PyTorch

AI Frameworks, Tools & Libraries

PyTorch is an open-source machine learning framework that began as a research project at Meta AI in 2016 and was publicly released in January 2017.

Reviewed by 7wData
Share Email

On this page

Profile

PyTorch is an open-source machine learning framework that provides tools for building, training, and deploying neural networks, primarily used for deep learning research and production AI applications.

PyTorch is an open-source machine learning framework that began as a research project at Meta AI in 2016 and was publicly released in January 2017. It is now governed by the Linux Foundation via the PyTorch Foundation, established in 2022. The framework is the dominant tool for deep learning research and production, competing primarily with TensorFlow and JAX.

As of June 2026, the PyTorch GitHub repository has over 101,000 stars, 27,900 forks, and 105,621 commits. The project does not report standalone revenue; its financial support comes from member contributions to the PyTorch Foundation, which includes Meta, Google, Microsoft, Amazon Web Services, and others. The most recent disclosed community event, the PyTorch Docathon 2026 (May 5–19, 2026), attracted over 260 registrants and resulted in more than 150 merged pull requests.

PyTorch's market position is reinforced by its adoption across major cloud providers and enterprises. In 2025 and 2026, the broader tech industry saw significant layoffs and restructuring, with companies like Intel, Microsoft, and Amazon cutting over 60,000 jobs between June and November 2025, partly to pivot resources toward AI. While PyTorch itself has not reported layoffs, its ecosystem is affected by these shifts.

The framework continues to see strong enterprise adoption, with case studies from LinkedIn (June 2026) and IBM Research (April 2026) published on the PyTorch site. Oracle's Q3 fiscal 2026 earnings (March 2026) showed cloud infrastructure revenue up 84% year-over-year, partly driven by AI workloads that often use PyTorch. Snowflake's fiscal 2026 results (January 2026) reported product revenue of $1.23 billion in Q4, up 30% year-over-year, with over 9,100 accounts using Snowflake AI features, many of which rely on PyTorch-based models.

Track PyTorch and 240+ vendors.

48k+ subscribers read the daily AI & data note. One email, both newsletters. Unsubscribe anytime.

Who buys this

  • Academic researchers and university labs conducting deep learning research
  • Enterprise AI/ML teams building production models for computer vision, NLP, and recommendation systems
  • Cloud service providers (AWS, Google Cloud, Azure, Oracle Cloud) offering managed PyTorch services
  • Startups developing AI-powered products, from generative AI to autonomous systems
  • Government and defense organizations using AI for analysis and simulation

Publicly disclosed clients

  • LinkedIn
  • IBM Research
  • Meta
  • Oracle
  • Snowflake

Strengths and what to watch

Strengths

  • Dominant market share in academic deep learning research, with a large and active open-source community (101k GitHub stars, 105k+ commits)
  • Strong integration with major cloud platforms (AWS, GCP, Azure, Oracle Cloud) and hardware vendors (NVIDIA, AMD, Intel), enabling broad deployment options
  • Active governance under the Linux Foundation's PyTorch Foundation, providing vendor-neutral stewardship and long-term stability

Watch for

  • Dependence on Meta for initial development and ongoing major contributions; any shift in Meta's AI strategy could affect PyTorch's direction
  • Intense competition from JAX (Google) and TensorFlow (Google), which continue to invest heavily in performance and ecosystem
  • Ecosystem fragmentation as specialized frameworks (e.g., vLLM, TensorRT-LLM) and hardware-specific backends (e.g., ROCm, CUDA) create complexity for users

Recent moves

Key Information

Industry
AI Frameworks, Tools & Libraries
Founded
2016

Frequently Asked Questions

What is PyTorch and what is it used for?

PyTorch is an open-source machine learning framework for building, training, and deploying neural networks. It is primarily used for deep learning research and production AI applications, including computer vision, natural language processing, and recommendation systems.

Who maintains and governs PyTorch?

PyTorch is governed by the Linux Foundation via the PyTorch Foundation, established in 2022. It began as a research project at Meta AI in 2016 and was publicly released in January 2017. The foundation includes members like Meta, Google, Microsoft, Amazon Web Services, and others.

How does PyTorch compare to TensorFlow and JAX?

PyTorch is the dominant tool for deep learning research and production, competing primarily with TensorFlow and JAX from Google. It has strong integration with major cloud platforms and hardware vendors, while JAX focuses on high-performance computing and TensorFlow emphasizes production deployment.

What are some notable companies using PyTorch?

Notable clients include LinkedIn, IBM Research, Meta, Oracle, and Snowflake. LinkedIn uses PyTorch for extreme-scale optimization problems, and IBM Research uses it in their RITS Platform. Snowflake reports over 9,100 accounts using AI features that rely on PyTorch-based models.

What was the PyTorch Docathon 2026 and what were the results?

The PyTorch Docathon 2026 was a community event held from May 5 to May 19, 2026. It attracted over 260 registrants and resulted in more than 150 merged pull requests, improving documentation and contributing to the framework's ecosystem.

Is PyTorch free to use and what is its GitHub activity?

Yes, PyTorch is free and open-source. As of June 2026, its GitHub repository has over 101,000 stars, 27,900 forks, and 105,621 commits, reflecting a large and active community. Financial support comes from member contributions to the PyTorch Foundation.

Sources

  1. github.com — GitHub repository stats (stars, forks, commits, branches, tags) as of June 2026
  2. pytorch.org — Docathon 2026 results: 260+ registrants, 150+ merged PRs, top contributors
  3. pytorch.org — Case studies page listing LinkedIn (June 2026) and IBM Research (April 2026) as notable users
  4. investor.oracle.com — Oracle Q3 FY26 earnings: cloud infrastructure revenue up 84% YoY, total cloud revenue $8.9B
  5. www.snowflake.com — Snowflake FY26 Q4 earnings: product revenue $1.23B, 30% YoY growth, 9,100+ AI accounts
  6. www.informationweek.com — Tech layoffs tracker: 60,000+ jobs cut June-November 2025, AI-driven restructuring at Intel, Microsoft, Amazon