Deal Timeline

Plotted by close date where disclosed, otherwise announcement. Select any marker to jump to the deal entry.

The Acquisition Playbook.

Three patterns run through NVIDIA's acquisitions — what it looks for, how it pays, and how it folds in what it buys.

01
Acquisition criteria
Buy software and interconnect that make the GPU more valuable.
Nvidia targets capabilities that deepen its platform — Mellanox for networking, Cumulus and Run:ai for the software stack, DeepMap for autonomous driving.
Bright ComputingExceleroRun:aiSchedMD (Slurm)MediaQ
02
Capital deployment
One landmark deal, many tuck-ins.
Mellanox ($6.9B) is the outlier; most others — Cumulus, Bright Computing, Run:ai, Excelero — are undisclosed capability buys, and the $40B Arm attempt was abandoned.
Bright ComputingExceleroRun:aiSchedMD (Slurm)MediaQ
03
Integration approach
Fold into the data-center platform.
Acquisitions are integrated to sell more full-stack accelerated computing, bundling networking, orchestration and software with the GPU.
Bright ComputingExceleroRun:aiSchedMD (Slurm)MediaQ

The Full Deal Book

28 acquisitions — each with the deal value, financing structure, target revenue, and executive commentary where disclosed.

01 Bright Computing · United States / Netherlands Not disclosed
Announced Jan 2022 Closed Jan 2022
Bright Cluster ManagerHPC cluster provisioning and orchestrationedge-to-cloud management

Bright Computing made software for provisioning and managing high-performance computing systems, with its Bright Cluster Manager used by more than 700 organizations. The software spans Arm and x86 CPUs, NVIDIA GPUs and Kubernetes containers, running at the edge, in datacenters and across hybrid clouds. NVIDIA acquired it to simplify HPC cluster deployment and management.

Why it was attractive
  • Widely deployed cluster manager and 700+ organization install base complementary to NVIDIA DGX/HPC
NVIDIA is changing the world as we know it, and we couldn't be more excited for our team and software to play a part in that.Bill Wagner — CEO, Bright Computing
02 Excelero · Tel Aviv, Israel Not disclosed
Announced Mar 2022 Closed Mar 2022
NVMesh software-defined block storageNVMe flash virtualizationKubernetes support

Excelero, based in Tel Aviv, built high-performance software-defined block storage, including its NVMesh software that pools virtual arrays of NVMe flash for cloud environments and supports Kubernetes. NVIDIA acquired the company to add block-storage capability to its enterprise software stack for HPC and AI clusters.

Why it was attractive
  • Proven high-performance storage software and an experienced Tel Aviv engineering team
The Excelero team is joining NVIDIA as demand is surging for high-performance computing and AI.Yaniv Romem — CEO and Co-founder, Excelero
03 Run:ai · Tel Aviv, Israel Not disclosed
Announced Apr 2024 Closed Apr 2024
Kubernetes GPU orchestrationcluster schedulingGPU pooling and utilization management

Run:ai built a Kubernetes-based workload management and orchestration platform that lets enterprises pool and schedule GPU resources across on-premises, cloud and hybrid environments to raise utilization. NVIDIA, a close collaborator with Run:ai since 2020, agreed to acquire the company to help customers run large, distributed AI compute more efficiently.

Why it was attractive
  • Open
  • Kubernetes-native GPU orchestration used by large enterprises managing datacenter-scale clusters
04 SchedMD (Slurm) · United States Not disclosed
Announced Dec 2025 Closed Dec 2025
Slurm open-source workload schedulingcluster resource allocationHPC/AI job management

SchedMD is the company behind Slurm, the open-source workload manager that queues, schedules and allocates compute across HPC and AI clusters and runs on more than half of the top 100 supercomputers. NVIDIA acquired SchedMD while pledging to keep distributing Slurm as vendor-neutral, open-source software and to accelerate its development for next-generation HPC and AI.

Why it was attractive
  • Stewardship of Slurm
  • the de facto scheduler across supercomputing and large AI training clusters
We're thrilled to join forces with NVIDIA, as this acquisition is the ultimate validation of Slurm's critical role in the world's most demanding HPC and AI environments.Danny Auble — CEO, SchedMD
05 MediaQ, Inc. · United States (Santa Clara, California) $70M
Announced Aug 2003 Closed Aug 2003 Cash acquisition

NVIDIA acquired MediaQ, a provider of graphics and multimedia processors for wireless and handheld mobile devices such as 3G and 2.5G handsets, PDAs and LCD displays. MediaQ's technology focused on enhancing visual display, improving connectivity and minimizing chip- and system-level power consumption, and shipped with drivers for major mobile operating systems. The deal extended NVIDIA's visual computing into the wireless mobile segment.

This acquisition supports NVIDIA's strategy of extending our platform reach and accelerates our entry into the wireless mobile markets.Jen-Hsun Huang — Chief Executive Officer, NVIDIA
06 iReady Corporation · United States (Santa Clara, California) Not disclosed
Disclosed in 8-K Asset acquisition

NVIDIA purchased assets from iReady, a developer of TCP/IP offload and network-interface technology. NVIDIA stated the acquired assets would help expand its server and workstation business opportunities tied to its NVIDIA nForce3 Professional media and communications processors.

07 Mental Images GmbH · Germany (Berlin) Not disclosed
Disclosed in 8-K

NVIDIA acquired Mental Images, a developer of professional rendering and 3D software technology, including the widely used mental ray rendering engine and related tools used in design, visualization and film production. Following the acquisition Mental Images operated as a wholly owned NVIDIA subsidiary and became the basis for NVIDIA's Advanced Rendering technologies.

08 AGEIA Technologies, Inc. · United States (Santa Clara, California) Not disclosed
Disclosed in 8-K

NVIDIA acquired AGEIA Technologies, the developer of the PhysX physics-simulation engine and dedicated physics-acceleration hardware. The acquisition brought PhysX software and the AGEIA engineering team into NVIDIA, and PhysX was subsequently integrated to run on NVIDIA GPUs. AGEIA became a wholly owned NVIDIA subsidiary.

09 The Portland Group, Inc. (PGI) · United States (Beaverton, Oregon) Not disclosed
Disclosed in 8-K

NVIDIA acquired The Portland Group (PGI), a developer of high-performance computing compilers and development tools, including Fortran, C and C++ compilers used in scientific and technical computing. The acquisition strengthened NVIDIA's software tools for GPU-accelerated and parallel computing.

10 SwiftStack · United States (San Francisco, California) Not disclosed
Disclosed in 8-K

NVIDIA acquired SwiftStack, a provider of object- and cloud-storage software and data-management technology for large-scale, multi-cloud environments. SwiftStack's technology and team supported NVIDIA's data-platform and AI-infrastructure efforts, and SwiftStack was reported as a wholly owned NVIDIA subsidiary in NVIDIA's SEC filings.

11 Stexar Corp. · United States (Oregon) Not disclosed
Disclosed in 8-K

NVIDIA acquired Stexar, a microprocessor engineering firm based in Oregon, adding a team of CPU and processor-design engineers. The acquisition primarily brought silicon engineering talent into NVIDIA rather than a commercial product line.

12 RayScale, Inc. · United States Not disclosed
Disclosed in 8-K

NVIDIA acquired RayScale, a developer of GPU-accelerated ray-tracing rendering software. The technology and team were folded into NVIDIA's advanced rendering and professional visualization efforts.

13 OmniML, Inc. · United States (Santa Clara, California) Not disclosed
Disclosed in 8-K

NVIDIA acquired OmniML, a startup specializing in machine-learning model optimization that miniaturizes and adapts AI models to run efficiently across environments from cloud platforms to edge devices. Members of the OmniML team joined NVIDIA's deep learning algorithm and software groups, contributing to model-optimization tooling such as the TensorRT Model Optimizer.

14 Deci AI Ltd. · Israel (Tel Aviv) Not disclosed
Disclosed in 8-K

NVIDIA acquired Deci, a deep-learning company whose platform combined Automated Neural Architecture Construction (AutoNAC) technology with hardware-aware optimization to help developers build, optimize and deploy more accurate and efficient models across edge, data-center and cloud environments. The Deci team was integrated into NVIDIA's AI software organization.

15 Shoreline.io, Inc. · United States (Redwood City, California) Not disclosed
Disclosed in 8-K

NVIDIA acquired Shoreline.io, a developer of automated incident-remediation and reliability software that helps operations teams detect, diagnose and automatically repair infrastructure issues. The technology and team supported NVIDIA's efforts to improve reliability and automation for large-scale computing environments.

16 Brev.dev, Inc. · United States (San Francisco, California) Not disclosed
Disclosed in 8-K

NVIDIA acquired Brev.dev, a startup that helps AI developers discover and provision GPU compute across different cloud providers, simplifying the process of finding and launching GPU instances for development and training. The team joined NVIDIA to support its developer-tooling and cloud efforts.

17 OctoAI (OctoML, Inc.) · United States (Seattle, Washington) Not disclosed
Disclosed in 8-K

NVIDIA acquired OctoAI, formerly known as OctoML, a company focused on model-serving and inference optimization that helps run generative-AI models efficiently and cost-effectively across hardware. The team, with roots in the Apache TVM compiler project, joined NVIDIA's AI software efforts.

18 Gretel Labs, Inc. · United States (San Diego, California) Not disclosed
Disclosed in 8-K

NVIDIA acquired Gretel, a platform for generating safe, high-quality synthetic data for enterprises, addressing privacy and data-scarcity challenges in AI development. Gretel leadership, including its co-founder, joined NVIDIA to lead work at the intersection of synthetic data, privacy and trustworthy AI.

19 CentML Inc. · Canada (Toronto, Ontario) Not disclosed
Disclosed in 8-K

NVIDIA acquired CentML, a company whose software sits between AI models and the chips that run them, optimizing model execution to improve efficiency and reduce inference cost. The CentML team and technology supported NVIDIA's AI compiler and inference-optimization work.

20 Lepton AI, Inc. · United States (Cupertino, California) Not disclosed
Disclosed in 8-K

NVIDIA acquired Lepton AI, a startup providing a platform for renting and managing GPU compute across cloud providers to run AI workloads. The technology and team informed NVIDIA's GPU-cloud and AI-platform initiatives.

21 DeepMap · United States Not disclosed
Announced Jun 2021 Closed Aug 2021
HD mappingcentimeter-level localizationmap update pipelines for AVs

DeepMap was a startup building high-definition maps that let autonomous vehicles localize themselves with centimeter-level precision, using continuously updated maps drawn from vehicle sensor data. NVIDIA acquired the company to fold its mapping and localization technology into the NVIDIA DRIVE self-driving platform.

Why it was attractive
  • Proven HD-map and localization technology plus an experienced AV mapping team
DeepMap is expected to extend our mapping products, help us scale worldwide map operations and expand our full self-driving expertise.Ali Kani — Vice President and General Manager, Automotive, NVIDIA
22 Cumulus Networks · United States Not disclosed
Announced May 2020 Closed Jun 2020
Cumulus Linux switch OSNetQ network analyticsopen networking automation

Cumulus Networks built open networking software, including the Cumulus Linux operating system for network switches plus analytics and management tooling, supporting more than 2,000 customers across roughly 130 hardware platforms. NVIDIA acquired it and placed it inside the networking business unit created by the Mellanox deal, pairing open network software with NVIDIA's switching hardware.

Why it was attractive
  • Mature switch operating system and a 2
  • 000+ customer base to complement Mellanox switching silicon
23 Mellanox Technologies · Yokneam, Israel $6.9B
Announced Mar 2019 Closed Apr 2020 All cash
InfiniBand interconnectshigh-speed Ethernetdatacenter networkingSmartNICs

Mellanox is a pioneer of high-performance datacenter interconnect technology, best known for originating InfiniBand and for a high-speed Ethernet product line. Its interconnects were used in more than half of the world's fastest supercomputers and across many hyperscale datacenters at the time of the deal. NVIDIA acquired it by merger for $125.00 per share in cash. The two companies had a long prior history of collaboration in high-performance computing.

Why it was attractive
  • Interconnect leadership complementary to NVIDIA GPUs
  • combined footprint powered over 250 of the TOP500 supercomputers
The emergence of AI and data science, as well as billions of simultaneous computer users, is fueling skyrocketing demand on the world's datacenters.Jensen Huang — Founder and CEO, NVIDIA
24 Icera · United Kingdom $367M
Announced May 2011 Closed Jun 2011 All cash
3G/4G baseband processorssoftware-defined cellular modemsRF technology

Icera developed baseband processors for 3G and 4G cellular phones and tablets, holding more than 550 patents granted or pending and shipping soft-modem products approved by over 50 carriers worldwide. NVIDIA acquired it for $367 million in cash. The plan was to pair Icera's modems with NVIDIA's Tegra application processor so NVIDIA could offer both of the main processors inside a smartphone.

Why it was attractive
  • 550+ patents and carrier-approved modem IP to complement Tegra in the fast-growing mobile market
25 PortalPlayer · United States $355M
Announced Nov 2006 Closed Jan 2007 All cash
Personal-media-player SoCsfirmwaresecondary-display software

PortalPlayer supplied semiconductors, firmware and software for personal media players and secondary-display-enabled computers. NVIDIA acquired it by merger under an agreement dated November 6, 2006, with each share converted to $13.50 in cash. The deal was positioned to accelerate NVIDIA's handheld product strategy, building on its low-power graphics, TV and video technology already used in portable media devices.

Why it was attractive
  • Handheld/media-player IP and engineering talent to build out a mobile product line
26 Hybrid Graphics Ltd. · Finland (Helsinki) Not disclosed
Closed Mar 2006

NVIDIA completed its acquisition of Hybrid Graphics, a developer of embedded 2D and 3D graphics software for handheld devices. NVIDIA stated the acquisition would enable customers of both companies to deploy rich graphics solutions for the worldwide handheld market.

27 ULi Electronics Inc. · Taiwan (Taipei) $52M
Announced Dec 2005 Closed Feb 2006 Cash acquisition

NVIDIA acquired ULi Electronics, a Taipei-based developer of PC core-logic (chipset) technology as well as multimedia and embedded products. The acquisition supported NVIDIA's platform-solution strategy around its nForce media and communications processors and GeForce and Quadro GPUs, and strengthened its sales, marketing and customer-engineering presence in Taiwan and China. ULi's president and CEO joined NVIDIA to lead MCP sales, marketing and support functions in Asia.

Alex and his team have built a company admired for creating innovative products in the core logic arena. The acquisition provides an opportunity for NVIDIA to leverage a very talented engineering team in order to expand our MCP initiatives while building closer relationships with customers in Asia.Jen-Hsun Huang — President and Chief Executive Officer, NVIDIA
28 3dfx Interactive, Inc. · United States (San Jose, California) Not disclosed
Closed Apr 2001 Asset acquisition

NVIDIA acquired the core graphics assets of 3dfx Interactive, a pioneering consumer 3D-graphics company best known for its Voodoo accelerator product line. The transaction was structured as a purchase of 3dfx's intellectual property, patents and related graphics assets, and it brought a substantial part of 3dfx's engineering team into NVIDIA. NVIDIA completed the acquisition through its wholly owned subsidiary NVIDIA US Investment Company.

More Acquirer Playbooks

See how VectorShift works for your firm

Request Demo