NVIDIA might be one of the most recognized names in AI and semiconductors today. After all, how could you not recognize a company when its stock has been mooning towards the $3 trillion valuation since 2025?
But NVIDIA is playing a much more strategic game than just GPUs and AI datacenter chips. In fact, it’s been making a series of investments into companies building on top of and around those two technologies.
Whether it’s healthcare AI startups or foundational chip companies, NVIDIA has been slowly and methodically taking positions in the companies that will define the coming decade. If you want to see where the world is heading, don’t look at NVIDIA’s products. Look at its portfolio.
Who are these companies? What are they building? And why is NVIDIA investing?
Before we jump in, here’s why NVIDIA’s investments matter NVIDIA isn’t just selling silicon. It’s building an entire platform—and that means championing the companies building on top of its technology.
In a lot of ways, every dollar NVIDIA puts into one of these companies often finds its way back into NVIDIA’s coffers. That’s because they’re all heavily NVIDIA-dependent. And while it may not own all of these companies outright, it’s working to exert influence where possible.
- Robots running on its CPUs and GPUs
- Infrastructure companies using NVIDIA chips and software
- Deep tech companies running their AI workloads in NVIDIA’s cloud
No one expects NVIDIA to sit on the sidelines of this exogenous growth. It’s building a fortress of allies.
NVIDIA Has Built a Fortress of Allies
Now, let’s dive into NVIDIA’s ten most notable investments across industries. While this is certainly not an exhaustive list, each of these companies plays a critical role in NVIDIA’s strategy—and yours if you follow its lead.
Arm Holdings (ARM)
One of the first—and largest—investment moves NVIDIA has made is toward one of its most direct competitors: Arm Holdings.
SoftBank had an initial $40B+ deal on the table to sell its majority stake in Arm to NVIDIA. But red tape and regulators in the UK and US blocked the deal, and SoftBank went another direction with its IPO. However, that hasn’t stopped NVIDIA from increasing its minority stake in the company through its IPO.
Arm’s CPU architecture powers nearly every smartphone on the planet and is increasingly making inroads into IoT and edge computing. NVIDIA chips are popular in data centers, but Arm dominates mobile and power-efficient computing devices.
Arm’s technical and legal access to IP in both companies’ chip business models also means cross-pollination at the architectural level is inevitable. NVIDIA, in its own Arm partnership, was already a member of the “dos 5” developing team. That means it is well-positioned to influence the mobile and edge market despite losing out on total ownership.
Arm isn’t an acquisition target. NVIDIA is trying to influence the larger CPU+GPU battleground, likely to its own benefit in the long run. NVIDIA’s competitors include other Arm licensees like Qualcomm, Apple, and MediaTek.
Recursion Pharmaceuticals (RXRX)
This investment from NVIDIA was one that many would not have expected. In 2023, NVIDIA poured $50 million into biotech company Recursion.
Recursion has already mapped out billions of the relationships between genes, proteins, and chemicals—and it’s doing it with NVIDIA AI and machine learning compute hardware.
Recursion is running its massive AI workloads directly on NVIDIA’s DGX cloud platform. The partnership with NVIDIA isn’t just about compute, though. It’s about entering into a mutually beneficial partnership with a biotech firm doing AI-driven drug discovery.
NVIDIA wants to make its GPUs essential to powering the future—not just in technology but in health and biotechnology. Bioinformatics and AI-run drug discovery are massive new frontiers with massive long-term tailwinds. NVIDIA is hedging its bets across both.
SoundHound AI (SOUN)
SoundHound is another investment in AI that NVIDIA was not shy to publicize. SoundHound has been working on a variety of voice AI technologies, and NVIDIA recently acquired a strategic minority stake in the company.
Voice is where generative AI is headed, and NVIDIA wants a seat at the table. That’s why NVIDIA has thrown its support behind SoundHound AI, a company that’s been on the ground floor of voice and conversational AI development.
The field of voice is analogous to the current LLM space. SoundHound AI is essentially a domain-specific LLM (LLaMA) for voice.
NVIDIA’s competition here is Google. The “google” in many searches is coming from AI.
Hugging Face
In a Series D funding round worth $235 million, NVIDIA was one of the first parties to Hugging Face.
Other notable backers of the firm include Google, Amazon, and Salesforce.
Hugging Face is the GitHub of machine learning models. It is the common ground where machine learning engineers and researchers go to find open-source AI models and share their own. NVIDIA’s backing of Hugging Face at such a young stage makes it a very favorable outcome for NVIDIA.
It’s less about having competition within generative AI—and more about owning the underlying technology stack required to train these models.
NVIDIA’s direct competitors here include Amazon Web Services and Microsoft Azure.
Runway ML
Runway is another company in the creative AI space that NVIDIA has had close partnerships with that has not always been disclosed in full publicly.
Runway, like NVIDIA, is working to create tools for video generation, editing, and creative production in AI models.
Video is an intensive GPU task, and any tool that NVIDIA has a stake in means the proliferation of NVIDIA hardware at its cores.
Own the creative industry with generative AI. The goal here is to get NVIDIA GPUs into the workflows of content creators and media producers as widely as possible.
Cohere
Cohere is a company that has very quickly set itself up as a potential direct competitor to OpenAI, particularly when it comes to enterprise-focused large language models (LLMs).
NVIDIA has participated as a cloud computing provider and has invested in the company as well as participated in multiple other fundraising rounds over the past several years.
Cohere is building out AI models to be used by enterprises, and NVIDIA is working hard to ensure that these models can be supported on its own cloud infrastructure. NVIDIA may not always pick one LLM over another in the end, but what it can do is ensure that it has a say in these companies’ early stages—and that their models are optimized for NVIDIA GPUs.
Own the infrastructure, not the gold. NVIDIA is not committing its coin to betting on only one LLM winner but rather supporting the entire ecosystem and making sure all the other pick and shovel companies are doing workloads on NVIDIA products.
CoreWeave
One of NVIDIA’s biggest moves recently has not necessarily been a direct investment but rather a partnership announced in 2023 and 2024 to make CoreWeave a primary provider of GPU-based cloud infrastructure.
CoreWeave became one of the largest purchasers of NVIDIA’s own H100 chips for the purpose of their data centers. In fact, NVIDIA helped CoreWeave finance the purchases to scale the business and add more GPU data centers.
Accelerate demand for its H100 GPU by enabling CoreWeave as an alternative for vertical-specific GPU cloud data centers outside of AWS or Azure.
Aeva Technologies
If NVIDIA is the brain, it also wants to be the long-range eyes for self-driving cars, too. In 2023, it acquired an equity stake and deepened its partnership with Aeva, an emerging LIDAR technology company.
NVIDIA has an autonomous driving division that’s been picking up speed. Aeva’s long-range sensing technology fits directly into that vision.
NVIDIA is playing a dual role here in two ways: NVIDIA wants to have great core software, but it also has been investing in a wide range of sensor hardware technologies (like LIDAR) to power the self-driving stack.
A competitor to NVIDIA here is most likely mobile chip rival Qualcomm.
AI21 Labs
AI21 is one of the front-runners in Israel in building LLMs and natural language processing tools that look to be direct competitors with both OpenAI and Anthropic, especially in regions outside of the US.
NVIDIA has invested in the company and partnered with it as a strategic partner.
If NVIDIA wants to make sure it has the pick and shovel over the other potential tech giants, NVIDIA will ensure these companies have the compute resources they need—on NVIDIA infrastructure, naturally.
Neutral but essential. NVIDIA will not pick AI21 as its favorite but will support any company and ensure its resources are centered on NVIDIA products.
The direct competition NVIDIA is staving off here is any and all generative AI companies with similar datacenter goals.
Perplexity AI
Perplexity is another search engine alternative being built to compete with Google that is LLM-powered and conversationally fast. NVIDIA has backed it indirectly as an investor in Perplexity’s fundraising, and like NVIDIA, Perplexity runs very deeply on NVIDIA’s infrastructure.
Perplexity is just one search engine among the next wave of AI-native applications and services built on the back of LLMs. If you’re looking for fast and frictionless searching that can answer many more open-ended and unique queries, you’re not going to use Bing or Google you’re going to use Perplexity.
Gain from the value creation as a result of building on top of their services.
One could also argue NVIDIA’s competitors are the current state of internet companies.
Map Out NVIDIA’s Winners
None of the companies above are haphazard. NVIDIA’s portfolio strategy reveals exactly where its “smart money” investment capital is being placed.
You may not have a billion-dollar-plus venture capital arm to throw around at your disposal, but NVIDIA’s portfolio can point you in the right direction.
If you’re trying to play this rally as a retail investor, here are three approaches you can take that will mirror NVIDIA’s actual portfolio strategy, close enough:
AI ETFs
AI is too big to just pick a few companies to back. ETFs like BOTZ (Global X Robotics & AI ETF) or even SMH (Semiconductor HOLDS in New Hartford, CT). These ETFs already give NVIDIA exposure and are comprised of a basket of AI leaders in hardware and software.
Pick NVIDIA Stocks
NVIDIA itself is the most obvious one here, of course. But if you’re looking at NVIDIA’s portfolio, then go for it. Stocks like Recursion, SoundHound AI, Arm, and even potentially Hugging Face are ones that NVIDIA has already flagged and investigated in depth as winners and direct partners.
Pick Infrastructure
A better strategy might be to go with infrastructure play. It’s a “buy the picks and shovel” play if you will. Instead of trying to pick a single winner, bet on the underlying infrastructure.
If NVIDIA is going to back it and build on it, it’s likely a stock you want in your portfolio too.
The “pick the infrastructure” play at NVIDIA level has been buying up public GPU inventories and pre-financing large data center construction efforts.
The picks and shovel strategy is sound investing because if NVIDIA is investing, they want to reduce your risk.
Final Take
If the internet was the gold rush, NVIDIA wants to be the company selling shovels. NVIDIA is cashing in not on the gold itself but on the power plants of the future.
Its entire M&A strategy and investment activity are trying to ensure NVIDIA is not just a part of these businesses but leading the charge wherever possible.
- NVIDIA Investments Is About Platform Dominance
- NVIDIA’s strategy is all about ensuring that it is not only on the right side of the AI revolution but that it controls as much of the underlying infrastructure as possible. It’s winning by any means necessary.
Watch NVIDIA in this space, and you are watching one of the most dominant and deeply integrated technology companies of all time playing in the deepest pockets—and most aligned incentives—to keep control over the biggest opportunities of our time.
That in itself is a signal that should tell you exactly who to follow and how to follow them when it comes to investment success.
Leave a Comment