All these predictions are taken from Forbes/Rob Toews' "10 AI Predictions For 2025".
For the 2024 predictions you can find them here, and their resolution here.
You can find all the markets under the tag [2025 Forbes AI predictions].
Note that I will resolve to whatever Forbes/Rob Toews say in their resolution article for 2025's predictions, even if I or others disagree with his decision.
I might bet in this market, as I have no power over the resolution.
Description of this prediction from the article:
In 2023, the critical physical resource that bottlenecked AI growth was GPU chips. In 2024, it has become power and data centers.
Few storylines have gotten more play in 2024 than AI’s enormous and fast-growing energy needs amid the rush to build more AI data centers. After remaining flat for decades, global power demand from data centers is projected to double between 2023 and 2026 thanks to the AI boom. In the U.S., data centers are projected to consume close to 10% of all power by 2030, up from just 3% in 2022.
The demand for energy to power AI data centers is skyrocketing. Our energy systems are not prepared.
Image source: Semianalysis
Today’s energy system is simply not equipped to handle the tremendous surge in demand coming from artificial intelligence workloads. A historic collision between these two multi-trillion-dollar systems—our energy grid and our computing infrastructure—is looming.
Nuclear power has gained momentum this year as a possible solution to this Gordian knot. Nuclear represents an ideal energy source for AI in many ways: it is zero-carbon, available 24/7 and effectively inexhaustible. But realistically, new nuclear energy sources won’t be able to make a dent in this problem until the 2030s, given long research, project development and regulatory timelines. This goes for traditional nuclear fission power plants, for next-generation “small modular reactors” (SMRs) and certainly for nuclear fusion power plants.
Next year, an unconventional new idea to tackle this challenge will emerge and attract real resources: putting AI data centers in space.
AI data centers in space—at first blush, this sounds like a bad joke about a VC trying to combine too many startup buzzwords. But there may in fact be something here.
The biggest bottleneck to rapidly building more data centers on Earth is accessing the requisite power. A computing cluster in orbit can enjoy free, limitless, zero-carbon power around the clock: the sun is always shining in space.
Of course, plenty of practical challenges remain to be solved. One obvious issue is whether and how large volumes of data can be moved cost-efficiently between orbit and Earth. This is an open question, but it may prove solvable, with promising work underway using lasers and other high-bandwidth optical communications technology.
A buzzy startup out of Y Combinator named Lumen Orbit recently raised $11 million to pursue this exact vision: building a multi-gigawatt network of data centers in space to train AI models.
As Lumen CEO Philip Johnston put it: “Instead of paying $140 million for electricity, you can pay $10 million for a launch and solar.”
Lumen will not be the only organization taking this concept seriously in 2025.
Other startup competitors will emerge. Don’t be surprised to see one or more of the cloud hyperscalers launch exploratory efforts along these lines as well. Amazon already has extensive experience putting assets into orbit via Project Kuiper; Google has a long history of funding moonshot ideas like this; even Microsoft is no stranger to the space economy. Elon Musk’s SpaceX could conceivably make a play here, too.
Update 2025-11-12 (PST) (AI summary of creator comment): Partial resolution framework:
If Rob Toews uses language like "Right-ish" or "Wrong-ish", the market will resolve to 75% PROB or 25% PROB respectively
Similar language ("mostly right", "nearly there", "in the right direction", or negative equivalents) will be treated the same way
If the resolution is ambiguous ("hard to say", "some right, some wrong", or other 50% answers), the market will resolve to N/A
Update 2025-11-12 (PST) (AI summary of creator comment): In cases where resolution is ambiguous or not clean cut, the creator may outsource resolution to Manifold moderators rather than resolving themselves.
Update 2025-12-05 (PST) (AI summary of creator comment): The resolution article from Forbes/Rob Toews is expected around December 8, 2025 (based on timing from previous years).
Update 2025-12-14 (PST) (AI summary of creator comment): If Forbes/Rob Toews does not publish a resolution article by end of Q2 2026, the market will:
Be extended indefinitely by default
Resolve to N/A (not NO) if capped at end of Q2 2026
1,000🏅 Top traders
| # | Trader | Total profit |
|---|---|---|
| 1 | Ṁ39,736 | |
| 2 | Ṁ1,200 | |
| 3 | Ṁ1,152 | |
| 4 | Ṁ500 | |
| 5 | Ṁ307 |
People are also trading
The Forbes article is over here: https://www.forbes.com/sites/robtoews/2025/12/14/grading-our-2025-ai-predictions-how-did-we-do/
The relevant quote from the article:
“Data centers in space” has become one of the trendiest and most consensus technology trends of 2025. Elon Musk has become a vocal champion of the idea, stating publicly that SpaceX is pursuing the opportunity. Same with Jeff Bezos and Blue Origin. Last month, Google announced a major new initiative named Project Suncatcher to put TPUs in orbit, with the first chips going up as soon as 2027. Eric Schmidt acquired launch company Relativity Space with the explicit goal of developing orbital data centers. Starcloud, an early pioneer of the concept of data centers in space, is working with Nvidia on the opportunity. Startups like Aetherflux are abruptly pivoting to get on the bandwagon.
Space is hard. It will take many years for the amount of computational power in space to scale to meaningful levels. But this is definitely happening—it makes too much sense not to do—and 2025 was the year that that became obvious to everyone.
I do think it's notable that Forbes didn't speak to the viability of data centers in space, only that money is being spent on it.
@cmiles74 but my point is that these aren’t serious attempts due to the viability. The money spent is a blatant investor hype scheme. If you understand a single thing about space you know ai will never exist there.
@Magnify Toews made a bunch of vague infalsifiable predictions and then Forbes asked him to come up with a resolution that preserved some credibility. He should be ignored.
@Magnify Agreed, the whole idea is nuts. I knew Forbes wasn't super interested in the science, I guess I should have realized that the decider would be dollars.
@MachiNi the predictions “Other startup competitors will emerge", and "one or more of the cloud hyperscalers launch exploratory efforts" were quite easy to falsify.
By the way, today Mastercard decided to invest in another space datacenters startup. There seems to be a lot of momentum among investors.
🙄
@ChurlishGambit , @AlanTennant , @uair01 – and for potentially others giving a low rating because you're disappointed in the resolution of this market: It has been made abundantly clear both in the description and in the comments that this is only about what the individual Forbes writer chooses to resolve the question to. I am sorry if you feel snubbed, but the information was definitely there.
@HenriThunberg What is the point of a market like this? How is this useful? How is it Futarchy? Next time title it "Will Forbes publish an article?" if that's what you're interested in predicting. & what a meaningless thing to predict
@HenriThunberg @MachiNi @ChurlishGambit
Actually the resolution criteria by Toews was quite explicit and easy to verify objectively:
“"Other startup competitors will emerge", and
"one or more of the cloud hyperscalers launch exploratory efforts"
Both were known to resolve yes since months ago.
We even alerted many times about that but everyone kept betting about a different thing
@HenriThunberg Also might be worth establishing what the resolution will be if Forbes doesn't publish an article looking back on their 2025 predictions this year/before the close, would it just resolve N/A?
@SisyphosDale Huh, I'd be very surprised but I guess not less than ... 2%?
By default I'd just extend it indefinitely, but would be happy to cap it by end of Q2 2026? And yes, N/A sounds way more right than NO in that case.
@HenriThunberg I'd be surprised too but also wouldn't sell 2%, especially conditional on the 2023 and 2024 being published on the 8th and the 10th. I think extending and capping makes sense, no strong opinion on when the cutoff should be, Q2 2026 sounds like it should be fine. And yeah since everyone's betting on what the article says N/A seems the most reasonable if there's no article.
@Jcd51
Please note the difference between your "development starts" and Rob Toew's "startup emerge" and "exploratory efforts launched". In uncertain domains we may see years of exploratory efforts in speculative bets before the development of the final solutions chosen start.
@MiguelLM Got it, the difference between we can make any progress at all on it, and startup teams can successfully defraud investors XD
@Jcd51 Investing can be simplified into to forecast a couple of numbers, and to multiply them:
how much profit is to be harvested
how likely it is this company will get how much of this profit
When the promised market is three orders of magnitude the current Earth datacenter capacity (baseline of around ~500 billion $ market in 2025), even if you give <0,01% chances of success the expected value is still very profitable.
@Jcd51 When the investors see big companies exploring the same space (Musk, Pichai) they reassure two things:
The chances are plausibly higher than 0%
They can probably sell their investment later to a big company
@Jcd51 Most companies evolve into doing something different than what they started doing. By the time they find out that datacenters in space will not work, they may have learned one or several valuable things the customers are willing to pay for. Maybe they will become experts in space computing for satellite usage, with a much smaller market than the original promise but still good enough to be profitable. Maybe they discovered something about materials durability or reliability that is applicable to components in Earth datacenters. When you work on the frontier of the current knowledge trying to overcome an almost impossible challenge, you may discover useful applications as a collateral result.
@Jcd51 When you put all of this into a model:
0.01% chances of success in becoming one of the biggest market in human history
10-20% chances of selling >10x to a big company in a few years
15-40% chances they discover something useful they can sell, different than their original business model
20-40% chances that they go bankrupt because 1) the original datacenter in space bet never happens; and 2) the company was not acquired by others; and 3) they were too late to realize that they have to seek other markets
… then it doesn’t seem so irrational investment. The numbers are completely invented, for illustrative purposes only.