For the most part, the gaming industry runs on a yearly cyclical cadence (there are large-scale decades-long trends and patterns, such as Joost Van Dreunen’s Play Pendulum, but the yearly cycles are how the industry self-organizes). So, for example, if it’s March, it’s time for GDC (and welcome once again, my friends, to the Game Revenue Optimization Mini-Summit). And if it’s December, it’s time for the pundits to gather round and tell us what will happen in the coming year.
Right now we’re in the prelude before the prediction bonanza, the calm before the storm, that long moment when the baby has been dropped on the floor but has not yet started screaming. It’s the time of the year when the people who made the predictions ‘fess up and talk about why they were right (and how it was reality that got it badly wrong).
That’s what this is. We made some predictions last year, and we’ve been thinking hard about what’s coming next year. But we owe it to you, dear reader, to let you know how we did and to let you draw your own conclusions about how seriously to take our next set of predictions.
How We Evaluated the Predictions
We evaluated our predictions along two distinct axes:
- Correctness. Were we right? Did our predictions come true? If what we said wasn’t true, you’d have good reason to ignore our upcoming set of predictions.
- Completeness. Did we miss anything important? Part of the value of predicting is not just being right, it’s covering all the important events. If we were running the country and we predicted an increase in street traffic but completely missed an attack by an army of CHUD (Cannibalistic Humanoid Underground Dwellers), you’d probably wonder whether we were the right leaders.
(as a side-note, we apologize to the fans of the undergraduate logic curriculum, who are no doubt saddened that we omitted compactness as a third evaluation criteria)
Separately, there’s also the question of who evaluates the predictions. The pattern in years past, and in most of the gaming industry press, is for the predictors to self-evaluate. That is, the people who make the predictions mostly get to decide whether they did a good job.
But we’ve been told that AI changes everything. Since 2025 is the year of agentic AI, we decided to have ChatGPT (5.1, pro) evaluate our predictions using three distinct personas.
Our expert panel was comprised of generative AI simulations of:
- A CEO of a small to midsize gaming company that is struggling to stay afloat during industry hard times.
- An industry pundit with deep knowledge of the space and a regular platform (e.g. blog or substack).
- An external observer, not working in the games industry but with some knowledge of the space and working in an adjacent industry, who feels somewhat skeptical about the value of year-end predictions.
Each of these personas was given the task of scoring us on both correctness and completeness.
We Were Mostly Correct
After a hearty breakfast, we convened the panel for the initial round of deliberations. Here’s the short version: We made 7 predictions. 6 came true; one came close.
(Taken During the Morning Session)
Here’s the detailed table:
| Prediction | CEO Evaluation | Pundit Evaluation | Skeptical Observer Evaluation | Our Reaction |
|---|---|---|---|---|
| There will be more revenue. | This prediction has basically come true in headline terms … However, that does not mean 2025 feels like a boom. | Accurate, but unexciting. | Global games revenue is higher in 2025 than in 2024, but 3 to 4 percent growth against similar levels of inflation yields only a small real gain. | Yes! We nailed it! |
| Mobile UA teams will keep leaning on MMM, often without proper validation. | Directionally right. | Mostly correct, but a bit pessimistic. | Agree with the prediction’s spirit that many teams cling to MMM without sufficient experimentation | Nailed it again! |
| No new Top 100 mobile games without web shops will add one in 2025. | Whether or not literally zero new Top 100 titles launched web shops in 2025 is hard to verify from public information and irrelevant. | Too strongly worded but directionally plausible. | The more meaningful observation is that web shops have rapidly become yet another layer of complexity in mobile monetization. | We weren’t wrong! |
| Web storefronts will see big gains from experimentation and personalization. | This prediction rings true in spirit. The 2025 ecosystem of DTC tooling, analytics integrations, and vendor case studies all pushes toward more experimentation and segmentation on web storefronts. | Partially correct but slightly overstated. | The prediction sounds plausible but I question whether, across the whole of mobile gaming in 2025, this really counts as one of the defining revenue growth engines. | Still not wrong! We’re 4 for 4 so far! |
| AR and VR will finally gain significant traction. | AR and VR do have more momentum in 2025 than a few years ago, but the impact on my business is still limited. | Broadly right. The data clearly show a meaningful upswing in AR or VR and smart glasses shipments in 2025, with AI enabled glasses emerging as a genuinely new category. | The prediction captures a real uptick in momentum but exaggerates how decisive 2025 is as a turning point. | Yes! Right again! |
| Alternative app stores will exceed 10 percent of western mobile gaming installs. | Overly optimistic about adoption speed. In 2025, alternative stores are a real strategic consideration, especially for EU focused titles and for Android in markets where OEM or carrier stores matter. | Almost certainly incorrect given the data available in late 2025. | A classic case of over extrapolating from regulatory headlines. Changing default distribution behavior for hundreds of millions of mainstream users is extremely difficult. | Ouch. These judges are tough. |
| Selling physical goods in game will stop being surprising. | Largely accurate but unevenly distributed across the industry and irrelevant to me. | Mostly correct given how 2025 has unfolded. | Agree that 2025 made the idea of buying physical goods inside games feel more normal. | Yes! Back on track and 6 for 7 overall! |
But We Missed Three Big Trends
After lunch, we reconvened the expert panel to discuss the harder question: What did we miss?
(Mainly About Where to Go for Dinner)
According to our experts, we missed three major trends.
| Prediction We Should Have Made | CEO Opinion | Pundit Opinion | Skeptical Observer Opinion | In Our Defense |
|---|---|---|---|---|
| AI-native game businesses. By late 2025, most commercially serious studios will treat AI as a default part of production and live ops, and the hard problems will be ROI measurement, workflow integration, and governance, not ‘should we use AI?’ | My P&L reality in 2025 is that AI is the only lever big enough to offset rising costs and shrinking margins. | 2025’s biggest structural shift is that AI is becoming the new production function. The bottleneck in games used to be content; now it’s taste and data quality. | AI is obviously big, but the ROI is opaque at this point. | How did we miss this? AI is huge and transformative and we’re doing a lot of work helping companies with their AI transformations. This omission is inexcusable. Especially since 30% of the respondents to a recent GDC survey said that they believe that generative AI is having a negative impact on the games industry. |
| Platformized Creator Economies. By end of 2025, creator platforms like Roblox and Fortnite will represent a third major commercial pillar alongside traditional PC/console/mobile – with built‑in A/B testing, regional pricing, and engagement-based payouts that make them some of the most data‑instrumented game economies on earth. | In 2025, a real strategic choice is: do we become a ‘studio on a platform’ (Roblox, UEFN) instead? This is new and a significant challenge. | This is the blind spot: you treated platforms mainly as distribution and webshops, not as competing economic systems. | Maybe. But are these ecosystems net-new value, or just reshuffling time and money away from other games? | What can we say? After VR and then Web3, we are skeptical of platform shifts and overlooked this one. |
| Hybrid Monetization is the New Norm. In 2025, the median successful game will be running at least two monetization models (e.g. ads + IAP, or IAP + sub), and platform-level subscriptions will keep pulling value out of pure à‑la‑carte spending. The hard problem moves from ‘which model?’ to ‘how do we optimize LTV across overlapping ones?’” | Your predictions focused on ‘more revenue’ and DTC mechanics but didn’t explicitly call out how messy monetization design has become. | Hybrid monetization and subscription stacking are now the design constraint for game businesses. | I don’t like it. This is starting to look like game design is just financial engineering. Games are not spreadsheets. | This is business as usual and not “prediction-worthy.” Hybrid monetization was well-established heading into 2025 and isn’t really a trend. And, anyway, Tiffany Keller covered it extensively at our Game Revenue Optimization Mini-Summit in 2025. |
In Conclusion
We were reasonably accurate — for the most part, the things we predicted did happen. At the same time, the events we didn’t predict are significant omissions. While our 2025 predictions captured the headline moves in revenue, UA, and monetization, our AI panel rightly called out the deeper structural shifts around AI‑native production, platformized creator economies, and hybrid monetization. It’s reasonable to say that we predicted the linear trends (that we forecasted correctly based on existing trends), but failed to anticipate the bigger non-linear shifts that will cause significant structural changes in the gaming industry.
As we head into the next prediction cycle, we’ll keep treating predictions as hypotheses to be tested, not pronouncements from on high—and we’ll try to be more explicit about the big picture changes and large-scale changes we missed this year.
