Bitcoin and AI: Ammous on Valuations and Scarcity
Saifedean Ammous questions AI valuations and explains why he sees Bitcoin as a scarce asset with a different role in the market cycle.

Bitcoin and AI: Why Saifedean Ammous Sees Different Bets
People love to frame Bitcoin and the AI boom as if they're competing investment stories: one is fast-moving technology and growth, the other is a scarce digital asset. In a BMTV interview, Saifedean Ammous - author of The Bitcoin Standard - pushes back on that kind of comparison.
His view is that AI companies such as Anthropic and OpenAI might be "massively overvalued." He also argues Bitcoin and AI don't directly compete, because AI can't create more Bitcoin.
That's a thesis, not a settled fact. AI businesses can absolutely grow into big valuations, and Bitcoin's fixed issuance doesn't automatically translate into higher prices. Ammous also suggested Bitcoin was likely near a bottom and could start a fresh run. That forecast depends on timing - without a clear market date and supporting evidence, it's best treated as his view at the time of the interview, not something you should read as a live signal.
This article lays out the argument, the monetary backdrop behind it, and the questions investors should test before acting. It also separates a clearer idea - Bitcoin's programmed scarcity - from the much more uncertain claim that a particular market cycle has already turned.
Why Ammous thinks AI valuations may be stretched
Ammous's warning isn't the same as saying "AI is useless" or "AI will fail." His claim is more specific: some of the biggest AI companies at the center of the boom - including Anthropic and OpenAI - could be priced far beyond what their long-term economics can support. That distinction matters. A technology can be genuinely useful and economically important, while particular companies or investments tied to it still look overpriced.
Valuation doesn't come down to demand for a product alone. Investors need to judge how much revenue a business can keep, what it costs to deliver that value to customers, how competition pressures pricing power, and how much capital is required to keep improving the product. With AI, these questions matter even more because training and running advanced models can be expensive. Growing usage by itself doesn't settle whether a business can turn that activity into durable, repeatable profits.
There's also a difference between what a technology creates and what any single company captures. Customers might benefit from cheaper or faster work, but competition can squeeze the prices AI services charge. Open models, new entrants, infrastructure providers, and shifts in customer behavior can all change who ends up owning the economics. Even if the technology reshapes workflows, it doesn't guarantee every current leader becomes a lasting profit machine.
Because of that, "AI is overvalued" is too vague to help. A more useful question is: which company is being priced, what assumptions about future cash flows are baked into the valuation, and what would need to be true for those assumptions to hold up? A high valuation might be reasonable if a company builds durable customer relationships and holds strong margins. It might also be fragile if it relies on endless funding, unusually high pricing, or growth that competition makes hard to sustain.
Ammous's interview puts that valuation issue in the same conversation as inflation, gold, the dollar, and investing in AI. The practical takeaway isn't "dismiss AI." It's to separate belief in AI's usefulness from belief in the price investors are paying for exposure to it.
How inflation and debt shape the AI investment story
The interview connects AI spending with inflation, interest rates, public debt, gold, and stablecoins. Underneath those topics is a shared concern: the value of money and the cost of capital. That helps explain why investors may gravitate toward assets they think are scarce - but it still doesn't prove that any one asset is priced correctly.
When interest rates rise, financing gets more expensive and investors tend to scrutinize businesses whose profits are expected far in the future. A company valued mainly on distant growth can be more sensitive to higher discount rates than one already generating steady cash today. That's a general valuation mechanism, not a specific claim that rates alone determine the value of an AI company. Expectations, competition, and actual earnings still matter.
Public debt adds another layer. Heavy debt can intensify disagreements over future taxes, spending, and monetary policy. Some investors see that uncertainty as a reason to hold assets they think are less vulnerable to currency dilution. Others argue that inflation and debt outcomes depend on political and economic choices - and that even "scarce" assets can drop hard when market conditions shift.
Ammous also points to Argentina as one way to think about inflation. The broader lesson is simple: when currencies feel unstable, people often make practical choices to protect purchasing power. But you can't just copy one country's experience and expect the same path elsewhere. Inflation history, institutions, access to financial services, and policy responses are different. The comparison can clarify how money behaves; it doesn't automatically predict what will happen in the United States or the UK.
Stablecoins sit in a slightly different category. They're designed to track a reference currency and can make digital-dollar transfers easier, but they aren't the same thing as Bitcoin's fixed issuance. Their value depends on the issuer, reserves, redemption terms, and the rules governing them. In short: the monetary backdrop helps explain why Bitcoin gets attention, but you still have to evaluate the investment case on its own merits.
Why he sees Bitcoin and AI as different bets
Ammous's core distinction is scarcity. AI systems can generate text, images, code, and other digital outputs at scale. Bitcoin, on the other hand, follows a monetary issuance schedule enforced by network rules. Demand doesn't change the supply schedule, and no company can decide to increase Bitcoin production just because it would be convenient.
That said, Bitcoin isn't totally frozen in time. Its rules are maintained by a decentralized set of participants, and protocol changes are possible if users and network operators adopt them broadly. The point is that changing Bitcoin's monetary rules is difficult and requires coordination - unlike a conventional company deciding to issue more shares or a software firm increasing output.
So these two assets are different types of exposure. An AI company is a claim on a business: customers, products, competitive position, costs, and future earnings. Bitcoin is a monetary asset: its value depends on demand, perceived usefulness, liquidity, network security, and confidence in its monetary properties. Neither category is inherently "better." They ask different questions.
This also explains why Ammous says Bitcoin and AI don't necessarily compete. Capital can flow into both. AI adoption could boost demand for computing infrastructure and software, while Bitcoin can appeal to people who want an asset with a predictable issuance path. The "competition" is indirect: investors have limited capital and may reallocate based on rates, risk tolerance, and expected returns. But that's not the same as saying the technologies are direct substitutes.
Still, the scarcity argument has limits. A fixed or predictable supply doesn't create demand by itself, protect you from volatility, or guarantee a fair market price. Bitcoin's price can fall even if its issuance schedule stays unchanged. Scarcity is one piece of an investment thesis, not the whole thesis.
The cleanest reading of Ammous's point is this: analyze AI company valuations and Bitcoin's monetary scarcity with different frameworks. One focuses on what a business can earn. The other asks whether people keep valuing and using a scarce, decentralized monetary asset.
Does a Bitcoin bear market mean a bottom is near?
Ammous said Bitcoin was in a bear market, described that period as likely being near the bottom, and expected Bitcoin to "start on its own run in a bit." The first part is his assessment at the time. The last part is a forecast. A forecast shouldn't be treated as proof that the market has already turned.
Bear markets often look "clear" only after the fact. Prices can drop further even after people start calling the bottom, and recoveries can stall before a real trend forms. Macro conditions, liquidity, leverage, regulation, positioning, and unexpected events can all change the path. A confident-sounding cycle story doesn't remove those moving parts.
It also helps to keep separate a cycle view and a long-term thesis. Someone can think Bitcoin has durable monetary properties and still expect a prolonged decline. Another investor might expect a short-term rally without believing Bitcoin will hold value over the long run. When people mix these time horizons, you can end up making bad calls - like treating a long-term belief as permission to ignore near-term risk.
One practical way to judge a "bottom" call is to ask what evidence would confirm it - or weaken it. Is the claim based on a price pattern, changes in liquidity, shifts in investor behavior, or just the idea that the decline has gone on long enough? What would make the analyst change their mind? If there's no timeframe and no falsifiable condition, the forecast may be more like an opinion about direction than a usable market signal.
The interview summary doesn't include a detailed model or specific indicators behind Ammous's bottom call. So readers should keep the uncertainty instead of filling in missing reasoning. His view is worth understanding as part of the debate, but it can't tell you where Bitcoin is today or what happens next.
How to assess the argument without treating it as a forecast
If you're going to take a strong market opinion seriously, the best move is to translate it into questions you can actually test. Ammous's comments include three different propositions: some AI companies may be overvalued; Bitcoin and AI have different economic characteristics; and Bitcoin may be close to a cycle bottom. Each needs its own kind of evidence.
For AI valuations, focus on business economics rather than headlines. Look at customer retention, revenue quality, operating costs, capital requirements, and whether the company can defend margins as alternatives emerge. A company can be strategically important and still fail to produce returns that justify its valuation. Likewise, a company can look "expensive" and still be fairly valued if its long-term earnings potential is higher than the market currently expects.
For Bitcoin, separate protocol properties from market price. The issuance schedule can be analyzed independently of whether today's demand supports a particular valuation. Then evaluate what could affect demand: adoption, liquidity, regulation, custody access, and investor risk appetite. Don't treat a supply characteristic as automatic proof of a price outcome.
For the cycle call, decide your time horizon and risk limits before you act. A long-term investor and a short-term trader can look at the same information and make different choices. If you're considering an allocation, set ahead of time how much volatility you can tolerate and what evidence would change your mind. That's more robust than leaning on a prediction with no clear point where it stops making sense.
Key takeaways:
- AI's usefulness doesn't automatically tell you whether specific AI companies are fairly valued.
- Bitcoin's predictable issuance makes it different from a business that can expand output, but scarcity doesn't guarantee appreciation.
- Bitcoin and AI can play different investment roles while still competing indirectly for capital.
- Ammous's bottom call was an opinion at the time of the interview, not a reliable indicator of current market conditions.
- Treat valuation, asset characteristics, and cycle timing as separate questions.
This framework is intentionally less about picking a winner between AI and Bitcoin. It's also more practical: it clarifies what evidence matters, what remains uncertain, and what assumptions could turn out wrong.
Conclusion
Ammous's argument works best as a way to separate unlike things. AI companies should be judged by how their businesses actually work and what they can earn, not just by how important AI might be. Bitcoin should be treated as a scarce monetary asset with a volatile market price, not as a guaranteed hedge or a catch-all substitute for other types of tech investment.
His expectation that Bitcoin was near a bottom is a separate, time-sensitive call. Keep that separate from the longer-term debate about Bitcoin's supply and role. The real question isn't which story sounds better. It's which assumptions you can back up, what could disprove them, and how much risk you're willing to take.