By mid‑2026, scientific publishing is in a transition phase: open access (OA) already accounts for roughly half of global article output, with gold/hybrid OA growing fast and diamond OA stable but still a small share [1][2][3]. AI is rapidly reshaping workflows, peer review, and evaluation, while funders and governments are tightening rules on costs and transparency [1][2][4][5].

Given these trajectories, the dominant models in 2027 are highly likely to be:

• APC‑funded gold and hybrid open access, heavily mediated by transformative agreements and read‑and‑publish deals

• Subscribe‑to‑Open (S2O) and other collective funding models in specific disciplines

• A growing but still minority role for diamond OA and community‑funded infrastructure

• AI‑integrated publishing workflows layered on top of these economic models (rather than replacing them)

Key factors

1. Current OA market structure

• Gold + hybrid OA dominates paid OA revenue

• Global OA journal revenue grew from $1.8B (2022) to $2.1B (2024) and is projected to reach $3.2B by 2028, driven largely by APC‑based gold and hybrid OA [2].

• Gold OA's share of OA output rose from 29% (2015) to 45% (2024), with hybrid OA gaining 8 percentage points over the same period [3].

• Among OASPA members, hybrid OA accounted for about 25% of OA articles in 2023, up from ~21% in 2022, while fully OA journals' share fell from 79% to 75% [4].

• Diamond OA remains niche but stable

• Diamond OA accounts for roughly 5–6% of publications, characterized in recent analysis as a ""plateau"" rather than a growth engine [5].

• Scaling is constrained by funding, infrastructure, and editorial capacity; successful models (e.g., LIPIcs, SciPost) tend to be discipline‑anchored, community‑governed ""boutique"" venues rather than mass‑market replacements [5].

2. Transformative agreements and S2O

• Transformative/read‑and‑publish agreements are now central in many national consortia (e.g., b‑on, CSU, multiple European consortia), covering thousands of hybrid titles and annual quotas through 2025‑2027 [6].

• These deals convert a large share of subscription reading spend into OA publishing spend, effectively locking in hybrid + gold OA as the default for many institutional authors through at least 2027.

• Subscribe‑to‑Open (S2O) is spreading:

• Royal Society moved eight subscription journals onto S2O from 2026, publishing OA when library revenue thresholds are met [7].

• Annual Reviews, EMS Press and others have successfully run S2O programs, and S2O appears frequently in 2025–2027 agreement portfolios [6][7].

• S2O is especially attractive where author‑pays APCs are politically or economically difficult, and thus is likely to dominate OA conversion in some society and math/physics communities rather than globally.

3. Pressure on the APC model and funder responses

• OA market analyses highlight rapid APC inflation (hybrid APCs topping $12,850; fully OA maxima near $8,900) and a first‑ever dip in OA share in 2023–24, partly from backlash against high‑volume OA publishers [3][8].

• Major funders (e.g., SNSF) are capping APCs and funding diamond OA platforms as a counterweight [9].

• However, forecasts through 2028 still show double‑digit OA revenue growth dominated by APC‑funded models, suggesting cost pressure will reshape pricing and eligibility more than overturn the model [2].

4. AI‑driven workflow changes (but same underlying economics)

• 2026–27 analyses anticipate:

• AI‑assisted screening, reproducibility checks, and ""reviewer agents"" to handle surging submissions [1][10].

• Structured, machine‑readable appendices, code/data sharing, and AI‑verifiable results becoming more common [1].

• Expansion of parallel ecosystems (preprint servers, AI‑focused venues) for AI‑generated/assisted work [1][10].

These changes mostly reshape peer review and integrity infrastructure, while the payment models (APC, S2O, diamond, transformative) remain in place.

5. Policy and equity dynamics

• Plan S and cOAlition S policies refuse to fund ""pure"" hybrid APCs outside transformative arrangements, but hybrid OA is still growing strongly due to those very agreements [3][4].

• Diamond OA is favored in equity and ""knowledge as a public good"" discourse (UNESCO, Plan S diamond initiatives) [9][11], yet real‑world scaling lags behind rhetoric [5].

• National OA agreements through 2027 (e.g., across Europe, b‑on) strongly favor continued hybrid+gold dominance over the next few years [6].

Evidence‑based outlook for 2027

Barring a sharp regulatory shock or sudden collapse in APC funding:

1. Dominant economic model (by volume and revenue)

• APC‑based gold and hybrid OA delivered via:

  • Transformative/read‑and‑publish agreements covering large hybrid portfolios

  • Stand‑alone gold OA journals (commercial and born‑OA)

• These will likely account for the majority of new articles and nearly all OA revenue by 2027.

1. Secondary but material models

• Subscribe‑to‑Open: Dominant within certain society/discipline niches; likely a mid‑single‑digit share of total article volume globally but prominent in math/physics and some social sciences [7].

• Diamond OA: 5–7% of total articles, with modest growth concentrated in well‑funded, community‑governed platforms and regional initiatives [5][9][11].

1. Workflow/format models cutting across economics

• AI‑integrated peer review (AI checks + human judgment),

• Mandated transparency (code, data, AI use disclosures, citation‑checking tools like Veracity),

• Machine‑readable appendices and preprint‑plus‑journal ecosystems [1][10].

These will not be revenue models in themselves, but defining features of ""dominant"" journals in terms of prestige and citation impact.

Counterarguments and risks

• APC model under pressure:

• Critics argue APCs are unsustainable and exclusionary, especially at $10k–13k levels [2][3][8].

• Funders capping APCs and backing diamond OA could gradually shift share away from APCs after 2027, but pipelines and multi‑year deals mean no abrupt collapse in just one year.

• Diamond OA scaling:

• Some policy reports call diamond OA ""the future,"" but empirical data show a long‑standing 5–6% plateau and operational bottlenecks [5].

• Expanded collective funding could change this over a decade, but not enough to dominate by 2027.

• AI disruption:

• Speculative scenarios suggest AI‑native platforms might fundamentally alter incentives. Current 2026 analyses, however, frame 2026–27 as a ""messy experimental"" period, not yet a new equilibrium [1][10].

Implications

• Authors & institutions: Plan for continued dependence on APCs and transformative deals for mainstream journals; build policies to manage APC budgets and equity (waivers, institutional funds, diamond partnerships).

• Societies & smaller publishers: Consider S2O, diamond OA, or consortial funding to avoid APC competition while meeting OA mandates.

• Funders & policymakers: Leverage APC caps and diamond funding to bend but not immediately break the APC‑driven system; push for transparency and AI‑robust integrity infrastructure.