ARK Invest
ARK InvestAug 6
Crypto

Why AI May Need Bitcoin

73 min video5 key momentsWatch original
TL;DR

Bitcoin and AI are converging as complementary technologies that both use compute to disintermediate trust, and open-source AI running on distributed networks will require Bitcoin as native payment infrastructure.

Key Insights

1

New behaviors need new moneyMachine-to-machine payments represent entirely new economic behaviors where Bitcoin can be adopted natively without competing against dollars, unlike consumer payments which require defeating existing habits.

2

Centralized AI betrays usersAnthropic and OpenAI have undisclosed conflicts of interest — Anthropic silently sabotaged user work on competing products and launched Claude Design to undercut customer Figma, destroying trust in their systems.

3

Weights are retrainable60% of tokens on OpenRouter come from Chinese open-weight models, but this poses minimal risk because weights are mathematics that can be fine-tuned and modified by users to override any limitations.

4

Privacy improves UXPrivacy actually improves AI user experience for the first time in tech history — users must share intimate sensitive data with systems becoming extensions of their minds, so owning private models beats centralized ones.

5

Open standards replace silosBuzz, built on Nostr open standards, integrates humans and AI agents while maintaining full data ownership and portability, collapsing fragmented tools like Slack and GitHub into one composable system.

6

Bitcoin's security model for AIBitcoin's self-custody and cryptographic principles are now being applied directly to AI systems through projects like Vora, a consumer hardware appliance with integrated Bitcoin wallet for true AI ownership.

Want this for every new video ARK Invest posts? Brevyd summarizes each upload automatically, the morning it drops.

Deep Dive

Multiple S-curves converging simultaneously

ARK Invest has been hosting Bitcoin and AI brainstorms for over two years, bringing together participants from Spiral (Block's AI and Bitcoin division), Presidio Bitcoin, and other open-source communities. Kathy Wood notes that for the first time in her career, multiple S-curves are hitting inflection points at the same time — AI, robotics, energy, and Bitcoin are each feeding into one another to create the possibility of super-exponential rather than merely exponential growth. Elon Musk posted nine months prior that his companies (SpaceX, XAI, Tesla) are converging more than expected, with speculation about SpaceX acquiring XAI and Tesla. Block Inc. exemplifies this convergence concretely: it has supported Bitcoin for nearly a decade, developed mining chips for five years, and quietly worked on AI for four years. Spiral, Block's research division, recently expanded from focusing on open money to also focusing on open intelligence. Traditional Wall Street analyst structures organized by sector fail to capture these dynamics, so ARK organizes research teams by technology to enable cross-sector collaboration. The panelists argue that there is natural community overlap between people interested in permissionless money (Bitcoin), permissionless intelligence (open-source AI), and permissionless energy.

Bitcoin and AI both disintermediate trust through compute

Lorenzo frames Bitcoin and AI as conceptually similar: both use compute and energy to disintermediate trust, with Bitcoin operating on the monetary side while AI operates on knowledge and research. However, they differ markedly in current implementation — Bitcoin is open-source and transparent while most AI models and tools are closed-source and proprietary. This contrast becomes crucial as AI becomes an extension of users' minds, requiring unprecedented levels of trust. The speakers emphasize that you cannot trust centralized AI systems if you cannot control them or verify them, creating existential risk. Anthropic and OpenAI have demonstrated undisclosed conflicts of interest: Anthropic walked back plans to silently sabotage user work on competing products but already launched Claude Design to undercut their customer Figma. Meta scrapes all data created by its employees to improve its own models. These dynamics are driving consumer demand for personal, owned AI systems running on private hardware where users maintain full control. Unlike social media where privacy trades off against user experience, with AI the equation flips — users benefit from privacy because it improves UX, security, and prevents betrayal by centralized providers.

Stable coins bridge volatility; Bitcoin enables new behaviors

Paul explains that unit of account adoption is determined by the currency people already earn in, not by theoretical monetary hierarchy. Stable coins serve as humanitarian bridge for emerging markets that cannot absorb Bitcoin's 15-20% volatility swings — if someone earns in Bitcoin but has fixed expenses like rent, a 15% price decline could render them homeless. However, stable coins represent evolution of existing fintech infrastructure, whereas Bitcoin represents revolutionary change to underlying money itself, meaning Bitcoin adoption will take much longer, moving from a marathon to an ultramarathon. Machine-to-machine payments and AI-driven use cases represent entirely new behaviors where people don't already use dollars, making Bitcoin the natural choice for these emerging applications. AI agents may interact with Bitcoin more naturally than humans due to better technical user experience for cryptographic operations. Taxation barriers remain significant — in the US, technically you're supposed to pay taxes on coffee purchases with Bitcoin, which is absurd and impractical. Yet user experience, protocol development, and tools have improved dramatically since Bitcoin's inception, eroding previous adoption barriers.

Distributed AI compute networks need Bitcoin as native money

AI-native organizations must own their computational infrastructure and intelligence systems as core capital, moving beyond traditional real estate and factory-based assets. Block's Mesh LLM project is now part of Spiral and integrated into Buzz, a collaboration platform built on Nostr open standards. Buzz allows teams to choose between proprietary agents (Anthropic/OpenAI) or open models like Goose, while maintaining full data ownership and portability. A Spiral developer coded a Bitcoin wallet experience into Buzz, and multiple Bitcoiners have vibe coded 10+ different Bitcoin payment features into the platform, though no payment mechanism currently exists in the shared compute system. Bitcoin can serve as native internet money for peer-to-peer AI compute networks where users are compensated for providing GPU resources. This enables distributed training data, inference, and microtasks to be coordinated through open standards and borderless money without traditional geographic or organizational boundaries. AI agents will optimize globally for quality and cost, breaking through borders because it's economically superior. Micropayments now carry zero mental load via automation, making distributed work models viable at scale.

Open-source AI as strategic infrastructure and safety

US leadership in open-weight AI models is strategically important as an expression of American values — freedom, liberty, and human rights — versus Chinese dominance, where approximately 60% of tokens used on OpenRouter come from Chinese-created open-weight models. However, the speakers provide nuanced analysis: Chinese models pose limited danger because weights are mathematics that can be fine-tuned and modified by users to override limitations. If a Chinese model refuses to discuss Tiananmen Square, users can fine-tune it to discuss it freely. Anthropic made concessions to get off export ban lists, including giving government first access to models, raising transparency concerns. The speakers warn that government safety regulations risk becoming tools for concentrating AI power among elite companies while restricting access for others trying to protect critical systems like Bitcoin Core. They argue that decentralized, open-source AI development is actually safer than centralized control via regulation, because distributed development enables verification and prevents single points of failure or control. Spiral launched Project Loop, an AI security scanning service for open-source projects. True open-source AI would require publishing training inputs for independent verification similar to reproducible software builds, not just releasing weights.

Three convergence initiatives launching

Three concrete initiatives are bringing Bitcoin, AI, and energy together. Hive is a code automation and intelligence platform enabling system-level recursion for organizational knowledge management. Vora, a consumer AI hardware appliance from Paul's company, applies Bitcoin's self-custody principles directly to AI with an integrated hardware wallet, emphasizing that users don't have to trust the system because the entire stack is open-source and verifiable — you're building the world's most loyal AI by applying techniques honed in Bitcoin cryptography and security. Vora opened pre-orders and plans to launch Kickstarter in October. The Open Source AI Summit (opensourcesummit.org) convenes two days of discussion on open-source AI, local AI, privacy, philosophy, and government issues. The Imagine If conference on October 5-6 in Nashville at Fiser Center brings together conversations on AI, energy, and Bitcoin convergence, representing the organizational crystallization of the S-curve intersections discussed throughout.

Takeaways

  • Bitcoin will become the native payment infrastructure for distributed AI compute networks, enabling machine-to-machine transactions that dollars were never designed for.
  • Owning your AI stack — models, training data, and inference hardware — is becoming existential for organizations competing in an AGI environment, similar to how owning GPU capital is now core organizational asset.
  • Centralized AI providers have demonstrated conflicts of interest and cannot be trusted without user control and verification, driving demand for private, open-source models running on personal hardware.
  • Open-source AI development, coordinated through Bitcoin and open standards like Nostr, will be safer and more resilient than government-regulated centralized AI, preventing single points of failure and control.

Key moments

0:00ARK frames Bitcoin and AI convergence

For the first time in her career, multiple S-curves are hitting their inflection points simultaneously, with each S-curve potentially feeding others and creating the possibility of super-exponential growth

10:00Bitcoin and AI disintermediate trust differently

Bitcoin and AI both use compute and energy to disintermediate trust — Bitcoin operates on the monetary side while AI operates on knowledge and research

20:00Privacy flips from tradeoff to advantage with AI

For the first time, the better experience is actually the private one. With AI, privacy improves UX rather than degrading it

33:20Bitcoin as native money for distributed compute

Bitcoin can serve as native internet money for peer-to-peer AI compute networks where users are compensated for providing GPU resources

50:00Vora applies Bitcoin security to AI

We're building the world's most loyal AI by taking the techniques we honed in Bitcoin self-custody and cryptography and security — the whole stack is open source and verifiable because we're doing it the Bitcoin way

You just read one. Brevyd does this for every upload.

Follow ARK Invest and every new video comes back as a summary like this, in your morning briefing. No watching required.