Nvidia

Nvidia was founded in Sunnyvale, California by Jensen Huang, Chris Malachowsky, and Curtis Priem, initially focusing on graphics chips for the PC gaming market.

Nvidia went public on the Nasdaq under ticker NVDA, raising capital to fuel its growth in the graphics chip market at a modest market cap of $73.5 million.

Nvidia launched the GeForce 256, which it marketed as the world's first 'Graphics Processing Unit' (GPU), introducing onboard transformation and lighting (T&L) to consumer 3D hardware.

Nvidia completed the acquisition of 3dfx Interactive's intellectual property and graphics assets after announcing the deal in December 2000. 3dfx was a pioneer in consumer 3D graphics and its Voodoo line had dominated gaming GPUs in the late 1990s. The acquisition eliminated Nvidia's primary rival and consolidated the company's position as the leading graphics chip maker.

Nvidia launched CUDA (Compute Unified Device Architecture), a parallel computing platform that allowed GPUs to be used for general-purpose computing, revolutionizing scientific computing and AI.

Nvidia introduced the Fermi architecture, the first GPU microarchitecture designed specifically for general-purpose GPU computing (GPGPU). Fermi brought full support for C++ and enabled high-performance computing (HPC) applications, establishing Nvidia's Tesla line as a serious platform for scientific computing and supercomputing workloads.

Alex Krizhevsky used two Nvidia GeForce GTX 580 GPUs to train AlexNet, a deep convolutional neural network that won the ImageNet competition by a landslide, reducing error rates by over 10 percentage points. This breakthrough demonstrated that GPU-accelerated deep learning could achieve superhuman accuracy in image recognition, sparking the modern AI revolution and making Nvidia GPUs the de facto standard for AI research.

Nvidia launched the GeForce GTX 980 and GTX 970 based on the Maxwell architecture (GM204), delivering dramatically improved power efficiency and performance over the previous Kepler generation. Maxwell introduced advanced voxel-based global illumination for more realistic lighting in games. The architecture became a dominant force in gaming GPUs and was widely adopted across desktop and laptop platforms.

Nvidia launched the DGX-1, the world's first purpose-built AI supercomputer, integrating eight Tesla P100 GPUs based on the Pascal architecture. The DGX-1 delivered the equivalent of 250 conventional servers in a single box and was offered to early AI researchers including a landmark delivery to the founding team of OpenAI.

Nvidia unveiled the Volta architecture and Tesla V100 GPU, featuring dedicated Tensor Cores for deep learning matrix operations. The V100 delivered 120 TFLOPS of deep learning performance and became the most widely deployed GPU in cloud and enterprise AI data centers worldwide, powering the majority of AI model training for years.

Nvidia launched the GeForce RTX 20 series based on the Turing architecture, the first consumer GPUs with dedicated RT Cores for real-time ray tracing and Tensor Cores for AI-powered rendering. The RTX platform introduced a new era of cinematic-quality graphics in real-time games, establishing ray tracing as the next frontier in gaming visuals.

Nvidia completed its $7 billion acquisition of Mellanox Technologies, gaining high-performance networking technology critical for data center and AI infrastructure.

Nvidia's proposed $40 billion acquisition of Arm Holdings from SoftBank was abandoned after facing intense regulatory scrutiny from regulators in the United States, United Kingdom, and European Union over antitrust and national security concerns. The collapse marked one of the largest failed semiconductor acquisitions in history, leaving Arm to pursue an independent public listing.

Nvidia launched the Hopper architecture and H100 GPU, featuring a Transformer Engine optimized for generative AI models like GPT. The H100 became the most sought-after computing product in the world during the AI boom of 2023-2024, with orders backlogged for months and a gray market premium, cementing Nvidia as the dominant supplier of AI infrastructure.

Nvidia became the seventh U.S. company to reach a $1 trillion valuation, driven by surging demand for its AI chips amid the generative AI boom.

Nvidia was added to the Dow Jones Industrial Average, replacing Intel, in a landmark recognition of the company's transformation from a graphics chip maker into a central player in AI and computing. The DJIA inclusion reflected Nvidia's trillion-dollar valuation and its status as a bellwether for the technology industry. Nvidia became the first semiconductor company focused on AI and graphics to join the blue-chip index.

During the AI boom, Nvidia became the first company in the world to surpass $4 trillion in market capitalization, cementing its status as a dominant force in the tech industry.

Nvidia became the first company in history to surpass a $5 trillion market capitalization, briefly displacing Apple and Microsoft as the most valuable company in the world. The milestone was driven by insatiable demand for AI training and inference GPUs, with Blackwell architecture chips powering the majority of new AI data center deployments worldwide.

At GPU Technology Conference 2026, Jensen Huang unveiled the Vera Rubin AI platform, the next-generation GPU architecture following Blackwell. Huang announced that Nvidia had $1 trillion in expected GPU orders through 2027 across Blackwell and Vera Rubin chips. AWS committed to deploying more than 1 million Nvidia GPUs across global cloud regions. The Vera Rubin R100 chip featured 336 billion transistors and 50 petaflops of AI performance.

Nvidia secured a long-term high-bandwidth memory supply commitment valued at approximately USD 750 billion with SK Hynix and other U.S.-led customers, part of USD 950 billion in combined South Korean AI chip supply agreements announced at a presidential summit in San Francisco. The agreement locks in HBM4 memory supply for Nvidia Vera Rubin-generation AI accelerators, with first AI data center deployment targeted for 2027, and includes a USD 500 billion-plus AI data center initiative between Nvidia and SK Group. The commitment secures Nvidia memory supply through the end of the decade amid fully sold-out HBM capacity.

Nvidia announced strategic partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR to establish independent compute financing platforms targeting over USD 500 billion for AI infrastructure buildout. The deal transforms Nvidia compute into an investable asset class, enabling long-duration usage-linked revenue while supporting the construction of AI factories globally. CEO Jensen Huang called it an important milestone, saying Nvidia has evolved from building chips to helping create a new class of productive, investable infrastructure.

Nvidia agreed to pay USD 6 billion to non-exclusively license AI coding models from startup Poolside and invested an additional USD 1 billion at a USD 12 billion pre-money valuation. About 109 Poolside employees also received job offers from Nvidia. The unusual deal structure combines technology licensing, investment, and talent acquisition without a formal company acquisition, potentially creating a new template for Big Tech acquisition of high-value AI startups. The arrangement deepens Nvidia involvement beyond chips and infrastructure into the software and model layer, particularly AI-assisted coding.

Nvidia announced NVHBM, a custom high-bandwidth memory architecture that integrates the memory controller into the HBM base die rather than the accelerator compute die. Compared to HBM4E, Nvidia claims NVHBM delivers 30 percent more memory bandwidth with 15 percent lower power consumption, and frees up to 25 percent more xPU die area for compute. NVHBM is available to NVLink Fusion partners, with AWS Annapurna Labs named as the first announced collaborator.

Amazon Web Services and Nvidia announced a major expansion of their strategic collaboration to deploy 2 million additional Nvidia GPUs across AWS global infrastructure during 2027 and 2028. The deployment will span Nvidia Blackwell Ultra, Rubin, and Rubin Ultra platforms. The deal also brings Nvidia Vera CPUs to AWS, includes construction of AI factories for the U.S. government featuring 100,000 GPUs on secure AWS infrastructure, and expands co-engineering across networking, open models, data processing, and robotics.

Nvidia confirmed it has acquired Hugging Face, the open-source AI platform hosting over 3 million models and used by more than 18 million developers, for USD 12.93 billion -- the largest acquisition in Nvidia's history and its biggest move beyond chip hardware into the AI software ecosystem. CEO Jensen Huang said Hugging Face would continue supporting open source and open-weight models, a commitment met with skepticism from parts of the open-source community wary of consolidation around Nvidia's expanding AI stack.

The U.S. Justice Department opened an investigation into whether Nvidia structured its licensing deal with AI-chip startup Groq, valued by later reports at USD 17-20 billion, to avoid antitrust merger review. As part of the December arrangement, Groq founder Jonathan Ross and COO Sunny Madra moved to Nvidia while Groq remained a separate company; the DOJ has sent Nvidia a formal demand for information. The probe tests whether a license-plus-talent deal that looks like an acquisition without an HSR filing constitutes a stealth takeover.