{"slug": "q2-5-2026-timelines-update-uplift-and-revenue", "title": "Q2.5 2026 Timelines Update: Uplift and Revenue", "summary": "AI Futures, the research group behind the AI Futures Model, updated its AI timelines forecasts, slightly shortening them and expressing increased confidence due to improved modeling and evidence. The group now incorporates coding uplift and revenue as alternative methods to anchor the arrival of Automated Coder (AC), alongside the original METR time-horizon approach, with all three methods predicting similar AC arrival dates. The latest all-things-considered forecasts are available at aifuturesmodel.com, and the group also re-evaluated AI 2027's predictions, finding reality is progressing at about 70-90% of the pace AI 2027 predicted.", "body_md": "*Tl;dr: Our timelines haven’t changed much (they got slightly shorter) but our modeling and evidence base have noticeably improved, so we feel somewhat more confident.*\n\nWe intend to regularly update our AI timelines forecasts as new evidence comes in and new analyses are done. Today’s “Q2” update was delayed by the crunch to publish [AI 2040: Plan A](https://ai-2040.com/), our [domestic regulation blog post](https://blog.aifutures.org/p/how-to-pace-the-us-frontier), and the time needed to implement and document changes to our model.\n\nThe original [AI Futures Model](https://www.aifuturesmodel.com) predicted when Automated Coder (AC), an AI for which the leading AI company would rather fire its human software engineers than forego AI usage for coding, would happen using [METR’s measurements of coding time horizon](https://metr.org/time-horizons/).** **(More precisely, time horizon anchors are used to set the [effective compute](https://www.aifuturesmodel.com/#section-modelingeffectivecompute) required for AC.) While serviceable, this method has huge weaknesses, including (a) it’s unclear what time horizon corresponds to AC (it’s even unclear whether any finite value would) (b) people strongly disagree about the extent to which we should expect the time horizon trend to be superexponential as a function of effective compute, in a way that can lead to vastly different predictions.\n\nSo we’ve been on the lookout for other methods for setting the effective compute required for AC, and now we have two candidates: coding uplift (i.e., how much of a speedup AIs are providing to software engineers at AGI companies) and [revenue](#Revenue). Coding uplift is our favorite method: the basic idea is to estimate the doubling time of the quantity (uplift - 1) and extrapolate that until AC-level uplift is reached. The [way we anchor the model](#Uplift) using uplift-relevant estimates is a bit complex, so we first explain a simpler [3-parameter uplift model](#A_3_parameter_uplift_model_for_predicting_when_Automated_Coder_will_arrive) that gives similar results. With coding uplift, the value corresponding to AC is much less uncertain than with time horizons, and while we also expect coding uplift to be somewhat superexponential this isn’t nearly as important as with time horizons.\n\nSurprisingly, [all 3 methods predict quite similar AC arrival dates](#Daniel). [1] We take this to be a somewhat encouraging sign about the robustness of our forecasts, though we are still very uncertain.\n\nYou can explore the uplift-anchored version of the model at [aifuturesmodel.com](http://aifuturesmodel.com) and the other options for AC forecasting via the dropdown at the top of the second graph.\n\nEach author assigned weights to the 3 AC anchoring methods which the model aggregates into an overall forecast. We also [re-evaluated AI 2027’s predictions](#Update_to_the_grading_of_AI_2027_s_predictions); reality seems to be going about 70-90% as fast as AI 2027 predicted.\n\nIncorporating all of the above, our latest all-things-considered timelines forecasts: ([link](https://www.aifuturesmodel.com/forecast/daniel-08-16-26?takeoff=ASI%2CTED-AI&cmode=forecaster&csim=eli-08-16-26%2Cbrendan-08-16-26&ctype=atc))\n\nHere’s how our forecasts have shifted recently:\n\n[https://www.datawrapper.de/_/4aHow/](https://www.datawrapper.de/_/4aHow/)\n\nWe describe in the [appendix](#Appendix):\n\nA simple method to predict the AC arrival date is to assume that (coding uplift - 1) grows exponentially. We think that this is the best simple model for predicting when AC will arrive.\n\nSpecifically, this model takes as input 3 parameters, for which we list Daniel’s median estimates:\n\n**Present day coding uplift, i.e. the speedup factor due to AI assistance: 2x. **Daniel thinks 1.04x is something like a lower bound given the [METR study](https://metr.org/blog/2026-02-24-uplift-update) which found 1.04x - 1.2x uplift, with METR thinking that these numbers were biased downward due to selection effects (people were less likely to participate in the study if they thought AI would be useful to them.) Otherwise, he’s integrating various sources of evidence including the [Apr 2026 Anthropic internal survey](https://www-cdn.anthropic.com/08ab9158070959f88f296514c21b7facce6f52bc.pdf#page=44) having a geometric mean of 4x, and a private estimate of 1.7x AI R&D labor uplift by Ryan Greenblatt (which was an estimate for AI R&D labor as a whole, so presumably coding-only would be higher).\n\n**Present day doubling time of the quantity (uplift-1): 5 months. **According to Anthropic employee surveys, coding uplift has gone from 1.25x to 4x in 7 months. [2] This would be a ~2 month uplift doubling time, but correcting for Mythos being above the long-term Anthropic ECI trend gives us a ~3.5 month doubling time.\n\nNow, probably their employees are biased towards overestimating coding uplift. But unless the bias has been significantly increasing over time, that’s still more than three doublings of uplift-1 in less than a year – a 3.6 month doubling time!** **Daniel’s median is longer, 5 months, because he’s partly deferring to the opinions of other researchers he respects (Eli and Ryan) whose subjective sense is that the doubling time is longer.\n\n**Uplift corresponding to AC: 20x. **The full AI Futures Model says 32x in the median case, but we expect the true uplift to be a little lower because the model doesn’t account that AIs can be used to accomplish coding tasks less efficiently than humans.\n\nComparing Daniel’s median estimates with Eli and Brendan’s:\n\n[https://www.datawrapper.de/_/ts1bS/](https://www.datawrapper.de/_/ts1bS/)\n\nThis simple model extrapolates the uplift trend (assuming the doubling time stays constant, i.e. the trend is exponential) [4] and sees when it reaches the uplift corresponding to AC.\n\nWhat does this method say? **See ****ac-arrival.vercel.app**** for a vibe-coded app in which you can play around with the simple extrapolation.**\n\nSee [below](#Explicitly_simulating_the_training_run_of_the_leading_AI_model) for a more complicated version that uses present day uplift and uplift doubling times as anchors for setting the behavior of the full AI Futures Model. Factors that are accounted for in the full model are:\n\nThe full model doesn’t have uplift at AC set as a parameter, instead it is inferred from model behavior.\n\nWe’ll now discuss how uplift and revenue estimates can be used to estimate the effective compute required for AC by anchoring the AI Futures Model.\n\nOur overall forecast is made by using each method separately and then aggregating the results via a weighted mixture. You can explore the uplift-anchored version of the model at [aifuturesmodel.com](http://aifuturesmodel.com) and the revenue (and time horizon) option via the dropdown at the top of the second graph.\n\nWe give the following weights:\n\n[https://www.datawrapper.de/_/1ycZM/](https://www.datawrapper.de/_/1ycZM/)\n\nWe give the most weight to uplift because (a) the value corresponding to AC is more clear than for revenue or time horizon and (b) the trajectory of (uplift - 1) seems closer to exponential in log(effective compute) than for time horizon. The main advantage of time horizon relative to uplift is that it’s more measurable, and the main advantage relative to revenue is that it’s a more direct measurement of coding capabilities.\n\nOur model already issues predictions about coding uplift, so we aren’t fitting an entirely new function and AC requirement like for the other two methods.\n\nThere is a module in the AI Futures Model (AIFM) which aggregates human labor and AI agents to produce an estimated “aggregate coding labor” (and therefore an estimated coding uplift) at each capability level.** **This module is generally calibrated by three degrees of freedom:\n\nIn time horizon and revenue mode, we use one of those trends to choose the capability level pinning down the third degree of freedom. In uplift mode, we don’t directly specify the capability level corresponding to AC, and instead we constrain the remaining degree of freedom by specifying the rate at which coding uplift is increasing today (specifically, the doubling time of uplift - 1).** **With the automation module calibrated, we can then read off the capability level corresponding to the AC definition, and therefore the AC date.\n\nWe fit a function from AI capabilities (operationalized as effective compute or [ECI](https://epoch.ai/benchmarks?view=graph&tab=eci)) to leading AI company annualized revenue (specifically, the leading AI model developer’s revenue; so not including Nvidia). In particular, we fit an exponential function from ECI to annualized revenue (equivalent in our model to an exponential function from log(effective compute) to annualized revenue). We extrapolate the function into the future, and make guesses about which level of AI company revenue would correspond to having just achieved the AC milestone.\n\nWe estimate the following median parameters:\n\n[https://www.datawrapper.de/_/N9NuP/](https://www.datawrapper.de/_/N9NuP/)\n\nModeling annualized revenue as an exponential function of ECI is a bit more sophisticated than modeling it as a function of time; it allows us to incorporate effects like a slowdown in datacenter growth or a feedback loop from AI R&D automation.** **Empirically, revenue has grown by 10x for every 15 ECI points so far. However, this method doesn’t take into account various other drivers of revenue growth besides capabilities (such as % of total compute allocated to inference and inference margins). It also doesn’t take into account that even holding those factors constant, revenue might not be an exponential function of ECI.\n\nWe try to intuitively take these factors into account by our choice of parameter values — even though Anthropic’s annualized revenue has grown 10x/yr for several years, we think it’ll slow down soon, and use 5-7x/yr as our median current growth rate.\n\nWe’ve updated our assessment of how the pace of AI progress has compared to AI 2027. Depending on what metrics you include and what aggregation method you use, reality seems to be going at roughly 70-90% the speed of AI 2027. That’s the quantitative assessment. The qualitative assessment will be discussed in the next section.\n\n[https://www.datawrapper.de/_/H4G09/](https://www.datawrapper.de/_/H4G09/)\n\nIf progress were to continue at 75% of the pace of AI 2027, Automated Coder would be reached in mid-2027.[[5]](#fnb48opeb6fo)\n\nPart of the reason that the relative uplift pace of progress is so much lower than the others is that since publication, we’ve revised our estimates downward for what AI software R&D uplift was at the *beginning* of AI 2027*. *This is reflecting a real way that we estimate reality is behind schedule, but it makes the “pace” of progress framing not as natural as the others.\n\nAs for the public salience metric, which is our biggest predictive error, we wonder if we should have picked a better operationalization. AI does seem much more salient today than it was a year ago, even if that particular survey isn’t showing any progress.\n\nA few minor methodological changes we’ve made since [our previous evaluation](https://blog.aifutures.org/p/grading-ai-2027s-2025-predictions):\n\nDetails about the estimates can be found in [this spreadsheet](https://docs.google.com/spreadsheets/d/1JfEJRIwGe7fSCIW14eC7q0sEhVWHEdetykVh9g9XnF8/edit?gid=1928509176#gid=1928509176).\n\nWhereas the early 2026 section in AI 2027 was very on-point (it was titled “Coding Automation”) the mid-2026 section seems more of a miss:\n\nWe aren’t China experts, but we probably would have heard by now if the CCP had consolidated the various Chinese AI projects and heavily prioritized acquiring compute. In general it seems that “China Wakes Up” has not yet happened. That said, we expect there has been *some *degree of AGI wakeup in China, as there has been across the world.\n\nOther notes:\n\nI’m struck by the fact that all three AC extrapolation methods gave basically the same answer, independently: ([link](https://aifuturesmodel.com/forecast/daniel-08-16-26?takeoff=ASI%2CTED-AI&breakout=1&show=model))\n\nI didn’t do the math in my head, I just made guesses about the parameters and then we calculated the results. I think this is some reason to be somewhat more confident in these predictions.\n\nAnother source of evidence I’d like to incorporate is the AI 2027 grading / tracking. In a nutshell, the methodology is:\n\nIt still seems like things are roughly on track for AI 2027, just going a bit slower. How much slower? About 75% speed, as mentioned [above](#Comparing_the_AI_2027_pace_of_progress_to_reality). This would predict AC happening in mid-2027. If we think it’s more like 60% speed, then that would predict early 2028. Again, interesting convergence with the other three methods.\n\nAre there any other major factors to consider, in forming my all-things-considered views? Well there are [many other things to say](https://blog.aifutures.org/p/ai-futures-model-dec-2025-update?open=false#%C2%A7why-our-approach-to-modeling-comparing-to-other-approaches), but overall I’m pretty happy with what I’ve said so far as a summary of the most important points. I’m not aware of any other arguments or considerations strong enough to push me significantly away from the above. So I’ll just go with what the model says for AC, except slightly more confident since all 3 anchoring methods give similar results, and one month sooner to incorporate the 75% AI 2027 speed method. ([link](https://aifuturesmodel.com/forecast/daniel-08-16-26?takeoff=ASI%2CTED-AI))\n\nHere’s how my forecast has changed since April: ([link](https://aifuturesmodel.com/forecast/daniel-08-16-26?cmode=forecaster&csim=daniel-04-02-26&ctype=atc))\n\nI increase the speed of post-AC takeoff for reasons [previously described](https://blog.aifutures.org/i/182911449/daniel). ([link](https://aifuturesmodel.com/forecast/daniel-08-16-26?takeoff=ASI%2CTED-AI))\n\nMy forecasts for the arrival date of AC, TED-AI, and ASI: ([link](https://aifuturesmodel.com/forecast/daniel-08-16-26?timeline=AC%2CTED-AI%2CASI&show=atc))\n\nThe top adjustments I apply to get my AC timelines are:\n\nMy all-things-considered adjustment: ([link](https://aifuturesmodel.com/forecast/eli-08-16-26))\n\nAnd a comparison vs. April: ([link](https://aifuturesmodel.com/forecast/eli-08-16-26?cmode=forecaster&csim=eli-04-02-26&ctype=atc))\n\nCompared to the model’s takeoff predictions, I speed mine up, primarily to take into account automation of hardware R&D, hardware production, and general economic automation: ([link](https://aifuturesmodel.com/forecast/eli-08-16-26?takeoff=ASI%2CTED-AI))\n\nMy forecasts for the arrival date of AC, TED-AI, and ASI: ([link](https://aifuturesmodel.com/forecasteli-08-16-26?timeline=AC%2CTED-AI%2CASI&show=atc))\n\nThis is my first set of parameters and all-things-considered views.\n\nVarious factors the model isn’t considering for timelines to AC:\n\nOverall, the model’s 70% on AC by Jan 2030 and nearly 90% by Jan 2035 feels too confident, so I reduce these to 60% and 80%.\n\nHere is my all-things-considered adjustment: ([link](https://aifuturesmodel.com/forecast/brendan-08-16-26))\n\nFor takeoff from AC to TED-AI:\n\nTo combine these changes to the model’s takeoff with my changes to the model’s timelines to AC, I also reweighted all the rollouts according to my all things considered distribution for AC. This yields the following distribution for TED-AI arrival: ([link](https://aifuturesmodel.com/forecast/brendan-08-16-26?timeline=TED-AI))\n\nBelow, we include plots that extend our analysis of [how our views have changed since publishing AI 2027](https://blog.aifutures.org/p/clarifying-how-our-ai-timelines-forecasts). When we refer to AGI in the below plots, we mean Top-Expert-Dominating AI: an AI that is at least as good as top human experts at virtually all cognitive tasks.\n\nZooming in on the changes since 2024:\n\nThis was originally motivated by our research for Plan A; see [Plan A Takeoff Forecast](https://ai-2040.com/supplements/takeoff-forecast)\n\nWe think this improves our takeoff speed estimates. But also, it allows us to predict the effect of various policies to [pace the frontier](https://blog.aifutures.org/p/how-to-pace-the-us-frontier) that involve reducing compute available for AI development.\n\nIncorporating this change leads to a somewhat slower takeoff; see the charts below for how it affected model predictions given each of Daniel and Eli’s parameter estimates.\n\nBased on a forthcoming research taste evaluation from P-Zero Research which updated us toward faster research taste progress, Eli adjusted his:\n\nBrendan’s median estimates of 2.69 and 4.35 were influenced by this research as well. Daniel didn’t update his estimates as the research taste slope estimate of 3 was already higher than Brendan and Eli’s, and his median to top taste multiplier estimate of 4 was very similar.\n\nWe have previously not been very clear on whether we’re forecasting when AI milestones will actually appear in the world, or whether we are assuming things go as fast as is technically feasible; e.g., assuming that there isn’t government intervention to slow down AI. In our supplementary materials, we implied that we were adjusting for non-technical slowdowns. But in practice, we hadn’t thought much about this and some of us were explicitly assuming the opposite.\n\n**We’ve discussed this issue, and we’ve decided that from now on our forecasts are for what will happen conditional on things going as fast as is technically feasible. **We think this is more informative to forecast than to attempt to account for the likelihood of various levels of slowdown. We’ve edited our supplementary materials to reflect this.\n\nWe had formerly been making use of an approximation that the rate of progress at the present day matches what it would have been in a counterfactual “human-only” trajectory, which is easier to simulate. But that assumption is increasingly false, since AIs are (in our estimation) starting to non-negligibly speed things up. In particular, this assumption would have caused us to underestimate the rate of effective compute growth that underlies today's observed progress rates on things like time horizon. We’d then be plugging in our actual (with-automation) model trajectory to that calibrated relationship, resulting in an incorrectly sped-up prediction that would underestimate time required to e.g. reach automated coder.\n\nWe reconfigured the front page a bit, adding a few extra metrics. We switched to showing Epoch Capabilities Index by default instead of “effective compute”.\n\n(Why privilege ECI like this, as opposed to time horizon or any other metric? Technically, effective compute *requires* an underlying capability metric to define, since you need to measure software efficiency in terms of training compute required to reach “equal capability level” (which requires a metric). Also, the idea of “training compute” as a single scalar that determines capability seems increasingly fraught).\n\nWe model ECI as the unique affine transform of log(effective compute) satisfying the properties that:\n\nIt’s possible we were subconsciously biased to confirm our existing views when choosing parameter estimates, but we did our best not to look at results when doing so. An exception is that Eli looked at the results of the simplified uplift model before setting his uplift parameters.\n\nSonnet 4.5 (Sep 25): median 1.25x (selected for top 30 Claude Code usage); Mythos (Apr 7 26, Feb 24 internal deployment): 4x geomean\n\nSince we’re using these numbers only to calibrate a relationship between uplift minus 1 and general capabilities, we obtain 3.5 months by reading off the release date of Mythos Preview as though it had been on the long term AECI trend rather than its actual release date.\n\nSteady exponential growth is a reasonable default assumption for many metrics in AI and roughly matches our past estimates. Our more sophisticated modeling suggests that progress will look exponential for some time until the trend goes superexponential as we approach AC.\n\n(2027-2025.25)*(1/.75)+2025.25\n\nOne example: until recently, we weren’t accounting for the time required to retrain models with new algorithms during takeoff. There might be similar things we haven’t thought of yet.", "url": "https://wpnews.pro/news/q2-5-2026-timelines-update-uplift-and-revenue", "canonical_source": "https://www.lesswrong.com/posts/ZPSsmRH5oMwLPXys4/q2-5-2026-timelines-update-uplift-and-revenue", "published_at": "2026-08-16 19:00:25+00:00", "updated_at": "2026-08-16 19:11:38.470219+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-research", "ai-policy"], "entities": ["AI Futures", "AI Futures Model", "METR", "Automated Coder", "AI 2040", "Daniel"], "alternates": {"html": "https://wpnews.pro/news/q2-5-2026-timelines-update-uplift-and-revenue", "markdown": "https://wpnews.pro/news/q2-5-2026-timelines-update-uplift-and-revenue.md", "text": "https://wpnews.pro/news/q2-5-2026-timelines-update-uplift-and-revenue.txt", "jsonld": "https://wpnews.pro/news/q2-5-2026-timelines-update-uplift-and-revenue.jsonld"}}