開發者數量腰斬之後：Crypto沒有死，只是把人才讓給了AI

Original Authors: Xinyang, Ethan, IOSG Ventures

In 2026, the GitHub activity curve of the Crypto open-source community completed a remarkable "bottom formation." After dropping from a peak of 45K monthly active developers in 2022 to approximately 23K, this apparent halving on paper sparked discussions on social media about "narrative exhaustion." However, when we dissect the cross-section of this curve, what we see is not the industry's contraction, but a profound "de-leveraging of talent."

▲ Source: Electric Capital Developer Report, based on Crypto Ecosystems Github

Who left? Who stayed?

Mostly newcomers left. In February 2024, the number of new developers in a single month reached a peak of 5,462, but subsequently plummeted, with a churn rate of 52% for those with less than a year in the industry. This cohort mostly flooded in during the bull market, working on NFT minting contracts, forking DeFi protocols, or building frontends for new L2s.

These roles are highly dependent on market hype. Once the hype fades, projects stop operating, and the positions disappear. Data shows that newcomers' code contributions never exceeded 25% of the total; from the start, they were never in the core circle of the industry.

▲ Newcomers surge in bull markets and exit in bear markets; established developers (2+ years of experience) hit an all-time high during the same period.

Source: Electric Capital Developer Report

On the other hand, developers with over two years of experience not only didn't decline during the same period but actually increased, hitting a new all-time high and contributing approximately 70% of the code volume. Maria Shen, GP at Electric Capital, put it bluntly: "When we look at the established developer group, it's growing, and it looks very healthy."

They stay, not because they have no other options.

Technologically, the core work in crypto now involves developing fundamental infrastructure that usually requires years of accumulation to understand: protocol development, security audits, cross-chain architecture. These tasks need years of groundwork and can't be dismissed by the market simply because the hype is gone.

Economically, many veterans hold unvested tokens, governance power, and equity in protocols. Their accumulated stake in the industry has created tangible barriers and returns.

Looking at ecosystem distribution, they are voting with their feet: Bitcoin developers grew by 64.3% over two years, Solana by +11.1%, while Cosmos saw a 51.1% decline and Polkadot a 46.9% decline. Veterans are consolidating into ecosystems with real users and revenue, leaving projects that still rely on narrative.

▲ Source: Coincub Web3 Jobs Report 2025

Data Source: Web3.Career

The changing structure of job positions confirms the same trend. In 2025, the highest proportion of new Web3 job postings was not for developers, but for Project & Programme Management, accounting for over 27%.

This might seem counterintuitive for an industry known for being technology-driven, but the logic behind it isn't complicated: the industry is moving from a construction phase to an execution phase. Over 100 chains need integration; institutional clients entering the space have completely different requirements for compliance and security; and DAO governance needs to find a balance among stakeholders with diverse interests.

This is not project management in the traditional sense, but coordination and judgment within an environment where the rules are still being formed.

On the surface, the industry is shrinking, but its core density is actually increasing. The bear market of 2018-2019 also saw a massive exodus of developers, yet it was followed by the emergence of phenomenal projects like Uniswap, Aave, and OpenSea, which defined the 2020-2021 bull run. The builders left this time around have more mature infrastructure, and the AI era presents them with an even bigger stage than the last cycle.

What capabilities do those who stayed possess?

What special abilities does the crypto industry precisely cultivate in its builders? To answer this, we need to return to the foundational principles of blockchain. Amidst the alternating cycles of bull and bear markets, this industry has always operated under the same basic rule: Code is Law, Execution is Final.

In the 2016 The DAO incident, an attacker exploited a recursive call vulnerability to steal $36 million. The code had no bugs; the logic executed exactly as intended, only the boundary conditions weren't anticipated by the designers. In 2021, the Poly Network cross-chain bridge was attacked, and $610 million was transferred within hours.

No platform could stop it, no institution could reverse it, and no legal clause could pursue recourse. This is crypto's structural feature that distinguishes it from almost all other industries: zero margin for error, and post-facto intervention is virtually non-existent.

This environment forces the development of a capability rarely needed in other industries: the ability to build a functioning system from scratch, one that strangers are willing to participate in, under conditions of absent rules and trust.

This capability comprises two levels. First, building trust from nothing, relying not on any external authority but solely on code and mechanisms to make strangers willing to deposit real assets. Second, making judgments under both technological and economic uncertainty, designing a working system without a regulatory framework, historical data, or industry standards for reference.

Both levels have concrete validation in crypto. Uniswap, with no corporate guarantee, no KYC, no customer support, achieves hundreds of billions of dollars in daily trading volume based solely on trust in a few hundred lines of code and an economic mechanism.

MakerDAO, with no central bank backing and no deposit insurance, maintains the stability of DAI purely through on-chain governance and collateral mechanisms.

The DeFi Summer period was even more extreme. Without a regulatory framework, audit standards, or historical data, builders designed AMMs, lending protocols, and yield farming, going from concept to tens of billions in TVL in just a few months. This capability manifests differently in builders focused on protocol, application, and governance layers, but the underlying principle is the same.

The AI era is creating a structurally similar problem. Model decision-making processes are opaque, and output results cannot be independently verified. AI agents are starting to execute trades and allocate funds autonomously, but the supporting rule systems and constraint mechanisms do not yet exist.

Large model companies control both the models and the evaluation standards, leaving users without effective verification means. Computing power is highly concentrated among a few top tech giants, creating monopolistic pricing when demand surges. These issues all point to the same core: the trust problem for autonomous systems, replayed on the larger scale of AI.

Crypto builders have been dealing with these types of problems for years, operating in environments without external authoritative constraints. Previously, the context was on-chain protocols; now it's AI. And there's already a cohort of people who have directly transferred their crypto-acquired capabilities into AI and produced results.

How are these capabilities being re-priced in the AI era?

The shift from crypto to AI is not uncommon in recent years, but upon closer examination, what they bring with them differs.

The most straightforward path is the direct transfer of hardware and experience. The three founders of CoreWeave, Michael Intrator, Brian Venturo, and Brannin McBee, started mining Ethereum with GPUs in 2017, scaling from one machine to thousands. They shut down the mining operation in 2022. Two months later, ChatGPT launched, and their GPUs were directly converted into AI compute supply. CoreWeave went public on the Nasdaq in March 2025 with an IPO valuation of around $23 billion, and its market cap peaked near $70 billion.

OpenSea co-founder Alex Atallah handled the aggregation and routing of highly heterogeneous assets in the NFT marketplace. He applied the same expertise to AI model routing, founding OpenRouter, which served over 5 million developers within two years and achieved a valuation of $500 million.

Another type of migration is more noteworthy. NEAR founder Illia Polosukhin is a co-author of the Transformer paper. After leaving Google, he initially aimed to build AI applications using natural language but encountered a practical problem: the need to make cross-border payments to data labeling workers worldwide, many of whom didn't have bank accounts. Blockchain technology became the optimal solution for this payment challenge.

Now, NEAR is pivoting to become an AI infrastructure platform, focusing on user-owned AI and decentralized confidential machine learning (DCML), allowing users to utilize AI services without exposing their data. The experience gained from building NEAR's decentralized architecture has become the hardest-to-replicate starting point in this direction.

Circle co-founder Sean Neville left to found Catena Labs, positioning it as an AI-native bank. He directly transferred his understanding of stablecoin infrastructure to the AI agent financial scenario, with a16z crypto leading an $18 million seed round.

Nader Dabit, a senior developer for Aave and Lens Protocol, moved to Cognition, bringing his extensive experience in building developer ecosystems across multiple crypto protocols into the AI agent tooling space.

What this cohort takes with them isn't just GPU hardware or user networks, but an intuition for mechanism design, experience in building developer ecosystems, and the judgment to construct trusted systems from scratch when rules are absent. These capabilities precisely correspond to three structural gaps encountered during the scaling of AI.

Aggregation and Optimization of Computing Power

Computing power is the most direct bottleneck for AI scaling. Training and inference require massive amounts of GPUs, demand is volatile, cloud providers are expensive and have queues, and enterprises don't want to hoard hardware themselves. This problem has two aspects: how to aggregate and allocate compute power, and how to use aggregated compute more efficiently. Crypto builders have directly transferable experience in both areas.

Hyperbolic tackles the issues of distribution and trust. Founder Jasper Zhang brought decentralized mechanism design to the AI compute track: tokens encourage dispersed GPU holders to contribute idle computing power, but the more core problem is trust.

Why should you trust that a computation result from an unknown node is correct? Their core innovation, PoSP (Proof of Sampling and Punishment), uses random sampling and game theory to make honesty the dominant strategy for nodes. It requires no full verification, is low-overhead, scalable, and produces reliable results. This mechanism is directly migrated from crypto's logic of verifying the behavior of untrusted nodes.

MoonMath addresses the efficiency problem. Its predecessor, Ingonyama, focused on ZK hardware acceleration, significantly boosting ZK proof generation speed under extreme computational constraints.

Now it has pivoted to a Physical AI performance layer, working on sparse attention acceleration for video diffusion models (LiteAttention), low-rank decomposition of FFN layers (LiteLinear), and training backpropagation acceleration (BackLite). From ZK acceleration to AI inference acceleration, the underlying skill is the same: making math run faster under extreme computational constraints. The track changed, but the accumulated expertise was not wasted.

AI Governance and Incentive Mechanism Design

When multiple AI agents start collaborating to execute tasks, how do you ensure they don't break the overall system while pursuing their individual goals? Each participant is optimizing its own objective function, with no guarantee that the sum of their actions will keep the system functioning properly, and agents operate at speeds far exceeding the window for human intervention.

This is precisely the type of problem crypto builders have repeatedly tackled in DAO governance and tokenomics design: getting participants with completely divergent interests to behave according to a system's pre-set direction without a central authority. Crypto's answer was an economic mechanism – violating rules incurs real economic costs, and the rules are written in code that executes automatically.

EigenLayer has directly migrated this mechanism to the AI context. Through its restaking mechanism, nodes must pledge assets before participating in collaboration; failure to perform or malicious action triggers automatic slashing. The rules are not suggestions but rigid boundaries backed by real economic consequences.

EigenCloud extends this logic to verifiable computation and collaborative governance for AI agents, forcing agents to stay within predefined bounds while pursuing their objectives. Using economic mechanisms to constrain agents is far more reliable than relying on ethical guidelines.

AI Agent Autonomous Payments

There's an even more fundamental problem: how do agents pay? Traditional payment systems are designed for humans. Credit cards require accounts, bank transfers need authorization, and every step assumes the operator is human, has an identity, and will wait. Agents don't wait. They might initiate a massive number of requests per second, each potentially involving a micropayment. Traditional payment rails become completely dysfunctional in this scenario.

Stablecoins and on-chain rules are infrastructure already built by crypto builders. They natively support programmability, permissionless access, and 24/7 operation. These three characteristics are the hard requirements for agent payment scenarios. What's missing is just a protocol layer connecting stablecoins to agent workflows.

x402, launched by Coinbase in May 2025, activates the HTTP 402 status code, embedding stablecoin payments directly into HTTP requests. An agent initiates a request and completes payment simultaneously, without needing an account, and settlement takes about two seconds.

As of April 2026, the x402 protocol has processed over 165 million transactions, with a cumulative volume of approximately $50 million and 69,000 active agents (source: x402 Foundation). Cloudflare, AWS, Stripe, and Anthropic MCP have all integrated it. Agent payment is already a track with real traffic.

These three directions correspond to three structural gaps encountered during AI scaling: aggregation and efficiency of computing power, incentive alignment for multi-agent collaboration, and infrastructure for autonomous payments. These problems have no ready-made answers in traditional software architecture, but the crypto industry has relevant experience dealing with them. The capabilities haven't disappeared; they've just found new scenarios.

Builder's New Role: From Writing Contracts to Setting Rules for AI

The scaling of AI is creating a functional gap that didn't previously exist. It's not a gap in technical talent, but a gap in people who can design trust mechanisms for autonomous systems. As the user shifts from humans to AI, the role of the crypto builder is being redefined.

The table below compares the dimensional changes in the specific functional paradigm shift:

The core difference between the two paradigms isn't the tech stack, but the method of establishing trust and the logic of rule enforcement. In the Pre-AI era, crypto builders dealt with human participants; rules were written into contracts, with zero margin for error, but the system's boundaries were relatively clear.

In the AI-Native era, when the interacting entity becomes an autonomously operating AI agent, the problem to solve is: agent behavior is unpredictable, execution speed far outpaces human intervention windows, and the system's boundaries themselves need to be redefined under greater uncertainty.

The functional positioning of the crypto builder is shifting from "writing secure contracts" to "designing trusted mechanisms for AI autonomous systems."

Hiring at major institutions already reflects this change:

▲ Q1 2026 core AI/data positions actively opened by top-tier trading platforms

Source: Gate Research Institute

The hiring landscape at top trading platforms and institutions in 2026 clearly reflects this trend: They are no longer just hiring AI engineers or crypto developers, but seeking individuals who can bridge the two domains – understanding on-chain incentive distortions and governance game theory, while also being able to deeply embed AI tools into crypto workflows and design mechanisms for long-term alignment between agents, regulators, and users.

Capital allocation strategies already reflect this assessment. Paradigm is raising a new fund of up to $1.5 billion, expanding its investment scope from crypto to AI and robotics.

Haun Ventures completed a $1 billion Fund II, focusing on financial infrastructure at the intersection of crypto and AI, specifically payment, stablecoin, and agent-to-agent economic systems that support autonomous AI agent transactions and coordination.

a16z crypto completed its $2.2 billion fifth fund (Crypto Fund V), explicitly stating it will be 100% invested in the crypto space. Facing the complexity and opacity of the AI era, they will focus on application directions leveraging crypto's transparency, verifiability, and decentralization.

Furthermore, according to PitchBook data, in 2025, approximately 40% of VC investment in the US crypto space flowed into companies also involved in AI businesses, a significant increase from 2024.

For crypto builders pivoting to AI, the chosen paths show significant differences depending on the market environment.

In the US, with the regulatory environment becoming relatively clearer, protocol-level innovation has gained real room to survive. The density of capital networks is high, the path from idea to funding is short, and there is a larger margin for error.

Common characteristics of projects like Hyperbolic, EigenCloud, Gensyn, and