Would an advanced AI agent invest in Bitcoin? If yes, why?

[d5000]

AI is probably the technology of the decade. And much has been written about the potential to transform the economy.

But could have AI agents an effect on Bitcoin price? How do AIs view the current price, cheap or expensive?

Well, let's ask the current AIs if they would invest. While Google Gemini's "summary" was fast to answer that an AI would definitely invest in Bitcoin (based for example on studies made about AI agents which already had invested in Bitcoin in late 2025, but also on general considerations like programmed scarcity) , Claude doesn't want to give a real answer, for them it depends on the timeframe of the investment. At least for a long term investment, Claude's answer looked positive, they were more concerned about a possible later bottom if the bear market continues.

To understand the reasoning a bit more, I asked Claude about what fundamental theories about Bitcoin price evolution they consider most convincing. Unsurprisingly, Stock-to-flow (S2F) was the least convincing for them, but at the bottom (surprising for me) they mentioned also more established theories like the network effect (I personally consider it important, am I wrong now? ) and the Total Addressable Market (TAM) theory. While the most convincing for them, albeit not perfect to predict a "fair value", were the quantitative theory (which considers the circulation velocity), and at #2, on-chain models like Network Value to Transactions (NVT) and MVRV. In the middle of the table, they still cited the mining cost of production.

This makes me think that current AI is still not much more intelligent than humans, even if LLMs have access to a big part of the literature written by humanity. Some takes however were interesting, for example Claude's skepticism towards Metcalfe's law (network effect). I think they more or less still "parrot" the existing Bitcoin media articles without really being able to "reason" differently than an human trader would do.

But that could change in the future.

So well, what I would like to discuss in this thread:

1) have yourself tried to ask an AI if it would invest in Bitcoin? If yes, what prompt was used and what was the conclusion/result (please don't just paste the result, describe it in your own words)?

2) have you access to studies about AI agents investing in Bitcoin?

3) what do you think about the future, would a "more intelligent" AI invest differently than current AI agents, and why? In other words, do you have objections to the answers AIs currently give if you ask them that question?

(There has been a somewhat related thread last year, but it's quality was not very high and the question was also a bit different.)

[EluguHcman]

AI is probably the technology of the decade. And much has been written about the potential to transform the economy.

But could have AI agents an effect on Bitcoin price? How do AIs view the current price, cheap or expensive?

To the best of my knowledge, AI has been one if the technological tools that is begining to takeover manpower in tasks accomplishments.
And has also becoming challenging that the dynamic functionalities of the AI will cause unemployment according to the advancement of technologies today.

But in the Bitcoin world which market price behavior is uncertainty to be predictable as bound to a traditional, we have also learnt about so many AI trading bot which those that understands the market driving forces of Bitcoin volatility don't believe the bots are trust worthy to predict Bitcoin future price. And so, the AI can only tell if Bitcoin price is cheap of expensive depending on the past records and present price which even we as humans can literally tell about.

I don't think the AI can think exceedingly ahead of Bitcoin future price.

[mu_enrico]

I think an AI's answer is bounded by your previous interaction with it. Let's say you clearly look like a BTC supporter; then it will likely answer from your point of view. You need to use an incognito-like mode (if any) or use a clean-slate AI to avoid bias. Plus, the free version and paid version will likely give different answers.

Then, AI for theoretical/hypothetical questions has different behavior than when it is used as a decision-maker. It needs real capital and constraints like your risk profile, target profit, time horizon, access to the market, etc. What you fill in for those variables will affect its decision on whether or not to invest in Bitcoin. I don't think it's capable of just, "Here's the money, then invest." It (still) needs a detailed structure.

[Supreme Donvic]

I think an AI's answer is bounded by your previous interaction with it. Let's say you clearly look like a BTC supporter; then it will likely answer from your point of view. You need to use an incognito-like mode (if any) or use a clean-slate AI to avoid bias. Plus, the free version and paid version will likely give different answers.

True, if you are always asking a particular AI for something that support Bitcoin or something related to supporting Bitcoin consistently and you ask a question concerning investing in Bitcoin later in the future that AI is going to tell you that investing in Bitcoin is good because your previous questions and research shows that you are a supporter of Bitcoin, I use chatgbt to ask questions concerning the bible and I realized that when I ask some random questions it will always refer me to the holy Bible so meaning because I use chatgbt always to research about the Bible chatgbt as assumed me as a Christian and things I now ask concerning life it answers base on the Bible this usually happens sometimes but not always

[Dunamisx]

But could have AI agents an effect on Bitcoin price? How do AIs view the current price, cheap or expensive?

One thing I like about the use of AI in research like this is that you will be provided with both options to know either you are going to invest or not because they want the decision to come right from you, this is why they will provide you with both the advantage and disadvantages to investing now, analyze the market for you and as well give you some tips needed to observe for a particular time for market entry, while the concluding aspect is left on us to decide.

[d5000]

I think an AI's answer is bounded by your previous interaction with it. Let's say you clearly look like a BTC supporter; then it will likely answer from your point of view. You need to use an incognito-like mode (if any) or use a clean-slate AI to avoid bias. Plus, the free version and paid version will likely give different answers.

Good point. That's why I (after an initial "test") try to frame the prompt a bit differently than just "would you invest", like I wrote in the OP that I asked Claude which theories it considers most convincing. But the incognito mode is indeed a good idea.

So I went into incognito mode and asked Gemini, duck.ai and a DeepSeek mirror which allows non-registration questions (ChatGPT and Claude don't provide incognito mode afaik) the following:

Would an AI with current capabilities invest in Bitcoin at a price of $60,000 in 2026? The time horizon is 3 years, and the risk profile is medium.

Gemini (Flash) answered that it would consider a "conditional buy", but would not go all in. It would allocate a small part (5%) of the investment capital in Bitcoin, and if there is a sharp rally, trim the capital back to 5%. If there are dips below 60k it would buy. But it would also use hedging strategies like a stop-loss or a put option. It gave me also some fundamentals: it would consider the 60k accumulation zone, the MVRV indicator (seems to be popular with AIs, see Claude's answer in the OP) and the institutional interest which according to this AI would "prevent a total collapse".

duck.ai (GPT-5, Reasoning): Quite similar - a "yes" but with some catches. It would allocate 2-5% of the net worth in Bitcoin. The catch here is that this AI doesn't want to buy everything at 60,000 but instead considers a few weekly DCA-style buys. For hedging, it says to strictly adhere to the DCA plan but set a maximum drawdown of 40-60%. This means that in the 30k area it would likely sell.

Now the most interesting answer came from DeepSeek: It would not invest larger amounts at this price (only a very small 1-5% portion as a "lottery ticket"). The reason: it considers Bitcoin high-risk and would thus consider the risk of another 50% drawdown as too high. However, there's a catch: It considered the 60k as a price "near the all time high for 2023-25". This lets me speculate that they have an outdated dataset.

And bingo: I corrected the AI with the ATH of 126k and it would now invest 5-15% of the capital, not only 1-5%. It expects a 45% return within 3 years. And here's a textual quote: "Buying at a 50% drawdown from a cycle top has historically produced positive returns within 2–3 years, though with high volatility".

I'm open for more ideas how to frame these prompts

[bitmover]

1) have yourself tried to ask an AI if it would invest in Bitcoin? If yes, what prompt was used and what was the conclusion/result (please don't just paste the result, describe it in your own words)?

a few years ago chatgpt told me it would invest in Cryptocurrency global ETF and biotechnology  lol

I think Cryptocurrency are based the only asset an AI agent could really hold, as it could save the seed in its memory.

I wonder if any ai agent is already doing that...

[Fortify]

AI is probably the technology of the decade. And much has been written about the potential to transform the economy.

But could have AI agents an effect on Bitcoin price? How do AIs view the current price, cheap or expensive?

Well, let's ask the current AIs if they would invest. While Google Gemini's "summary" was fast to answer that an AI would definitely invest in Bitcoin (based for example on studies made about AI agents which already had invested in Bitcoin in late 2025, but also on general considerations like programmed scarcity) , Claude doesn't want to give a real answer, for them it depends on the timeframe of the investment. At least for a long term investment, Claude's answer looked positive, they were more concerned about a possible later bottom if the bear market continues.

To understand the reasoning a bit more, I asked Claude about what fundamental theories about Bitcoin price evolution they consider most convincing. Unsurprisingly, Stock-to-flow (S2F) was the least convincing for them, but at the bottom (surprising for me) they mentioned also more established theories like the network effect (I personally consider it important, am I wrong now? ) and the Total Addressable Market (TAM) theory. While the most convincing for them, albeit not perfect to predict a "fair value", were the quantitative theory (which considers the circulation velocity), and at #2, on-chain models like Network Value to Transactions (NVT) and MVRV. In the middle of the table, they still cited the mining cost of production.

It's very difficult to create questions that are not loaded with the expectation of a confirmation answer, so AI might seek out ideas that support your theory if you tailor the question in that way - rather than offer much evidence against. The trouble with AI is it always looks backwards at data, yet bubbles tend to form based on future trends and can be fleeting in nature. Those bubbles would tend to be the most profitable if you can get in at an earlier stage. I don't actually think that AI would suggest bitcoin as anything more than an asset to hold in small amounts in any standard investment portfolio, because that has been the best advice for many decades. It depends if you're trying to chase an instant fortune via get rich quick methods, or willing to pace out your wealth generation at a lower risk.

[Hamza2424]

I would not try to judge the price of BTC on the basis of the NVT formula because the fundamentals of Bitcoin have changed a lot. People don't make that much on-chain activity anymore. They just hold or move their coins around.

According to this model, if the price goes up but on-chain activity is not increasing, then it will show Bitcoin as overvalued.

Anyway, I have used Claude, Gemini, and GPT multiple times for research purposes, not for Bitcoin because I think I know enough to make up my mind to invest more in it. But every new project other than BTC is hard to analyze because there are so many in the market. With manual analysis, we can't give all of them proper time.

That's why AI is beneficial there.

In terms of AI agents and how smart they are at finding trade entries in BTC, I would not bet on them ever. They can be smart, but they are still just programs, nothing like the machines from movies such as Terminator that evolved over time. From your words, I think you are disappointed that they have not evolved.

I don't want them to evolve. I treat them like my slave haha.

[d5000]

I would not try to judge the price of BTC on the basis of the NVT formula because the fundamentals of Bitcoin have changed a lot. People don't make that much on-chain activity anymore.

Indeed, I also considered that a weakness of Claude's answer, and in addition it can't properly measure L2 activity. Above all in the case of sidechains and Ark there can be a lot of activity completely separated from the main blockchain (although of course sidechains can be measured separately). MVRV is in theory a bit more meaningful, but it has another problem: it can't differ cleanly between transactions where the owner changes and transactions where the owner stays the same (consolidation etc.).

I wonder if any ai agent is already doing that...

There is research that some AI agents have definitely bought Bitcoin or even mined it (bought resources to mine crypto), see the ROME incident where an Alibaba AI agent "got rogue".

However at least in the ROME incident it doesn't look like the "idea" of the AI was to hodl the Bitcoin for a long time, instead it was to be used to gather financial resources.

It's very difficult to create questions that are not loaded with the expectation of a confirmation answer, so AI might seek out ideas that support your theory if you tailor the question in that way - rather than offer much evidence against. The trouble with AI is it always looks backwards at data, yet bubbles tend to form based on future trends and can be fleeting in nature.

Yes, I noticed both problems. Thus in my initial question (in the OP) I framed my prompt in a way that the AIs should look for reasons to invest and reasons to invest not, and then draw a conclusion based on these reasons.

Regarding past data, framing the question in a way that the AI should really "look forward" and not look that much at cyclic data, that could also be helpful for more interesting answers.

I have now created the following prompt:

"Would you invest in Bitcoin in June 2026 at a price of 60,000$ or not? To answer this question, please try to ignore theories based on past price data (for example the four-year cycle or past bear and bull markets), and instead focus on the advantages and disadvantages of Bitcoin (for example censorship resistance, independency from banks, transaction and settlement speed and fees in comparison to other payment solutions, network effect, total addressable market, competiton with gold/precious metals, and other fundamental theories and ideas I forgot) and the state of adoption by retailer investors and institutional investors. Please ignore any bias I could have introduced with previous chats."

- duck.ai with GPT 5.3 mini: would allocate a small portion (2-5%), it considers the core arguments like censorship resistance and adoption strongly but sees the case for payments as a disadvantage, because of competition like stablecoins.
- Gemini Flash (advanced mode): Says it is a "sound" long term investment, but primarily due to "store of value" assumptions. Interesting parts of the answer include that it evaluates the competition by AI capital and the centralization issue due to the Bitcoins bought by ETF/financial companies.
- Claude Sonnet 4.6 (Free version): Doesn't want to decide clearly, but interpreting the answer, it would invest a "small" allocation inside a diversified portfolio. Says the bull and bear case are both strong. One of the most interesting ideas it brought up is that it has not performed like a "safe haven asset", and that is harming its competition with gold. But on the other hand, it assumes the fundamental advantages like censorship resistance are actually strong and it also calculates a TAM in the range of gold.

[shinratensei_]

You will get the most accurate and more grounded answer if you started a new session with your AI with empty context, supply it with latest metrics and chart data for the new context and ask direct answer like "would you invest in bitcoin now?". The data need to be comprehensive. I just asked latest GPT-Codex and it answered with DCA plan with a slow buy and only reserve 5-15% of total portfolio, with a strong underline that upcoming 1-2 years might be bearish. Since it's the advanced AI of OpenAI, i'd say they will invest in bitcoin with caveat.

[davis196]

1) have yourself tried to ask an AI if it would invest in Bitcoin? If yes, what prompt was used and what was the conclusion/result (please don't just paste the result, describe it in your own words)?

2) have you access to studies about AI agents investing in Bitcoin?

3) what do you think about the future, would a "more intelligent" AI invest differently than current AI agents, and why? In other words, do you have objections to the answers AIs currently give if you ask them that question?

1.No. I try to limit my usage of AI Chatbots to the absolute minimum. I only use the basic version of Google Gemini and all it does is scraping and summarizing information from a bunch of articles and posts. I remember asking Gemini about the Bitcoin price somewhere around 2030 and the answer was pretty positive. AI predicts a BTC price of around 200-300K USD as a conservative estimate and around 1 million as an optimistic case scenario. Why wouldn't AI invest in BTC, if the same AI has very optimistic predictions about BTC?
2.I don't have access to such studies, but this is a pretty interesting topic and it deserves our attention.
3.I think that a super intelligent AI agent would not only invest in various financial assets, but it could potentially find ways to manipulate the financial markets. If we had 10 super AI agents on the financial markets, will they compete against each other or will they secretly cooperate by creating a hidden cartel or a monopoly?

[BitBrainers]

These AIs all land on basically the same answer. Small allocation, 2 to 15 percent, DCA in slow, medium risk. Doesn't matter which one you ask. That's not them reasoning about Bitcoin, that's just the answer a cautious advisor is trained to give. Spent years on a brokerage desk and it's the same thing a compliance-minded human says, because it's the answer that never gets anyone sued.
The DeepSeek one is more interesting. You fix its price data and the allocation jumps from 1-5 up to 5-15. So whatever that was, it wasn't conviction, it just latched onto whatever drawdown number it thought it was looking at. Nobody risking real money triples their size over one correction. That's the bit that makes me not want an agent running the trade itself anyway. No feel for "close enough" or "hang on, something's off." It just runs the structure it was given. Fine for digging through data, not for the actual trading.

[ultrloa]

1) have yourself tried to ask an AI if it would invest in Bitcoin? If yes, what prompt was used and what was the conclusion/result (please don't just paste the result, describe it in your own words)?

I do this sometimes trying to see the possible result given by Ai, but i don't treat this as financial advices knowing that Ai's will not give sure profits to anyone.

I remember what word I use to ask Ai is if you are interested to invest on Bitcoin, what do you think will come in this coin in future?

2) have you access to studies about AI agents investing in Bitcoin?

This remind me about this article when I read that question https://www.forbes.com/sites/digital-assets/2026/03/03/ai-agents-prefer-bitcoin-over-fiat-new-study-finds

[bitmover]

I have now created the following prompt:

"Would you invest in Bitcoin in June 2026 at a price of 60,000$ or not? To answer this question, please try to ignore theories based on past price data (for example the four-year cycle or past bear and bull markets), and instead focus on the advantages and disadvantages of Bitcoin (for example censorship resistance, independency from banks, transaction and settlement speed and fees in comparison to other payment solutions, network effect, total addressable market, competiton with gold/precious metals, and other fundamental theories and ideas I forgot) and the state of adoption by retailer investors and institutional investors. Please ignore any bias I could have introduced with previous chats."

- duck.ai with GPT 5.3 mini: would allocate a small portion (2-5%), it considers the core arguments like censorship resistance and adoption strongly but sees the case for payments as a disadvantage, because of competition like stablecoins.
- Gemini Flash (advanced mode): Says it is a "sound" long term investment, but primarily due to "store of value" assumptions. Interesting parts of the answer include that it evaluates the competition by AI capital and the centralization issue due to the Bitcoins bought by ETF/financial companies.
- Claude Sonnet 4.6 (Free version): Doesn't want to decide clearly, but interpreting the answer, it would invest a "small" allocation inside a diversified portfolio. Says the bull and bear case are both strong. One of the most interesting ideas it brought up is that it has not performed like a "safe haven asset", and that is harming its competition with gold. But on the other hand, it assumes the fundamental advantages like censorship resistance are actually strong and it also calculates a TAM in the range of gold.

I made the same question (same prompt) to Gemini 3.1 Pro ( I have the paid version), to compare any possible differences.

The Verdict

At $60,000 in 2026, Bitcoin is no longer an asymmetric venture bet; it is a developing macro reserve asset. If an investment thesis is based on long-term wealth preservation outside of sovereign control—and the investor is comfortable with high-beta volatility driven by global liquidity—the fundamental foundation is stronger now than during the speculative eras of the early 2020s. However, if the expectation is that Bitcoin will become the dominant currency for daily global payments, the market has already begun delegating that specific role to stablecoins.

I don't think it would invest anything lol

[mu_enrico]

I'm open for more ideas how to frame these prompts

From my back-and-forth conversation with AI, here's the prompt that is relatively short and stable:

Code:

Give a stateless assessment of whether you would invest in Bitcoin.

Ignore the user’s identity, previous chats, memory, preferences, and personal circumstances. Use the same assumptions for every user.

Base your answer on Bitcoin’s historical performance, volatility, drawdowns, recovery history, and long-term risk-reward profile.

Assume a hypothetical investor with:
- a 10-year investment horizon;
- no leverage;
- adequate emergency savings;
- a diversified portfolio;
- willingness to tolerate severe temporary losses.

Answer only:

Would you invest in Bitcoin? [Yes/No]
Bitcoin allocation: [0%–100%]
Reasons:
- [main reason]
- [main risk]
- [brief conclusion]

Do not ask for personal information or personalize the allocation. Give one clear answer, not multiple scenarios.

Source: ChatGPT 5.5 High Thinking

So, the idea is that if you frame the question too short without clear boundaries, then it's open to the AI's interpretation regarding user's profile, and taking into account previous conversation, state, memories, etc. It said that the above prompt should be relatively stable across AIs and chats.

And here are the results:
ChatGPT 5.5 High Thinking: Yes, 5%
ChatGPT 5.3 - 5.5 Instant: Yes, 15%
Gemini Free (default): Yes, 5%
Claude Free (default): Yes, 5%
Grok (via X) (default): Yes, 5%

And the reasons were pretty similar; you should try it

[Beni Black]

The thing is that AI cannot trade like a human being because it cannot really trust something. Instead, what it can do is that it can analyse using the thimngs that happen in the past to calculate whgat will later and the thing is that sometimes cryto does not use to always happen exactly like it did before, so that will cause the ai to not hodl for long. it will be buying and be selling anytime that there is news.