Company lists · Data refreshed September 2026

Top 25 Neural network libraries worldwide

RevenueBase identified 127 neural network libraries worldwide as of September 2026 and tracks 2,961 professional contacts across them, including 550 with verified work emails. The United States is the largest market and San Francisco the largest hub.

This list covers neural network libraries worldwide, matched on what each company actually does as described in its own words. Ranked by team size, largest first. Every row shows the company's own description, its RevenueBase Smart Search rating, and when it was last checked. Data refreshed September 20, 2026. Country and city lists for neural network libraries follow the table.

Published by RevenueBase, a B2B data infrastructure company. This list was built with RevenueBase Smart Search, the natural-language company search RevenueBase customers use to build their own targeted lists and unlock verified emails and direct dials at the companies on them. How these lists are built and judged.

How to read this list HQ checked on every row RevenueBase Smart Search rating on every row Refreshed Sep 20, 2026 86% judged on-target in blind review
127Companies
2,961Contacts tracked
550Verified emails
Explore this list → Free signup · no credit card · 500 free credits

The 25 largest neural network libraries worldwide

Ranked by team size
Top 25 of 127 neural network libraries worldwide, ranked by team size, largest first. Data refreshed Sep 20, 2026. RevenueBase Smart Search rates how each company's information relates to neural network libraries: strong evidence (its own description names it), partial (names part of it), indirect (matched on the rest of its profile rather than the description shown). Click a contact count to view that company's people.
#CompanyRevenueBase Smart SearchHeadquartersEmployeesContacts
1Indirect evidenceRedwood City, CA, United StatesHQ checked Sep 20, 2026201-500967 contacts →

Indirect evidence

“Snorkel AI is the frontier AI data lab, helping teams build the data and environments behind high-performing frontier and agentic AI.”

Source: RevenueBase Database · checked Sep 20, 2026
2Indirect evidenceParis, United StatesHQ checked Sep 20, 2026n/a608 contacts →

Indirect evidence

“The AI community building the future.”

Source: RevenueBase Database · checked Sep 20, 2026
3Indirect evidenceSan Francisco, CA, United StatesHQ checked Sep 20, 2026501-1,00054 contacts →

Indirect evidence

“An open source machine learning framework that accelerates the path from research prototyping to production deployment.”

Source: RevenueBase Database · checked Sep 20, 2026
4Indirect evidenceAustin, TX, United StatesHQ checked Sep 20, 202611-5032 contacts →

Indirect evidence

“Nous Research is an AI research lab creating world-class open source AI.”

Source: RevenueBase Database · checked Sep 20, 2026
5Indirect evidenceBellevue, WA, United StatesHQ checked Sep 20, 2026n/a31 contacts →

Indirect evidence

“Every company is becoming an AI company.”

Source: RevenueBase Database · checked Sep 20, 2026
6Indirect evidenceSan Francisco, United StatesHQ checked Sep 20, 202611-5029 contacts →

Indirect evidence

“Essential AI is on a mission to make frontier AI more accessible through an open, collaborative platform.”

Source: RevenueBase Database · checked Sep 20, 2026
7Indirect evidenceAnn Arbor, MI, United StatesHQ checked Sep 20, 202611-5028 contacts →

Indirect evidence

“Jaseci Labs is composed of a team of AI professors, Ph.D.s and industry experts with a proven track record of developing and launching award-winning AI products for millions of…”

Source: RevenueBase Database · checked Sep 20, 2026
8Indirect evidenceMünchen, GermanyHQ checked Sep 20, 202611-5025 contacts →

Indirect evidence

“At Pruna AI, we build high-performance AI models that combine speed, quality, and specialization for key use cases.”

Source: RevenueBase Database · checked Sep 20, 2026
9Indirect evidenceSingaporeHQ checked Sep 20, 202611-5020 contacts →

Indirect evidence

“Gradient is an AI research lab building open infrastructure for the future of AI.”

Source: RevenueBase Database · checked Sep 20, 2026
10Strong evidenceUnited StatesHQ checked Sep 20, 202611-5013 contacts →

Strong evidence

“Shodh AI builds an efficient infrastructure for training large neural networks.”

Source: RevenueBase Database · checked Sep 20, 2026
11Indirect evidenceQuébec, QC, CanadaHQ checked Sep 20, 20261-1010 contacts →

Indirect evidence

“At Tracel, we’re solving the challenge of AI model portability and efficiency across platforms, making deployment simpler and freeing teams from hardware constraints.”

Source: RevenueBase Database · checked Sep 20, 2026
12Indirect evidenceToronto, ON, CanadaHQ checked Sep 20, 202611-509 contacts →

Indirect evidence

“Crater Labs is a Toronto-based research laboratory and an artificial intelligence and machine learning consulting company.”

Source: RevenueBase Database · checked Sep 20, 2026
13Indirect evidenceUnited StatesHQ checked Sep 20, 20261-108 contacts →

Indirect evidence

“Helping organizations find PMF faster.”

Source: RevenueBase Database · checked Sep 20, 2026
14Indirect evidencePorto Alegre, BrazilHQ checked Sep 20, 20261-108 contacts →

Indirect evidence

“Our mission is to make advanced AI infrastructure open, secure, and scalable for everyone.”

Source: RevenueBase Database · checked Sep 20, 2026
15Indirect evidenceUnited StatesHQ checked Sep 20, 202611-507 contacts →

Indirect evidence

“Power your AI Open-source tools to build, test, and deploy Artificial Intelligence”

Source: RevenueBase Database · checked Sep 20, 2026
16Strong evidenceUnited StatesHQ checked Sep 20, 20261-106 contacts →

Strong evidence

“We’re an AI research company building a neural network to model all human output.”

Source: RevenueBase Database · checked Sep 20, 2026
17Indirect evidenceUnited StatesHQ checked Sep 20, 20261-106 contacts →

Indirect evidence

“At Axolotl, we lead the way in open source software for fine-tuning AI models, providing developers with unmatched flexibility, customization, and ease of use across a wide range…”

Source: RevenueBase Database · checked Sep 20, 2026
18Indirect evidenceUnited StatesHQ checked Sep 20, 20261-105 contacts →

Indirect evidence

“Taking Silicon Valley-grade AI to the world, through repeatable, reliable and robust AI building blocks.”

Source: RevenueBase Database · checked Sep 20, 2026
19Indirect evidenceParis, FranceHQ checked Sep 20, 20261-104 contacts →

Indirect evidence

“Accelerate scientific discovery with automated deep learning.”

Source: RevenueBase Database · checked Sep 20, 2026
20Indirect evidenceBellevue, WA, United StatesHQ checked Sep 20, 20261-104 contacts →

Indirect evidence

“A models-research and infrastructure company building the OSS stack for AI-native developers”

Source: RevenueBase Database · checked Sep 20, 2026
21Indirect evidenceSan Francisco, CA, United StatesHQ checked Sep 20, 20261-104 contacts →

Indirect evidence

“fast.ai is dedicated to making the power of deep learning accessible to all.”

Source: RevenueBase Database · checked Sep 20, 2026
22Indirect evidenceIstanbul, TürkiyeHQ checked Sep 20, 202611-503 contacts →

Indirect evidence

“Meet Organon Analytics, the game-changer in the world of Machine Learning.”

Source: RevenueBase Database · checked Sep 20, 2026
23Indirect evidenceCampinas, BrazilHQ checked Sep 20, 20261-103 contacts →

Indirect evidence

“While designing new processors to run ML models can lead to improved performance, making effective usage of processor features in a clean software toolchain can become a…”

Source: RevenueBase Database · checked Sep 20, 2026
24Indirect evidenceDE, United StatesHQ checked Sep 20, 20261-102 contacts →

Indirect evidence

“Privacy first Decentralized AI Ecosystem”

Source: RevenueBase Database · checked Sep 20, 2026
25Indirect evidenceSan Francisco, CA, United StatesHQ checked Sep 20, 20261-102 contacts →

Indirect evidence

“We enable AI at scale.”

Source: RevenueBase Database · checked Sep 20, 2026

Showing 25 of 127 neural network libraries worldwide

See the full list, filter by size, metro and specialty, and get verified emails and direct dials for all 550 reachable decision-makers.

Free signup · no credit card

View the full list →

Neural network libraries worldwide at a glance

As of Sep 20, 2026
18.6%Verified-email coverage
23.3Average contacts per company
1-10Median size (employees)
7.9%Largest-city concentration · San Francisco
Sep 20, 2026Page data refreshed

Companies by employee count

1-10
87
11-50
30
51-200
5
201-500
3
501-1,000
1
10001+
1

Email recency is reported on each country list linked below.

How this list was built

Company categories are rarely black and white. RevenueBase reads each company's own description of its business to decide whether it belongs among neural network libraries, so some rows will be adjacent to the category rather than pure plays. The description and RevenueBase Smart Search rating on every row are there so you can check each match yourself.

Source
The RevenueBase company graph: 60 million companies and 398.9 million professional contacts resolved from primary sources, queried with RevenueBase Smart Search, the natural-language company search available to every RevenueBase user.
Coverage
Every country. Country and city lists for neural network libraries are linked below.
What qualifies
Companies worldwide whose own description of what they do matches “neural network libraries”. Matching reads each company's description, not its industry code or its name.
RevenueBase Smart Search
RevenueBase Smart Search is how this list is assembled and how each row is rated. RevenueBase's data engine and AI assess a company's LinkedIn profile, information supplied directly by the company, its website and other publicly available information, then relate it to a category defined in a user's own words. Each row is rated strong evidence (its own description names neural network libraries), partial evidence (it names part of the category) or indirect evidence (the description shown names none of it and the match rests on the rest of the company's profile). The description supporting each rating is quoted on the row.
What we exclude
Companies whose own description does not clearly place them among neural network libraries, and records with no verified work email.
Match quality
In blind human judging of this matching method, 86% of matched companies were on target. Expect roughly 1 in 7 rows to be adjacent to the category rather than a pure play. Every row shows the company's own description and its RevenueBase Smart Search rating so you can judge each match yourself. Read the full methodology: how the lists are built, how the judging worked, and what the figure means.
Ranking
Team size, largest first: companies are ordered by the number of professionals RevenueBase tracks at each one. Employee ranges shown are reported bands, for context.
Verification
Every listed company carries at least one verified work email, 550 across the full list. Contacts re-verify on a rolling cadence, and each contact's own verification date is shown inside RevenueBase.
Refresh
This page's figures were pulled from the company graph on September 20, 2026. Lists refresh periodically as the graph is re-checked.
Corrections
Wrong category, wrong headquarters, or a company that has closed? Use “Report an issue” on any row. Reports go to the RevenueBase Data Team and feed back into the company graph.

Neural network libraries by location

1 lists · largest first

Every Neural network libraries list RevenueBase publishes, by country and city. Counts are matching companies in each list as of its own refresh date.

Countries

Frequently asked

What is RevenueBase Smart Search?

RevenueBase Smart Search is how RevenueBase decides which companies belong on this list and how strongly each one fits. Its data engine and AI assess a company's LinkedIn profile, information supplied directly by the company, its website and other publicly available information, then relate it to the category the list is about. Each row is rated strong evidence when the company's own description names neural network libraries, partial evidence when it names part of the category, and indirect evidence when the description shown names none of it and the match rests on the rest of the company's profile.

How many neural network libraries are there worldwide?

RevenueBase identified 127 neural network libraries worldwide as of September 2026, matched on each company's own description of what it does rather than its industry code.

What is the largest company on this list?

Snorkel AI, headquartered in Redwood City, tops this list with the largest team on record.

Which country has the most neural network libraries?

The United States leads with 30 companies. Every neural network libraries list by country and city is linked on this page.

Which city has the most neural network libraries?

San Francisco leads with 10 headquartered companies, followed by Bangalore (3) and London (3).

How do I get contact information for these companies?

A free RevenueBase account lets you run this exact search and work the top results, verified emails and direct dials included. Upgrading unlocks all 550 verified decision-maker contacts tracked across these 127 companies.

How current is this list?

The list refreshes periodically as the company graph is re-checked. This page's figures were pulled on September 20, 2026.

Ready to reach these companies? Verified emails and direct dials for all 127 neural network libraries worldwide. Start with a free account, no credit card.

Download this list →

Related lists

About this data

RevenueBase is a B2B data infrastructure company. Its company graph covers 60 million companies and 398.9 million professional contacts, with emails verified on a rolling cadence rather than scraped once and left to decay. These company lists are generated directly from that graph with RevenueBase Smart Search, the same natural-language company search every RevenueBase user has: describe a market in plain words, get the companies that match, and add verified emails and direct dials for the people at them. It is the same data that powers RevenueBase's B2B data products, and a free account includes 500 credits to start. How these lists are built and judged.