Sustaining our stewardship of the open-source data science community
Apr 30, 2020By Peter Wang
This week we instituted a change to the Terms of Service for the Anaconda package repository at repo.anaconda.com, as a response to the dramatic increase in usage and adoption of Python and Anaconda among commercial users over the last few years. Moving forward, we are asking commercial users who heavily rely on our packaging & distribution efforts to purchase our commercial repository software, available in Anaconda Team Edition or Enterprise Edition.
By "heavy commercial usage", we are referring to those who are doing whole-scale mirroring of our repository, or whose CI/CD and deployment systems are regularly making thousands of package download requests to our repository. (Simple traffic analysis shows that this is a large portion of the Global 2000.) At this point, Anaconda solely bears the cost of packaging, testing, and petabytes of monthly bandwidth costs. I trust that our significant commercial users understand that it is no longer tenable for us to provide this service at zero cost.
This does not apply to non-commercial users (e.g. students, academics, hobbyists). Even if you are using Anaconda for commercial purposes, if you are just doing small one-off research projects or just learning data science, this is no real burden to our systems, and these changes do not apply to you either.
Similarly, our customers who have paid installations of Anaconda Team Edition or Anaconda Enterprise Edition are not impacted by this change. This change only affects non-licensed, significant commercial usage of our public cloud repository.. We will soon start reaching out to such companies to help them see the benefits of using our commercial products, which provide best-in-class security and governance of open data science software, in any cloud, or on-prem in even the strictest of air-gapped environments. We will engage constructively with them to understand their needs and create a plan to support their usage.
Since taking over the CEO role several months ago, I have been refocusing the company around a mission of stewardship and support of the open source data science community. Much has been written about the "sustainability crisis" in open source, and I believe Anaconda can play a key role in helping fix the broken economics.
For almost a decade now, Anaconda has invested millions of dollars every year into OSS innovation and maintenance of projects such as Pandas, Dask, Numba, Bokeh, HoloViews, Panel, and Intake. I have always wanted to support more projects, but it wasn't clear what kind of monetization the market would support.
Based on current traffic and usage patterns we see, we anticipate that our repository & enterprise products will soon reach a point of profitability. As this materializes, we plan to invest a significant percentage of those profits back into the broader open-source community, though at this early stage we cannot yet commit to an exact dollar amount. It is my sincere hope that over time, this investment will grow to become a significant source of sustainable funding for the open data science and ML ecosystem.