nvidia’s war machine, turning them into “ecosystem nodes” spread across various vertical industries.

In the long run, this may not only weaken market incentives to explore alternative compute-power solutions, but also reinforce a market structure dominated by one player.

Nvidia is no longer just the “shovel-seller” of the AI era. From its latest investment in Revolut to its capital map spanning cutting-edge technology industries, NVentures’ growth trajectory is both a snapshot of Nvidia’s ongoing expansion and a main lever by which Nvidia turns its technical advantages into industry control.

In the future, at every key inflection point in new industries, NVentures may appear. And Nvidia’s shadowy tentacles—using capital as ink—are rewriting the power map of the global tech sector.