Cardano Founder Hoskinson Says Smartphones, Not New L1s, Will Decide Crypto’s Next Decade

Cardano founder Charles Hoskinson argues that crypto’s next decade will be decided by smartphones and AI agents, not by new Layer-1 blockchains. Hoskinson just crossed out a huge chunk of the industry’s favorite roadmap. His read is blunt: crypto’s future runs through the smartphone, not another Layer-1, and AI agents will run most of the internet by 2035. My take: that lands harder than another TVL chart because it points at distribution, not just infrastructure.

He made the comments across a wide-ranging session covering crypto’s original mission, JPMorgan’s blockchain pivot, and what he sees as the real bottleneck for adoption. This was not a price call. It was a strategy call. And yes, the targets look like a quiet shot at the L1 arms race that has pulled capital toward Solana, Sui, Aptos, and the rollup field since 2021.

The thesis comes in five parts:

1. Mission: Crypto, he says, was built to change the world, not to enrich the people who broke the global economy in 2008.
2. Institutional pivot: JPMorgan, which once cut off customers’ bank accounts, now ships a blockchain product. He calls that a win for the industry, even if it stings ideologically.
3. UX bottleneck: User experience in crypto is still too hard, and the answer is not yet another base chain.
4. Abstraction: What’s needed is account-level and chain-level abstraction.
5. Distribution: It has to live where users already are. The phone.

The smartphone thesis: secure enclaves beat hardware wallets

Hoskinson’s core claim is that modern smartphone secure enclaves now offer stronger key protection than dedicated hardware wallets like Ledger or Trezor. This is the sharpest claim in the whole argument. Modern phones already ship with a hardened secure enclave for sensitive data, and in his estimate that environment is stronger than what Ledger or Trezor offer in dedicated hardware. That is a direct challenge to a roughly $1B hardware wallet market. It also puts mobile-native wallet stacks in the center of the map.

Then comes the longer horizon. AI agents are becoming more important to the internet than humans, and by 2035 most search and commerce will route through agents. People, in his telling, end up in a separate lane he calls the “human internet.” Why does this matter? Because privacy, agents, and blockchain cannot be glued together at the end like a compliance PDF. I’ll be honest: that last part is the line many traders will skip, and it is probably the line with the longest shelf life.

Adoption angle: the bottleneck has moved from throughput to mobile UX

The investment implication is that mobile wallet experience and account abstraction now matter more for adoption than transaction throughput. The first crypto angle here is adoption, and it cuts against the prevailing trade. Capital has flowed relentlessly into new L1s and L2s. Solana. Sui. Aptos. The rollup field. The bet was simple: better infrastructure unlocks the next user wave. Hoskinson is saying the bottleneck moved.

Most guides still frame adoption as a throughput problem. That’s only half right. If the next billion users come through smartphone-native wallets with abstracted accounts, the prize goes to whoever owns the mobile experience, not whoever ships the fastest VM. Coinbase’s Smart Wallet push, MetaMask’s mobile-first redesign, and the steady pressure on Apple and Google to loosen wallet restrictions start to look less like product polish and more like positioning for the only distribution channel that scales. ADA itself trades light on this narrative. Cardano’s mobile story has been quieter than its academic one, and that gap is not a footnote for ADA holders.

Institutional angle: JPMorgan’s blockchain pivot signals settlement-layer repricing

According to Hoskinson, JPMorgan shipping a blockchain product proves the cost of not having a blockchain rail now exceeds the cost of building one. The second angle is institutional, with regulation sitting right beside it. JPMorgan is the tell. When the largest US bank ships a blockchain product after years of public skepticism from its own CEO, the signal is not ideological conversion. It is cost accounting.

That is the same logic pushing BlackRock’s BUIDL past $500M in tokenized treasuries and pulling banks like BNY Mellon and State Street into custody. Counter to the usual crypto purity test, JPMorgan showing up on-chain is not automatically bad for the original thesis. It means the rails won enough that the incumbents had to use them. For traders, the implication is straightforward: regulated rails are absorbing the infrastructure premium that retail-facing chains used to capture, and tokens whose value depends on speculative throughput face a harder bid than tokens tied to settlement, custody, or compliance plumbing.

AI-agent angle: privacy must be built into the agent stack

Hoskinson predicts that by 2035, AI agents will conduct most online search and commerce, requiring crypto-native wallets, identity, and zero-knowledge privacy tooling built in from the start. The AI-agent prediction is the wildcard. It will get pushback, and some of that pushback is fair. A 2035 world where agents run most search and commerce is no longer fringe, given public roadmaps from Anthropic, OpenAI, and Google. But the crypto-native version of that thesis is still thin in production.

The agent infrastructure requirements break down as follows:

• Transactions: If agents transact, they need wallets.
• Identity: If they negotiate, they need identity.
• Rails: If they operate at machine speed across thousands of services, they need permissionless rails and programmable money.
• Privacy: Agents acting on behalf of humans cannot leak the human’s data into every endpoint they touch.

Is this overkill? For a 50-page site, maybe. For agents touching wallets, purchases, search history, and identity across thousands of services, no. That’s the lane projects like Virtuals, Fetch.ai, and the broader on-chain agent stack have been pitching, with mixed execution and plenty of hype. Hoskinson’s useful addition is forcing privacy into the mandatory bucket. That points toward zk-tooling, selective disclosure, and confidential computing as part of the agent stack rather than separate verticals. Honestly, that’s the part of his read I’d argue with the least.

What Hoskinson didn’t say

Notably, Hoskinson made no ADA price call, no Cardano product pitch, and no competitor attack. The remarks read as an industry strategy memo, not a founder talking his book. Worth noting what he did not say. He did not talk ADA price. He did not pitch a Cardano product release. He did not take the easy lap on Ethereum or Solana. The remarks read more like a strategy memo aimed at the industry than a founder talking his book.

That tone is itself a signal. Yes, this contradicts the usual founder-playbook read, where every public appearance gets translated into token promotion. Bear with me. Cardano’s leadership is publicly framing the next cycle around mobile UX, abstraction, and agent infrastructure rather than TPS or ecosystem grants. Whether ADA’s roadmap actually delivers against that framing is a separate question, and one the market will answer in basis points, not press cycles.

What this means for traders and investors

The actionable signal is a quiet repricing: capital should rotate from pure-throughput L1s toward mobile wallet, account abstraction, and on-chain agent infrastructure. The signal is a quiet repricing of where infrastructure capital should sit. If the bottleneck is mobile UX and account abstraction rather than a faster L1, then the long tail of new base chains looks expensive next to wallets, abstraction tooling, and agent infrastructure.

Here is where I would look for relative strength and weakness:

• Relative strength: Tokens tied to smart-account standards (ERC-4337 ecosystem), mobile wallet distribution, the early on-chain agent stack, and privacy tooling that can actually serve agent workflows.
• Relative weakness: Pure throughput plays that don’t have a mobile or abstraction story.
• The ADA read: Hoskinson’s framing flatters Cardano’s research-heavy posture, but ADA needs to ship a credible mobile and agent story to convert the narrative into flows. The token has spent most of 2026 range-bound while Solana and the L2 set captured the speculative bid. That gap won’t close on talk.

Watch four things next, in this order:

1. Cardano deliverables: Any Cardano roadmap update or Input Output post that translates Hoskinson’s smartphone and agent thesis into specific deliverables and dates. Without that, the remarks sit as commentary rather than catalyst.
2. JPMorgan cadence: Each new institutional rail that goes live raises the floor for compliance-aligned tokens and lowers it for chains that depend on regulatory ambiguity.
3. On-chain agents: The agent cohort, Virtuals, Fetch.ai, and the smaller agent-tooling names, for early signs of real agent-driven transaction volume rather than narrative volume.
4. Mobile policy: Any Apple or Google policy move that loosens restrictions on mobile wallets and in-app crypto payments. That’s the gate Hoskinson is implicitly betting gets opened.

We tried to reduce this to a normal L1-versus-L2 trade. It doesn’t fit. If even one of those four breaks the right way in the next quarter, the smartphone-and-agents thesis stops being a Hoskinson talking point and starts becoming a tradable rotation.