NVIDIA is not building a conventional, diversified venture-capital portfolio. It is investing across the AI stack—model developers, GPU-cloud providers, robotics companies, developer platforms and data infrastructure—to strengthen an ecosystem that consumes NVIDIA chips, networking, systems and software.
The strategy is broader than a list of startup stakes. NVIDIA’s filings show $17.5 billion in investments in private companies and infrastructure funds during fiscal 2026, followed by $18.6 billion disclosed in the first quarter of fiscal 2027. Those figures include more than traditional startup equity, so they should not be added up as though they represent a complete venture portfolio. NVIDIA’s filings also disclose infrastructure guarantees, while many individual private-company stakes remain undisclosed.
As an Amazon Associate I earn from qualifying purchases.
What NVIDIA is actually investing in
NVIDIA can have several different relationships with the same AI company. Those relationships are economically important, but they are not interchangeable.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →| Relationship | What it means | Example |
|---|---|---|
| Equity investment | NVIDIA buys shares in a private or public company. | CoreWeave and Nebius |
| Strategic partnership | The companies agree to technical integration, product deployment or commercial cooperation, with no necessary equity transaction. | Robotics companies using Isaac, Cosmos or Omniverse |
| Infrastructure commitment | A capacity purchase, deployment agreement, guarantee or other support for land, power, data centers or cloud infrastructure. | Planned gigawatt-scale deployments by AI-cloud providers |
| Public-company stake | NVIDIA invests in a listed corporation rather than a conventional startup. | Intel, Synopsys, Coherent and Nokia |
| Inception participation | A startup receives NVIDIA technology, training, marketing or ecosystem support. | NVIDIA Inception members |
NVIDIA Inception is explicitly a startup-support program, not proof that NVIDIA owns a company. NVIDIA says it supports tens of thousands of startups, but program membership should not be described as venture investment. The official Inception page explains the program’s current benefits and eligibility.
#1 Best Overall
- AI Performance: 767 AI TOPS
- OC mode: 2632 MHz (OC mode)/ 2602 MHz (Default mode)
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Axial-tech fan design features a smaller fan hub that facilitates longer blades and a barrier ring that increases downward air pressure
- A 2.5-slot design maximizes compatibility and cooling efficiency for superior performance in small chassis
The most consequential bets and relationships
There is no complete public ledger that identifies every NVIDIA private-company investment, ownership percentage and check size. The companies below are therefore a curated map of the most strategically significant publicly reported investments and relationships—not a definitive ranking of NVIDIA’s entire portfolio.
OpenAI: the clearest model-and-infrastructure bet
OpenAI is the most prominent example of NVIDIA combining capital with infrastructure alignment. NVIDIA reportedly invested $100 million in OpenAI’s 2024 financing round, according to technology-press coverage. The companies later announced an intended strategic partnership involving at least 10 gigawatts of NVIDIA systems and a potential NVIDIA investment of up to $100 billion.
That potential investment should not be presented as completed. In its fiscal 2026 filing, NVIDIA said it was still finalizing the agreement with OpenAI and warned that no transaction was guaranteed to close. The 10-gigawatt figure is also a forward-looking deployment plan rather than proof that the capacity is already operating. NVIDIA’s announcement describes the proposed arrangement, while the SEC filing supplies the qualification.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallStrategically, OpenAI can provide NVIDIA with a major customer, a high-profile validation of its platform and a source of future demand for GPUs, networking and complete data-center systems. It also raises the question of circular economics: NVIDIA may invest in a company that uses capital and financing support to purchase NVIDIA infrastructure. That possibility is worth examining, but it is not evidence by itself of improper accounting or artificial demand.
CoreWeave: financing the specialized AI cloud
CoreWeave connects NVIDIA’s capital directly to a company that buys and deploys NVIDIA systems at large scale. NVIDIA invested $2 billion in CoreWeave Class A common stock at $87.20 per share in January 2026. The companies also announced plans for CoreWeave to accelerate deployment of more than 5 gigawatts of AI factories by 2030, adopt NVIDIA CPUs and storage platforms, and provide related software to cloud providers and enterprises. CoreWeave’s filing and the joint announcement document the transaction.
CoreWeave became a public company in 2025. It is therefore more accurate to describe the relationship as a startup-era investment followed by a later public-company stake, rather than simply calling CoreWeave an NVIDIA startup.
The investment supports NVIDIA’s broader effort to cultivate specialized AI clouds. These providers rent GPU clusters to startups and enterprises that cannot—or do not want to—build their own data centers. The model can expand NVIDIA demand beyond a small number of hyperscale customers, but it carries substantial risks: expensive facilities, electricity and cooling requirements, customer concentration, debt and financing exposure, and the possibility that GPU utilization falls.
Nebius: expanding AI capacity across regions
NVIDIA announced a $2 billion investment in Nebius in March 2026 alongside a strategic partnership to develop full-stack AI-cloud infrastructure. Nebius said the relationship would support deployment of more than 5 gigawatts of NVIDIA systems by the end of 2030. The filed announcement describes the investment and target.
Nebius demonstrates that NVIDIA’s strategy is not limited to backing famous model companies. Regional and international cloud capacity can make NVIDIA-powered computing available to developers and enterprises that are outside the largest U.S. cloud ecosystems.
The 5-gigawatt figure is a planned target, not installed operating capacity. Investors and customers should distinguish among an announced goal, a signed purchase commitment, financed construction and deployed infrastructure.
Mistral AI: supporting an open frontier-model ecosystem
Mistral AI is one of Europe’s leading frontier-model companies and has reportedly received NVIDIA participation in multiple financing rounds. TechCrunch reported that NVIDIA participated in Mistral’s approximately €1.7 billion Series C in 2025. The size of that financing round is not the same as NVIDIA’s individual check.
Recommended Free Tools
NVIDIA later included Mistral in its Nemotron Coalition, alongside other model and developer companies. Mistral helps illustrate several strategic goals: supporting European AI capability, encouraging an open or partially open model ecosystem and increasing demand for training and inference infrastructure.
Investment does not mean NVIDIA controls Mistral or that Mistral must use NVIDIA hardware exclusively. Model companies can use multiple chip vendors and cloud platforms.
Rank #2
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5060
- Integrated with 8GB GDDR7 128bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
Perplexity: backing the application layer
Perplexity gives NVIDIA exposure to an AI-search product rather than only a foundation-model lab or infrastructure company. TechCrunch reported that NVIDIA first invested in Perplexity in 2023 and participated in later rounds, including a $500 million financing closed in December 2024.
Perplexity matters because successful AI applications can become major destinations for inference—the repeated process of running a trained model to answer user requests. That creates a different demand profile from one-time model training and may require sustained computing capacity as usage grows.
NVIDIA also included Perplexity in the Nemotron Coalition. The company’s investment participation, technical relationships and any separate acquisition speculation should be treated as distinct claims.
Cohere: enterprise AI demand
Cohere develops enterprise-focused AI models and applications. NVIDIA identified Cohere in its discussion of recent investments and its wider AI ecosystem. NVIDIA’s investment overview lists the company among its portfolio relationships.
Cohere is strategically useful because it shows that NVIDIA’s portfolio is not built solely around consumer chatbots. Enterprise deployments may generate demand for inference systems, model-serving software, secure private-cloud installations and data-center networking.
Public reporting on a financing round should not be interpreted as proof of NVIDIA’s exact ownership percentage. Unless a company or filing discloses the check, the safest description is that NVIDIA participated in or was reported to have participated in a financing.
Scale AI: the data and evaluation layer
Scale AI operates in data labeling, model evaluation and AI-development infrastructure. Its importance lies in the “picks-and-shovels” layer: the data preparation and testing processes that help companies build and operate useful AI systems.
Scale is not equivalent to a GPU-cloud provider or a frontier-model company. Its capital needs, hardware consumption and revenue model are different. Including it in the map shows that NVIDIA’s ecosystem strategy reaches beyond chips and model training into the services required to make AI products reliable and deployable.
xAI: large-scale infrastructure and frontier models
xAI represents NVIDIA’s exposure to another frontier-model builder and large-scale AI-infrastructure operator. NVIDIA publicly identified xAI among partners involved in building U.S. AI infrastructure. The announcement supports the infrastructure relationship.
xAI illustrates why AI capacity involves more than GPUs. Large clusters require power, cooling, networking, storage, construction and software. The available evidence does not establish a precise NVIDIA equity amount, so it should not be stated as one.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteThe robotics and physical-AI portfolio
NVIDIA’s investment and ecosystem activity increasingly extends into machines that perceive, reason and act in the physical world. Its NVentures page highlights companies including Skild AI, Generalist AI, Agility Robotics, 1X, Bedrock Robotics, Flexion and Vention. Technology coverage has also identified Figure AI as a major NVIDIA-related startup investment.
These companies span humanoid robots, industrial automation, autonomy and robotics software. Their strategic value is different from that of a chatbot company: they need simulation, synthetic data, real-time inference, sensor processing and systems that can operate reliably in factories, warehouses and other physical environments.
NVIDIA supplies several enabling technologies:
- Isaac for robotics development and simulation.
- Cosmos for physical-AI and world-model tools.
- Omniverse for simulation and digital twins.
- GPUs, networking and accelerated computing for training and inference.
There is an important evidence distinction here. NVIDIA’s NVentures and startup pages may present portfolio companies, technology partners and ecosystem participants together. A company appearing on the page is not automatically proof of an equity investment.
Rank #3
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5070 Ti
- Integrated with 16GB GDDR7 256bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
The broader picks-and-shovels network
Other publicly discussed relationships help complete the portfolio map:
Free tools Windows power users keep installed
One-click scans. No signup required.
- Hugging Face: a central distribution and collaboration layer for open-source models, datasets and tools. Its importance is partly developer adoption: open models may be hardware-neutral in principle but are often deployed on NVIDIA systems because of CUDA support, optimized libraries and cloud availability.
- Crusoe and Lambda: specialized AI-cloud providers that rent access to GPU clusters and serve startups and enterprises.
- Runway: an AI media-generation company that represents application-layer demand for model inference and video workloads.
- Inceptive: a company working at the intersection of AI and drug discovery, showing NVIDIA’s interest in scientific computing.
- Waabi: an autonomy company that connects the portfolio to transportation and physical-world AI.
- SandboxAQ: a specialized AI and quantum-technology company, broadening the strategy beyond general-purpose language models.
- Cursor, CodeRabbit and LangChain: developer and model-building tools associated with NVIDIA’s wider ecosystem and coalition activity.
These companies should not be presented as equally large, equally funded or equally dependent on NVIDIA hardware. Their strategic significance comes from the different layers they occupy: model distribution, software development, scientific applications, autonomy and inference.
Inflection AI shows why startup outcomes can change
Inflection AI is a useful cautionary example. The company attracted substantial investment and built a large NVIDIA-powered cluster, but Microsoft later hired its founders and licensed its technology. That left Inflection’s future as an independent model company less clear, according to reported coverage.
Calling the investment a “failure” would depend on the metric. Inflection may not have become a dominant standalone consumer AI platform, but its technology, talent, infrastructure and financing could still have generated value for employees, founders, investors or strategic partners.
The case also shows why a portfolio company can remain strategically useful even when its original corporate plan changes. NVIDIA’s ecosystem may gain from the diffusion of talent and technology, not only from a single startup becoming a lasting winner.
What NVIDIA’s filings reveal about scale and risk
$17.5 billion in fiscal 2026
NVIDIA’s fiscal 2026 filing says it invested $17.5 billion in private companies and infrastructure funds, primarily supporting early-stage startups. The figure is broader than a list of individually announced venture checks. It should not be described as the total amount NVIDIA invested in named AI startups.
$18.6 billion in the first quarter of fiscal 2027
NVIDIA’s quarterly filing disclosed $18.6 billion in private-company and infrastructure-fund investments in the first quarter of fiscal 2027. Because this figure comes from a broader financial category, it is not directly comparable to adding together the $2 billion CoreWeave and $2 billion Nebius transactions.
$3.5 billion in infrastructure guarantees
The fiscal 2026 filing also disclosed $3.5 billion in land, power and shell guarantees to early-stage companies, generally over multi-year periods. These are not equity investments. They matter because they show NVIDIA supporting the physical conditions required to operate AI data centers—and accepting exposure to delays, construction problems, power constraints and weak utilization.
NVIDIA warns in its filings that private-company investments can be illiquid, that startups may never become profitable and that it could lose some or all of an investment. The company also reports a developer ecosystem of more than 7.5 million CUDA and related-software developers, which helps explain why its influence extends beyond the ownership stakes disclosed in financial statements.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →The infrastructure flywheel
NVIDIA’s strategy can be understood as an ecosystem flywheel:
- NVIDIA supplies GPUs, networking, systems and software.
- It invests in model developers, application companies and specialized cloud providers.
- Those companies expand AI capacity and make more products available to end users.
- Startups, enterprises and developers consume more NVIDIA-powered training and inference.
- More usage strengthens CUDA, NVIDIA’s libraries, cloud availability and developer familiarity.
- The resulting ecosystem can make it more difficult and expensive for customers to switch platforms.
This is an analytical framework, not proof that every investment was made solely to create hardware sales. Companies have genuine compute needs, and many operate across multiple vendors. But the structure helps explain why NVIDIA may invest in both the company building a model and the cloud provider supplying its compute.
Why the strategy could benefit NVIDIA
- Demand creation: portfolio companies may buy NVIDIA systems directly or rent them through cloud partners.
- Platform adoption: startups using CUDA, TensorRT, networking and NVIDIA deployment tools can reinforce developer familiarity.
- Distribution: companies such as Hugging Face can help NVIDIA-optimized models and tools reach a wide developer audience.
- Early visibility: investment relationships can give NVIDIA insight into emerging workloads, model architectures and infrastructure requirements.
- Physical-AI expansion: robotics, simulation and digital twins open new markets beyond data-center language models.
- Cloud diversification: specialized AI clouds can expand capacity outside the largest hyperscalers.
The risks, conflicts and objections
Concentration risk
NVIDIA’s ecosystem investments may reinforce demand, but they can also concentrate exposure to the same AI-spending cycle. A slowdown in training budgets, falling cloud utilization, model commoditization, custom ASIC adoption or better competing accelerators could affect many portfolio companies at once.
Supplier and investor conflicts
NVIDIA may simultaneously be an investor, supplier, technical partner, customer, guarantor or source of commercial referrals. That combination can create conflicts when NVIDIA helps finance companies that purchase its products or compete with its customers. The existence of a relationship does not establish wrongdoing, but it makes transaction terms, accounting treatment and disclosures important.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Rank #4
- Powered by the NVIDIA Blackwell architecture and DLSS 4 OC mode: 2640MHz/Default mode: 2610MHz (Boost Clock)
- Military-grade components deliver rock-solid power and longer lifespan for ultimate durability
- Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
- 3.125-slot design with massive fin array optimized for airflow from three Axial-tech fans
- Phase-change GPU thermal pad helps ensure optimal thermal performance and longevity, outlasting traditional thermal paste for graphics cards under heavy loads
Circular-financing concerns
The concern is straightforward: NVIDIA invests in an AI lab or cloud provider; that company uses its capital or financing support to expand NVIDIA-based infrastructure; the expansion increases demand for NVIDIA products and may help support the private company’s valuation.
There are also counterarguments. AI labs and cloud providers have real demand from customers, data-center infrastructure requires capital from many sources, and commercial partnerships can reduce deployment friction. The right conclusion is that the arrangement creates an economic feedback loop worth scrutinizing—not that every transaction is artificial or improper.
Dependence on one platform
NVIDIA’s technical support can lower barriers for startups, but close integration with CUDA and NVIDIA-optimized software may increase switching costs. Buyers evaluating AI services should ask whether they need NVIDIA-specific performance and tooling or whether portability across AMD, Google TPU, AWS Trainium, custom silicon and other accelerators matters more.
Illiquidity and startup failure
Private-company investments can be difficult to value and sell. A company can have an impressive funding round and still fail to build a durable business. It can also be acquired, restructure, lose its founders or pivot without becoming worthless to every stakeholder.
How to evaluate claims about NVIDIA’s portfolio
When reading an announcement or investment article, use this checklist:
- Identify the relationship: Is it equity, a partnership, a deployment agreement, a guarantee or Inception membership?
- Separate the check from the round: A $1.7 billion financing round does not mean NVIDIA invested $1.7 billion.
- Check the date and status: Is the transaction closed, announced, planned or still subject to definitive agreements?
- Distinguish capacity from operations: A future gigawatt target is not installed capacity.
- Look for ownership disclosure: Do not infer a percentage stake from a company appearing on an NVIDIA webpage.
- Assess the revenue link: Does the company buy NVIDIA systems, rent NVIDIA-powered compute, or merely use an NVIDIA development tool?
- Consider alternatives: Is the company multi-vendor, or is it structurally dependent on NVIDIA?
- Separate startup and public-company deals: CoreWeave’s later public-company investment is different from an early-stage venture check.
What this means for investors and businesses
For investors, NVIDIA’s startup activity is evidence of an attempt to extend its platform advantage, not a guarantee that every investment will appreciate or that NVIDIA will dominate every AI segment. The most direct financial link may come from infrastructure companies that purchase large amounts of NVIDIA equipment, while model and application investments offer longer-term exposure to usage growth and ecosystem influence.
For businesses choosing AI infrastructure, NVIDIA’s ecosystem can provide mature software, broad developer familiarity and access to specialized cloud capacity. The trade-off is potential vendor lock-in, uncertain pricing during shortages, storage and networking costs, egress charges, and less flexibility if a cheaper or more efficient accelerator becomes available.
GPU hourly pricing alone is not enough for a sound comparison. Businesses should also examine utilization, minimum commitments, reservation terms, geographic availability, data privacy, orchestration support, model-serving tools, egress and the engineering cost of migrating between platforms.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Strategic corporate investments are a separate category
NVIDIA has also made or announced strategic investments involving larger public companies such as Intel, Synopsys, Coherent and Nokia. These may strengthen manufacturing, semiconductor design, optics, networking or telecommunications relationships, but they are not startup investments in the conventional sense.
Keeping them separate prevents a misleading picture in which every strategic corporate stake is added to a venture portfolio. They belong in an analysis of NVIDIA’s broader industrial strategy, not in a simple ranking of AI startups.
Bottom line
NVIDIA’s most important investment is arguably not any single stake. It is the combination of capital, GPUs, networking, CUDA software, cloud capacity, developer distribution, robotics platforms and infrastructure commitments.
OpenAI and other model companies can expand AI usage. CoreWeave and Nebius can provide the capacity to run it. Hugging Face and developer tools can spread the software. Robotics companies can open entirely new workloads. The resulting strategy resembles ecosystem construction more than conventional venture diversification.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →That ecosystem can reinforce NVIDIA’s position, but it also creates concentration, valuation, execution and conflict risks. The clearest way to understand any individual deal is to ask two separate questions: What did NVIDIA actually invest? And what commercial or technical relationship does the investment create? The answers are often related, but they are not the same.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
