# What the people building AI said would happen, and whether it did.

Markdown copy of https://realalpha.app/predictions/ on Real Alpha, the best minutes from long interviews with the people building AI. Cite and link the page, not this copy.

Predictions ledger

Of 24 predictions with deadlines, none has come due yet. 7 more have no deadline. The next comes due on Dec 31, 2026, from Gavin Baker on BG2 (Jun 11, 2026): Inference revenue will end this year well over $200 billion.

[Download the CSV](https://realalpha.app/predictions.csv) How it's built

Who expects what, a year, five years and a decade out, with each moment to play: [What will AI look like in the next few years?](https://realalpha.app/q/ai-predictions/)

## Coming due 24

1. [Gavin Baker](https://realalpha.app/people/gavin-baker/) Managing partner & CIO, Atreides Management BG2 · Jun 11, 2026
   [Inference revenue will end this year well over $200 billion.](https://realalpha.app/s/why-1-5t-of-ai-capex-can-still-pencil-out-against-300b-of/?t=3926)
   > “Yeah, exactly. I think we end this year well over 200 billion in inference revenue, well over.”
   [1:05:26](https://realalpha.app/s/why-1-5t-of-ai-capex-can-still-pencil-out-against-300b-of/?t=3926) They said “this year” AI business models
   Due Dec 31, 2026 in 2 months
2. [Emad Mostaque](https://realalpha.app/builders/emad-mostaque/) Founder, Intelligent Internet Moonshots · Oct 17, 2026
   [The big unsolved problems of physics, such as quantum gravity and unified field theory, will be solved by another method within half a year.](https://realalpha.app/s/gpt-6-handled-research-level-physics-problems-for-about-3/?t=1953)
   > “I don't think they'll be solved this way, actually. I think they'll be solved another way, but I don't think it'll be more than half a year.”
   [32:33](https://realalpha.app/s/gpt-6-handled-research-level-physics-problems-for-about-3/?t=1953) They said “not more than half a year” Science & math
   Due Apr 17, 2027 in 6 months
3. [Dylan Patel](https://realalpha.app/people/dylan-patel/) Founder & CEO, SemiAnalysis Big Technology Podcast · Oct 9, 2026
   [OpenAI will become profitable on revenue minus compute costs within two or three quarters.](https://realalpha.app/s/ai-revenue-may-justify-infrastructure-spending-sooner-than/?t=739)
   > “And OpenAI gets there, you know, within the next couple of quarters, maybe two or three quarters.”
   [12:19](https://realalpha.app/s/ai-revenue-may-justify-infrastructure-spending-sooner-than/?t=739) They said “within the next couple of quarters, maybe two or three quarters” AI business models
   Due Jul 9, 2027 in 9 months
4. [Bernardo Aceituno](https://realalpha.app/builders/bernardo-aceituno/) President, Stack AI MTS · Oct 17, 2026
   [Within a year, AI agents in enterprises will expand beyond web-based use to more product surfaces, such as Teams, Asana and Confluence, where they can be triggered.](https://realalpha.app/s/enterprise-agents-will-spread-across-apps-and-act/?t=1073)
   > “ai agents are going to continue expanding to product surfaces to expect like any single surface you have today to be able to interact with your core ai agents whether if it's teams whether if it's asana whether if it's you know confluence whatever you're working with being able to trigger them from there”
   [17:53](https://realalpha.app/s/enterprise-agents-will-spread-across-apps-and-act/?t=1073) They said “in a year from now” AI agents Hard to check
   Due Oct 17, 2027 in 12 months
5. [Sundar Pichai](https://realalpha.app/builders/sundar-pichai/) CEO, Google and Alphabet Sources · Oct 21, 2026
   [Waymo will expand internationally and into other weather conditions and highways over the next 12 months.](https://realalpha.app/s/waymo-is-scaling-through-partnerships-and-broader/?t=2955)
   > “And I think that is what the next 12 months is going to be, more of it, international, other weather conditions, highways, and everywhere scaling up with this high safety bar we have”
   [49:15](https://realalpha.app/s/waymo-is-scaling-through-partnerships-and-broader/?t=2955) They said “the next 12 months” Self-driving
   Due Oct 21, 2027 in 12 months
6. [Tanishq Abraham](https://realalpha.app/people/tanishq-abraham/) Co-founder & CEO, Sophont MTS · Oct 21, 2026
   [Within a year, some medical AI tools will reach retrospective clinical trials, and possibly prospective trials.](https://realalpha.app/s/cancer-ai-tools-could-reach-clinical-trials-within-a-year/?t=1306)
   > “I mean, I think what we'll start to see is some of these tools could potentially reach these clinical trials, at the very least, definitely like retrospective trials, potentially even prospective trials.”
   [21:46](https://realalpha.app/s/cancer-ai-tools-could-reach-clinical-trials-within-a-year/?t=1306) They said “a year from now” Health & medicine
   Due Oct 21, 2027 in 12 months
7. [Arvind Krishna](https://realalpha.app/builders/arvind-krishna/) Chairman & CEO, IBM All-In · Oct 21, 2026
   [Somebody will claim to have demonstrated error correction or mitigation on a running quantum computer by the end of 2027.](https://realalpha.app/s/quantum-safe-security-needs-to-start-before-quantum/?t=187)
   > “And so I'll predict you'll see somebody make claims by next year.”
   [3:07](https://realalpha.app/s/quantum-safe-security-needs-to-start-before-quantum/?t=187) They said “by next year” Chips & GPUs
   Due Dec 31, 2027 in 14 months
8. [Arvind Krishna](https://realalpha.app/builders/arvind-krishna/) Chairman & CEO, IBM All-In · Oct 21, 2026
   [One or two companies will make progress on quantum error correction, though not at scale, by 2027.](https://realalpha.app/s/quantum-safe-security-needs-to-start-before-quantum/?t=193)
   > “you'll find one or two will make progress by 2027 error correction not at scale but error correction”
   [3:13](https://realalpha.app/s/quantum-safe-security-needs-to-start-before-quantum/?t=193) They said “by 2027” Chips & GPUs
   Due Dec 31, 2027 in 14 months
9. [Dylan Patel](https://realalpha.app/people/dylan-patel/) Founder & CEO, SemiAnalysis Big Technology Podcast · Oct 9, 2026
   [U.S. CapEx, including data centers, chips and the wider supply chain, will be on the order of $2 trillion next year.](https://realalpha.app/s/ai-related-u-s-capex-could-approach-2-trillion-next-year/?t=117)
   > “So, like, I mean, if you look at, like, next year, CapEx across the U.S. will be on the order of $2 trillion, not just data centers and chips, but also all the rest of the supply chain that people are investing in.”
   [1:57](https://realalpha.app/s/ai-related-u-s-capex-could-approach-2-trillion-next-year/?t=117) They said “next year” Data centers & energy
   Due Dec 31, 2027 in 14 months
10. [Dylan Patel](https://realalpha.app/people/dylan-patel/) Founder & CEO, SemiAnalysis Big Technology Podcast · Oct 9, 2026
    [OpenAI and Anthropic together could reach $1.5 trillion in revenue by the end of 2027.](https://realalpha.app/s/ai-revenue-may-justify-infrastructure-spending-sooner-than/?t=683)
    > “I'm really in that you can get to, you know, one and a half trillion, maybe even by 2020 end of 2027 with just those two companies.”
    [11:23](https://realalpha.app/s/ai-revenue-may-justify-infrastructure-spending-sooner-than/?t=683) They said “by end of 2027” AI business models
    Due Dec 31, 2027 in 14 months
11. [Misha Laskin](https://realalpha.app/builders/misha-laskin/) Co-Founder & CEO, ReflectionAI No Priors · Oct 9, 2026
    [Frontier model training runs will cost on the order of ten billion dollars going into next year.](https://realalpha.app/s/frontier-training-costs-rise-4x-per-generation-and-ai-speeds/?t=567)
    > “Going into next year, I think order 10.”
    [9:27](https://realalpha.app/s/frontier-training-costs-rise-4x-per-generation-and-ai-speeds/?t=567) They said “next year” Scaling laws
    Due Dec 31, 2027 in 14 months
12. [Chamath Palihapitiya](https://realalpha.app/people/chamath-palihapitiya/) Founder & CEO, Social Capital All-In · Oct 2, 2026
    [Within 12 to 18 months, governments will regulate data center access and usage, allocating GPUs, inference and chips to sectors such as defense and financial services.](https://realalpha.app/s/cyber-defense-will-force-governments-to-allocate-gpus/?t=1349)
    > “What I think is going to end up happening, and this is a prediction no one may like, but I do think we're going to end up in a world where governments are going to regulate data center access and usage.”
    [22:29](https://realalpha.app/s/cyber-defense-will-force-governments-to-allocate-gpus/?t=1349) They said “12 to 18 months from now” Policy & regulation
    Due Apr 2, 2028 in 17 months
13. [Josh Gao](https://realalpha.app/people/josh-gao/) Founder & CEO, Mecka AI Training Data · Oct 20, 2026
    [Pick-and-place and packaging tasks on assembly lines that traditional robotics could not solve will be solved within the next two years.](https://realalpha.app/s/constrained-jobs-will-come-before-household-robots/?t=1392)
    > “A lot of these pick and place, these packaging tasks that you see on assembling lines that had to be manipulation bots that couldn't be solved by traditional robotics, we will get there within the next year or two.”
    [23:12](https://realalpha.app/s/constrained-jobs-will-come-before-household-robots/?t=1392) They said “within the next year or two” Robotics
    Due Oct 20, 2028 in 2 years
14. [Arvind Krishna](https://realalpha.app/builders/arvind-krishna/) Chairman & CEO, IBM All-In · Oct 21, 2026
    [Western open-weight models as capable as the Chinese ones will come out within the next year or two.](https://realalpha.app/s/open-weight-models-could-handle-most-enterprise-ai-at-lower/?t=846)
    > “But I think that over the next year or two, I'm very doubtful that there will not be great models that come up in the West that are also open weight and that are as capable as the Chinese.”
    [14:06](https://realalpha.app/s/open-weight-models-could-handle-most-enterprise-ai-at-lower/?t=846) They said “over the next year or two” Open-weight models
    Due Oct 21, 2028 in 2 years
15. [Matthew Prince](https://realalpha.app/people/matthew-prince/) TBPN · Oct 22, 2026
    [Within two years, it will be almost impossible to write code with classic bugs such as divide-by-zero, off-by-one, or unconstrained-memory errors.](https://realalpha.app/s/ai-could-make-common-coding-bugs-nearly-impossible-within/?t=1264)
    > “But I think we're going to come out of this, and two years from now, I think it's going to be almost impossible to write buggy code, at least in terms of the sort of classic divide by zero or off by one or unconstrained memory glitches.”
    [21:04](https://realalpha.app/s/ai-could-make-common-coding-bugs-nearly-impossible-within/?t=1264) They said “two years from now” Cybersecurity
    Due Oct 22, 2028 in 2 years
16. [Alexander Wissner-Gross](https://realalpha.app/people/alexander-wissner-gross/) Founder, Reified Moonshots · Oct 15, 2026
    [AI with a virtual cell will propose in silico small-molecule drug interventions that cure the top 5,000 diseases within two to three years at most.](https://realalpha.app/s/in-silico-cures-for-5-000-diseases-may-arrive-in-2-to-3/?t=3465)
    > “there are in silico proposed say small molecule drug interventions that cure the top 5,000 diseases thanks to a virtual cell I think we're going to see that in the next two to three years maximum.”
    [57:45](https://realalpha.app/s/in-silico-cures-for-5-000-diseases-may-arrive-in-2-to-3/?t=3465) They said “in the next two to three years maximum” Biology & drugs
    Due Oct 15, 2029 in 3 years
17. [Alexander Wissner-Gross](https://realalpha.app/people/alexander-wissner-gross/) Founder, Reified Moonshots · Oct 15, 2026
    [AI will solve the search problem of planning intervention paths from a diseased state to a healthy state within the next two to three years.](https://realalpha.app/s/curing-disease-is-a-path-planning-search-problem-not-a/?t=4215)
    > “How do I, you know, turn left and then turn right, take this pill, then take this pill, then climb this, you know, hill to get from A to B and the AI will solve. That's just a search problem. So AI will solve the search problem in, I think, next two to three years.”
    [1:10:15](https://realalpha.app/s/curing-disease-is-a-path-planning-search-problem-not-a/?t=4215) They said “next two to three years” Biology & drugs Hard to check
    Due Oct 15, 2029 in 3 years
18. [Chamath Palihapitiya](https://realalpha.app/people/chamath-palihapitiya/) Founder & CEO, Social Capital Moonshots · Oct 21, 2026
    [There will be a massive compute shortage in the next three or four years.](https://realalpha.app/s/power-shortages-will-constrain-ai-more-than-chips/?t=2675)
    > “we are probably going to see a massive, massive, massive compute shortage in the next three or four years.”
    [44:35](https://realalpha.app/s/power-shortages-will-constrain-ai-more-than-chips/?t=2675) They said “in the next three or four years” Data centers & energy
    Due Oct 21, 2030 in 4 years
19. [Jensen Huang](https://realalpha.app/builders/jensen-huang/) Founder & CEO, NVIDIA BG2 · Oct 15, 2026
    [By the end of the decade, annual AI infrastructure build-out will probably be between $3 trillion and $4 trillion, closer to $4 trillion.](https://realalpha.app/s/why-the-infrastructure-build-out-doesnt-need-trillion-dollar/?t=4139)
    > “I said a couple of years ago that as we exit this decade, infrastructure build-out will probably be somewhere between $3 to $4 trillion a year.”
    [1:08:59](https://realalpha.app/s/why-the-infrastructure-build-out-doesnt-need-trillion-dollar/?t=4139) They said “as we exit this decade” Data centers & energy
    Due Dec 31, 2030 in 4 years
20. [Jake Cooper](https://realalpha.app/people/jake-cooper/) Founder & CEO, Railway Knuckle Up with Nakul · Oct 13, 2026
    [Within five years there will be 100 times more people writing 100 times more software, about 10,000 times more software overall.](https://realalpha.app/s/agent-growth-could-make-compute-the-next-infrastructure/?t=3308)
    > “Well, I think you're going to have, you know, a hundred times more people writing a hundred times more software.”
    [55:08](https://realalpha.app/s/agent-growth-could-make-compute-the-next-infrastructure/?t=3308) They said “in five years” Coding agents Hard to check
    Due Oct 13, 2031 in 5 years
21. Alex Rives Head of Science, Biohub The Cognitive Revolution · Oct 20, 2026
    [First drafts of cellular atlases of health and disease will exist within two years, and a pretty comprehensive map within five years given a focused coordinated effort.](https://realalpha.app/s/a-coordinated-effort-could-map-cellular-health-and-disease/?t=2874)
    > “first drafts of that within the next one or two years, you know, within five years with a really focused coordinated effort, we can have, you know, a pretty comprehensive map there.”
    [47:54](https://realalpha.app/s/a-coordinated-effort-could-map-cellular-health-and-disease/?t=2874) They said “within the next one or two years; within five years” Biology & drugs
    Due Oct 20, 2031 in 5 years
22. [Arvind Krishna](https://realalpha.app/builders/arvind-krishna/) Chairman & CEO, IBM All-In · Oct 21, 2026
    [Enterprises will consolidate from hundreds of software vendors to five to ten main platforms within six years.](https://realalpha.app/s/ai-could-consolidate-hundreds-of-enterprise-software-vendors/?t=2325)
    > “I believe that's going to play out whether it takes six months or six years.”
    [38:45](https://realalpha.app/s/ai-could-consolidate-hundreds-of-enterprise-software-vendors/?t=2325) They said “whether it takes six months or six years” Enterprise AI Hard to check
    Due Oct 21, 2032 in 6 years
23. [Alan Chan](https://realalpha.app/people/alan-chan/) Research Fellow, GovAI The Cognitive Revolution · Oct 14, 2026
    [An AI takeover, via controlling government or physically overpowering militaries, will not happen within the next five to ten years.](https://realalpha.app/s/why-an-ai-takeover-is-much-harder-than-the-rogue-agent-panic/?t=4316)
    > “Both those pathways, I think, are quite implausible to me in the next, you know, like five year, 10 year horizon or something.”
    [1:11:56](https://realalpha.app/s/why-an-ai-takeover-is-much-harder-than-the-rogue-agent-panic/?t=4316) They said “in the next five year, 10 year horizon” Safety & alignment
    Due Oct 14, 2036 in 10 years
24. [Bill Thompson](https://realalpha.app/people/bill-thompson/) Founder & CEO, Spartan Forge The Joe Rogan Experience · Oct 21, 2026
    [Within ten years, AI will leave no use for people in a lot of the things people do now, eliminating many jobs.](https://realalpha.app/s/ai-could-make-local-production-and-human-made-work-more/?t=9497)
    > “I think there's going to be no use for people for a lot of things that people do right now.”
    [2:38:17](https://realalpha.app/s/ai-could-make-local-production-and-human-made-work-more/?t=9497) They said “Ten years maximum” Jobs & the economy Hard to check
    Due Oct 21, 2036 in 10 years

## No deadline 7

1. [Pierre Stock](https://realalpha.app/builders/pierre-stock/) VP of Science, Mistral AI The MAD Podcast · Oct 22, 2026
   [Mistral's next model, ML5, will be able to do meaningful work for a few days, up from a few hours now.](https://realalpha.app/s/mistral-says-its-models-speed-senior-researchers-up-two-to/?t=4523)
   > “Tomorrow with ML5 is going to be a few days, for sure, of meaningful work.”
   [1:15:23](https://realalpha.app/s/mistral-says-its-models-speed-senior-researchers-up-two-to/?t=4523) AI agents
2. [Tiago Sada](https://realalpha.app/builders/tiago-sada/) CPO, World MTS · Oct 20, 2026
   [Bots will make up 99.999% of online traffic very soon, rather than humans.](https://realalpha.app/s/real-time-deepfakes-can-cause-lasting-financial-damage/?t=277)
   > “I expect that to be 99.999% very soon will actually be bots, not humans.”
   [4:37](https://realalpha.app/s/real-time-deepfakes-can-cause-lasting-financial-damage/?t=277) They said “very soon” AI agents
3. [Scott Aaronson](https://realalpha.app/people/scott-aaronson/) Professor of computer science, UT Austin The Peterman Pod · Oct 12, 2026
   [Very soon, OpenAI's backlog of AI solutions to significant open problems will be published, hundreds of them, mostly unchecked by any human.](https://realalpha.app/s/openais-15m-navier-stokes-proof-and-how-credit-will-work/?t=1060)
   > “but but my my expectation is that at some point you know very soon in fact you know all these things are going to be published all right like you know we're going to start seeing hundreds of solutions to significant open problems that in most cases no human has gone through”
   [17:40](https://realalpha.app/s/openais-15m-navier-stokes-proof-and-how-credit-will-work/?t=1060) They said “very soon” Science & math
4. [Dylan Patel](https://realalpha.app/people/dylan-patel/) Founder & CEO, SemiAnalysis Big Technology Podcast · Oct 9, 2026
   [Anthropic will likely release a new model in the coming weeks, ahead of its IPO.](https://realalpha.app/s/frontier-labs-are-holding-back-models-before-public-release/?t=3425)
   > “So I think, you know, Anthropic will likely release a new model in the coming weeks, you know, for their IPO.”
   [57:05](https://realalpha.app/s/frontier-labs-are-holding-back-models-before-public-release/?t=3425) They said “in the coming weeks” Frontier models
5. [Dylan Patel](https://realalpha.app/people/dylan-patel/) Founder & CEO, SemiAnalysis Big Technology Podcast · Oct 9, 2026
   [OpenAI will likely release a new model in the coming weeks.](https://realalpha.app/s/frontier-labs-are-holding-back-models-before-public-release/?t=3431)
   > “I think OpenAI will likely have to release a new model in the coming weeks as well, just because it's such a competitive environment.”
   [57:11](https://realalpha.app/s/frontier-labs-are-holding-back-models-before-public-release/?t=3431) They said “in the coming weeks” Frontier models
6. [Andy Pavlo](https://realalpha.app/people/andy-pavlo/) VP of Database Research, ClickHouse The MAD Podcast · Oct 8, 2026
   [Agents will potentially drive a 10 to 100x increase in the volume of queries hitting databases.](https://realalpha.app/s/agents-could-mean-10-to-100x-more-database-queries/?t=831)
   > “So I think there's going to be potentially a 10 to 100x increase in volume and the number of queries that are going to hit up against databases.”
   [13:51](https://realalpha.app/s/agents-could-mean-10-to-100x-more-database-queries/?t=831) Data & databases
7. [Asha Sharma](https://realalpha.app/people/asha-sharma/) Lenny's Podcast · Aug 28, 2025
   [Spending on post-training will match spending on pre-training and later exceed it.](https://realalpha.app/s/post-training-could-draw-more-investment-than-pre-training/?t=2701)
   > “I believe we will see, you know, just as much money spent on post-training as we will on pre-training and in the future more on post-training.”
   [45:01](https://realalpha.app/s/post-training-could-draw-more-investment-than-pre-training/?t=2701) They said “in the future” Reasoning & RL

## How it's built

Real Alpha cuts the best minutes from long interviews with the people building AI, and every one of those moments is read for predictions: 31 so far, from 23 episodes. A model (GPT-6 Luna) reads each moment's transcript and lists every forecast that says what will happen by when, or names an outcome that plainly will or won't. A second model (Claude Sonnet) keeps only the ones said as the speaker's own view, by someone it can name, and drops directions with nothing to measure. Code checks every quote word for word against the transcript and links the second it starts. We read all 24 of the first ones against their transcripts by hand on Oct 22, 2026: 23 were right in every field, and we put the other right. A time frame becomes a deadline counted from the day the episode came out: a year means its last day, “in 18 months” counts from the episode, and when the words allow two dates we take the later one, so no one is judged early. Once a deadline passes, a model searches the web for what happened and proposes a status, and a person approves it before it changes here, with the source linked. New interviews add their predictions within the hour of being cut. The moments themselves, grouped by how far out they look, are on [AI predictions for the next few years](https://realalpha.app/q/ai-predictions/).

The ledger is free to reuse under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/): download [the CSV](https://realalpha.app/predictions.csv) (31 rows, as of Oct 23, 2026) and credit “Real Alpha predictions ledger (realalpha.app/predictions)” with a link to [this page](https://realalpha.app/predictions/). Each row's moment_url plays the second it was said.

Spotted a wrong date, quote or status? Write to [hello@realalpha.tv](mailto:hello@realalpha.tv) and we'll fix it. Last updated Oct 23, 2026.