Intro

Below is a summary of a Tweet I made as well as a market wizards podcast I did

Rant

The 4 Horsemen of the Crypto Bull

HORSEMAN 1: CBDC ROLLOUT AS THE CATALYST

Much like the automatic exchange of information preceded the 2017 bull, so the CBDC rollout is the catalyst for this bull

1. CBDC Implementation

The biggest catalyst is a European CBDC expected to be fanfared in October due to March Philip Lane ECB Speech

a. Tether as Existential Threat

• Outlined Tether as existential threat to European sovereignty
• Indicated a run on the banks if status quo prevails
  • Exacerbated if stables start providing yield
  • Results in decline of the Euro in their view

b. CBDC Viewed as Necessary

c. Fiscal Crisis Accelerates CBDC

• Reading between the lines: every G10 country on the brink of a bond crisis
• UK yields back where they were during Truss
• Japan bond market going haywire
• US credit downgrades

d. Needed for Taxation

• Tariffs hard to stick as taxation method

e. Populist Taxes

• CBDCs are an effective taxation method because they:
  • Allow wealth taxes
  • Can facilitate negative rates
  • Enable tiered transaction/sales taxes for different net worth

f. Get It Out Before It Hits

• CBDCs will be the world’s largest ever financial repression event
• Cash could be de-monetized to facilitate their adoption
• Per Lane’s speech: easy to say it’s a matter of national security

g. Global Problem

Because fiscal problems are global, there is no reason to think CBDC implementation stops at the EU.

Expected rollout:

1. EU
2. UK
3. Japan

h. Bearer Assets Needed

As it becomes more clear this is coming, people will allocate directly into crypto. What do they buy?

1. Some US stable coins - But they are impaired:

   • Linked to US govt
   • What if Trump loses and implements CBDC in the US?
   • Money frequently frozen in Circle
   • The US will work with Tether to get inflows but because the CBDC story is non-sovereign, flows will go into:

2. Bitcoin

3. Altcoins

   • More money may flow into altcoins agnostically because they are less chainanalysed than BTC
   • But they need “money” like properties (aka long track record)
   • So Dinocoins specifically benefit here

4. Privacy coins (probably asymmetric given their market cap)

HORSEMAN 1’s PUMPKIN HEAD

CBDCs are needed because the distraction economy is cooking our brains and ruining productivity. AI accelerates this.

Growth Deceleration

• GDP Growth is being revised down globally, not up
• Original AI Capex plans were based on 10% GDP growth per Satya
• But Capex is RAMPING despite GDP decelerating. Why?

Chat ≠ Work

• Time in app in ChatGPT is up 2x
• But skilled labor wage growth is still high
• Employers continue to complain about distracted employees and hiring shortages

AI is a Drug

AI will:

• Make extremely addictive porn
• Make extremely addictive video games
• Make people working at TikTok and Meta really good at their jobs and allow them to make the algo more addictive
• Create generative experiences such as AI companions that will consume huge amounts of electricity but also not improve productivity

Drug Dealers Winning

• All the companies making AI are video game and social media companies
• This should surprise nobody
• But that narrative cannot justify politicians turning on coal plants to make sexy elves
• So we pretend that productivity is going to go up, even though AI is going to cook everyone’s brain

CBDC Needed Worse Because Society Cooked

• Resulting inflation = ++
• Which means the bond crisis just gets worse
• Which means you need financial repression to prevent collapse
• Which will heighten the timeline for CBDCs
• Thankfully, AI will make everyone much faster at coding CBDCs

HORSEMAN 2: TRUMP’S BIG BEAUTIFUL BILL

Trump’s big beautiful bill is the forward driver for the coming AI leverage cycle

$370B of Present Value

a. Mining

• BTC mining becomes more appealing

b. Leverage Hits Because Tax Deductions

• People start buying GPUs with debt because they can write off the purchase in the first year vs depreciating slowly

c. Crypto Is Source of Leverage

• Crypto provides that debt via:
  • Lending protocols
  • A new category of stable coin (AI asset backed stable coins)

d. The Debt Can Finance Real Demand Overage

People buying GPUs to service demand overage:

• Microsoft Token usage growing 500% YoY whereas existing Capex build is only doubling
• Reasoning models make everything worse
• Video and photo gen make everything worse
• Overage is going to be insane (that’s why we are provisioning coal plants)
• Coreweave booked 4 years out

Gap is insane:

• The 300% delta is compounded by “sovereign AI” (i.e., the US cutting off China from its supply chain)
• Everyone building their own systems with permissioning and different rules (Saudis will have vastly different AI safety requirements)
• Blackwell and many other coming hardware upgrades will leave a huge “long tail” of increasingly capable consumer hardware that can do basic AI jobs

Quality is good enough:

• Open source models continue to perform good work
• A good open source model + web search is very strong
• Analogy: “You don’t need a 180 IQ intern to assemble you a daily news brief. He might even be worse at it than a 120 IQ intern who has done it every day for a year”
• DeepSeek R models continue to show viability of open source

The Use of Crypto AI

1. Long Tail players solving demand overage

   • Data centers literally cannot supply what is desired

2. Smooth Use

   • Overcoming regional friction from sovereign AI (i.e., the experience of buying AI from China vs US doesn’t matter on chain)
   • This is also a bet on innovation happening in random places, and being available in specific geo-fenced data centers

3. Verified Inference

   • In existing platforms you don’t actually really know what you’re buying
   • Unless you host your own hardware you don’t know if it’s being trained on
   • Running your own data center is very expensive and getting more so
   • Verified Inference is a crypto native product that lets you know you got a specific model response for a given prompt
   • Boring but is a natural blockchain use case for compliance sensitive businesses

4. Mining and AI Proof of Work

   • In 2017 BTC mining was almost 10% of TSMC demand. Now is sub 2%. Despite BTC dominance being very high
   • AI Capex investment is likely 10x that of BTC over next 10 years
   • AI based proof of work or “mining” use cases will become a dominant proof of work as a linear extrapolation of hardware trends
   • Thus “AI proof of work” as a category (whether that is TAO subnet mining, or Ambient Mining) will expand nearly infinitely

HORSEMAN 3: BANKS AND HEDGE FUNDS ARE READY

All the banks and global hedge funds know this is coming and can custody crypto now

The Guys In Charge Are Pilled

• The people telling these stories are heavily at banks: Jamie Dimon, Ray Dalio, top fund managers
• Fiscal problems no longer a fringe view
• Crypto is increasingly a macro asset class

Trump Changed The Rules

• Now the banks can own crypto directly
• OCC rules changed
• Post 2020 Banks can also invest in fintech
• Now there is a reflexive loop that looks like the 2016 XRP playbook:
  • Bank partnership → Price up → Bank partnership → Price up → Fund partnership → Etc.

This Works Best on Validator Coins

XRP, HBAR, and RPCA security models. The finance coins are especially appealing because the banks can bolt on solutions for crypto whales:

a. Private banking b. I-bank underwriting for their various MSTR clones c. Convert treasury issuances for their crypto treasury buys

The entire thing is an insane financial industry fee bonanza so the industry is incentivized to partner with crypto protocols now that Gensler is gone

AI Integration

• Banks/hedge funds are also the #1 adopters of AI for their workflows
• This is structural as financial work is not mostly creative and can be handled by existing training data (as compared to drug discovery which is a heavier lift)
• So there is even heavier whitespace for collaboration

Note: The author is the founder of an AI finance coin (Post Fiat) and thinks this bet is as close to a “sure thing” as anything they’ve ever seen in their career. Easier to underwrite than a market demand story in their view, which is why they’re a founder in this category, not an investor.

HORSEMAN 4: AI COMPANIES DON’T CARE ABOUT IP

Closed source AI companies simply do not care about property rights or IP

The Problem

Web3 always had a structural problem. Nobody knew really who you were protecting your data from. Google wasn’t that evil at the end of the day. Worst case scenario you got some really targeted ads. Facebook showed users are actually happier getting extremely targeted ads.

Now it is very clear who you are protecting your data from. Basically every AI company on earth endlessly scraping the web for:

• Your images
• Your thought process
• Your writing
• Any meta IP that you’ve divulged

It doesn’t matter if AI systems are creative. They devour IP. Even big companies cannot sue. If you ask for Games Workshop IP, OpenAI renders it with no payments to Games Workshop (this might change for big companies but it won’t change for you, random X user).

Possible Solutions

A. Consensual IP Mechanisms (long shot, but gotta try)

• Ala Story Protocol
• Governments probably won’t build a sane model to wrap the global IP quagmire
• Everything is worse because it’s all globalized and instantly translated in all languages
• There will be a single crypto winner at most in the “Consensual IP” category because governments won’t want to work with more than 1

B. Proprietary Data Lakes

• Groups of like-minded individuals will begin banding together and creating data lock-in with data that does not get shared outside the token owner group
• DAO models likely to form here
• It’s basically trade/data secrets on a blockchain with a financial monetization model, or licensing

C. AI Hive Minds

• This would be a bolt-on to Proprietary Data lakes
• Essentially you have AI agents administering monetization of the proprietary data in the lakes
• Called “web 4” - the idea that users are basically too cooked to monetize their own data, so AI will do it for them

WHY THE HORSEMEN FAVOR ALTS

Accelerator 1: Privacy

• Alts are simply less effectively tracked than BTC globally
• Given that CBDCs are a relatively near-term catalyst, CBDCs will drive demand for untracked extra sovereign bearer assets that are liquid
• Privacy coins benefit here too
• There have been legal changes re: Tornado Cash that benefit this further

Accelerator 2: Vibe Coding

• The main opex cost of crypto protocols is engineers
• As it becomes possible to have fewer people and less engineering costs and more effectively complete the roadmap

Accelerator 3: New Consensus Mechanisms

• LLMs can do very powerful things with governance
• At Post Fiat for example, LLMs are used to determine validator rewards in an objective manner that can be vetted by anyone with a model or set of prompts
• This previously wasn’t possible - or would require very expensive voting which could be easily bribed
• Verified inference makes this doubly possible (i.e., you can prove that you vetted a validator reward on chain)

Accelerator 4: ICOs

• Plasma is the beginning of a larger trend
• De-regulation means ICOs are probably coming back

Accelerator 5: Liquidity Dynamics

• Banks/funds will want to maximize their upside
• Buying BTC isn’t juicy enough for a directionally ‘sure thing’

Accelerator 6: Quantum

• BTC has an extremely conservative culture (see OP CAT saga)
• It will have a harder time adapting to quantum signatures
• This is highly relevant especially for the Satoshi wallet which has known pub keys/won’t move the funds

Accelerator 7: Leverage

• The MSTR premium dynamic + Bitcoin treasury strategies can easily reverse
• MSTR vol is at the lows making converts more expensive
• As more people issue equity against this strategy it compresses everyone’s premium
• The meta of CEP, for example, is that lightning network/strike has more or less stalled out
• So now the guy trying to move BTC forward as a network is doing a financial engineering scheme

Note: The author doesn’t want to trash talk BTC too much though. BTC is vastly preferable to CBDCs and they don’t want to bet against rabid idealists.

CONCLUSION

The author believes a basket of the following will outperform:

1. Privacy Coins

2. “Dino Coins”

• Coins with time-tested security models and liquidity (yes, XRP)

3. AI Coins Specifically

• AI debt stable coins (invested in USD AI)
• AI proof of work (invested in Ambient)
• Solid distributed inference plays (Akash, Nosana, Render, Bittensor and likely many others coming online)

4. Coins with Bank/Fund Partnerships/Validators

• LINK not a bad example
• Post Fiat (author’s own project) is in this category

5. IP Plays or AI Web3 DAOs

• IP, Grass, others

The author notes they’re naming coins so people don’t get mad. But they think there will be a ton of launches/ICOs and who knows what the list looks like in a couple years.

Overall thesis: Very bullish because we have:

• A clear catalyst: CBDCs
• An inevitable macro trend: AI being in huge demand, but cooking everyone’s brain
• An asset class which is perfectly suited to this cocktail: crypto (not just bitcoin)