I was at first excited to see this announcement by the ACM. The good news is that the ACM Digital Library is now “Gold Open Access,” meaning that anyone can now access all articles for free. It used to be paywalled, with a noticeable fee per article or subscription cost per year. If I read the policy correctly, images published can also be reused, depending on Creative Commons licensing. See this page for more on that.
The not-so-great news is that the ACM has put Article Processing Charges (APCs) in place. For each article accepted to many ACM publications, the authors are charged from $700 to $1800. Many institutions have joined ACM OPEN, in which case fees are waived. Some have not, in both the U.S. and in other countries – see the map. It used to be free to publish. We’ll see what effect this change has on where submissions come from.
I understand that the ACM needs to fund the Digital Library somehow. Only a few publications in our field are Diamond Open Access, where there are no fees for readers or submitters. The three I know are JCGT, JoCG, and WSCG. “Gold Open Access” means free for readers, not for creators.
What bothers me about the recent ACM letter is what’s left out. The passage, “As of January 1, 2026, ACM completed its transition to full open access” is putting a lot of work on the word “full.” Yes, for readers, Gold Open Access is definitely better than the paywall, though graphics people could usually find preprints online by searching Google Scholar or our page (still useful for finding related materials to the research). I would have appreciated seeing a more realistic assessment, e.g., “we’ve shifted the costs; there are likely new winners and losers, and we will continue to work on understanding and implementing what the community considers to be fair.” This (long!) post is about how costs have shifted and what it might mean to authors.
Overall I massively appreciate the work that people at the ACM, many of them volunteers, did to make free-for-consumers happen. It’s a win-win in that more people can access articles, and authors’ work then has a greater impact. Over the past few years things at the ACM have been moving in generally good directions, e.g., in 2023 authors no longer had to sign over to the ACM their copyright (including the right to reuse their own images). This latest move to Gold Open Access is good for consumers; I’m not so sure about all creators, as fees can limit those at some institutions, and certainly those authors who are independent.
I wrote the ACM Director of Publications, Scott Delman, about the shifting of costs from consumers to creators. He provided thoughtful and extensive replies to my queries. He notes that there are a few important modifiers to the new APCs is that. First, these APCs will be discounted for 2026. Many developing nations, listed here, will not be charged at all. Finally, creators can apply for financial hardship. More on this below, in his reply.
Scott also points out these two documents on the ACM’s publication finances: 2022 and 2023&2024. He gives an in-depth run-down of the changing finances in his latest reply. I appreciate his responsiveness, along with the thought and effort he and others have put into this overarching question.
Most of Scott’s first reply follows.
Since that transition [to Gold Open Access] began in 2020, we have transitioned approximately 75% of the roughly $28M in annual expenses to run ACM’s Publications and Digital Library from institutional subscriptions to read and download ACM Publications in the DL over to ACM Open institutional licenses to support the publication of OA articles affiliated with those institutions so that anyone can read and download articles from the ACM DL. There are now just under 3,100 institutions paying to support the new OA model.
Unfortunately, there remain 300-400 institutions that have decided not to transition over to the new model (yet) globally, requiring the payment of APCs for approximately 25% of the peer reviewed research articles we will publish in 2026. To minimize the impact on authors in 2026 and to give more time for the remaining 300-400 institutions to join ACM Open, ACM took several important steps.
- ACM decided to significantly subsidize the APC costs in 2026. Instead of the $700 / $1,000 cost (which is the amount required to be financially sustainable) of a conference article, ACM is charging $250 / $350 in 2026. For journal articles, the APC pricing of $1,300 / $1,800 is being subsidized at the rates of $950 / $1,450 in 2026.
- ACM implemented a geographic waiver and discount policy to address concerns about equity and inclusivity for authors from over 80 developing countries. See https://www.acm.org/publications/policies/policy-on-geographic-apc-waivers-and-discounts.
- ACM implemented a Financial Hardship Waiver program to address concerns about authors in unique and difficult financial situations. See https://www.acm.org/publications/policies/policy-on-discretionary-open-access-apc-waivers.
It is our goal to continue bringing more institutions into the ACM Open program with the help of the community over the next few years to minimize the number of authors that will need to pay APCs. One of the real challenges we have had over the past 5 years is convincing Chinese universities to join the ACM Open program. As you may know, Chinese authors represent over 30% of the papers ACM published in 2024 (and the number likely grew in 2025), so Chinese authors represented a significant percent of the remaining 25% of papers not covered by ACM Open. But since there is no national-level Open Access mandate in China, it has been difficult to convince institutions to join. That said, all indications to date are that Chinese computer scientists have funding to pay APCs.
So, why are we requiring APCs? If ACM were to not require APCs from authors at institutions that have yet to join, most or all of the 3,100 institutions that have joined ACM Open would likely cancel over time, and the model would be unsustainable. Without approximately, $7-8M in APC income in 2026, ACMs Publications program would not be financially sustainable.
I hope the above is helpful and explains why APCs are necessary at this stage, even though we all agree it would be far better to avoid them in favor of institutional licenses.
I replied, noting, “One concern I’ve heard from someone associated with one educational institution (I haven’t been able to confirm it) is that the new fee structure for them is overall more expensive than the previous ACM DL institutional subscription.” Much of Scott’s reply follows.
It is absolutely true that many institutions are paying more now than when they previously subscribed to the ACM DL when there was a subscription paywall. Some pay substantially more. Let me explain.
As you’ll see from the two links above to articles about ACM Publications Finances, it costs ACM about $28-29M annually to run its publications program currently, including all of the costs of running the ACM Digital Library. This goes up every year due to inflation, significant increases in the size of the program, costs associated with running the Digital Library, etc. Back in 2020 when we started the transition to OA, we published about 20k research articles. We now publish more than twice that amount, and there are real costs associated with that increase.
For ACM to be financially sustainable over the long term, the Publications program needs to, at the very least, cover its expenses. Generally speaking, the only way to do that in the past was by selling DL subscriptions to institutions (universities, companies, government research institutes, etc). We did this consistently for decades and had approximately 2,400 – 2,500 institutions that subscribed with a very high renewal rate of 96-97%. Out of this group, about 800 of these were heavy users of the DL. They accessed and downloaded papers from the DL consistently. The other 1,600 used the DL but not nearly as much as the top 1/3. But, together all of these institutions generated enough income for ACM to be sustainable. At the same time, the top 1/3 generated about 1/3 of the income and the bottom 2/3 generated about 2/3 of the income. Our pricing over the years was not based on usage. Also important to note is that the top 1/3 published about 85% of the papers published by ACM each year with the bottom 2/3 publishing only about 15%. This reality was the largest driving force in the model we developed in close collaboration with MIT, CMU, University of California, and a few other large research institutions between 2018-2020.
When we announced the transition to full Open Access over 5 years ago, the value proposition completely changed, since suddenly the bottom 2/3 of institutions that generate 2/3 of our income annually would have no incentive to continue paying ACM (since the DL would become OA in 2026), so we had to assume that many of the non-publishing institutions would either cancel or only agree to pay significantly less, which meant that approximately 2/3 of the income required to sustain ACM Publications was at risk. This meant that the top 1/3 of institutions would need to shoulder a much larger percentage of the financial burden.
The ACM Open model had two main challenges….the first was to convince institutions that publish a lot with ACM to cover those publication costs, so that individual authors affiliated with those institutions wouldn’t need to pay APCs. Institutions that publish more pay more and ACM’s pricing table reflects this transparently (see here).
The second challenge was to convince the 2/3 of institutions that don’t publish with ACM actively to continue supporting the overall model by providing a value proposition that makes sense to these institutions. This is the main reason ACM took the decision to create two versions of the DL platform. Without this decision a large % of the 2/3 of institutions would likely cancel over time and the sustainability of the transition to Open Access would be at risk. It is also worth noting that while the top 1/3 of institutions are paying “more”, 2/3 of institutions are paying less than they historically paid for access to the DL. Generally speaking most institutions in developed countries were paying between $6,000 – $9,000, and the bottom 2/3 are now paying between $2,500 – $6,000. We also have about 300-400 institutions that have not yet decided what they will do, and there is a real risk of cancellations from this group. For these institutions, APCs need to be paid by authors to make up for the lost income if they ultimately decide not to participate in ACM Open.
I hope this long-winded explanation makes sense to you, even if you or others may not agree with the explanation or all of the decisions ACM has made throughout this transition.