The anxiety of Binance: Cryptocurrency struggles to retain innovators
Author: Gu Yu, ChainCatcher
In the past few years, what crypto giants have lacked the least is money. After experiencing the collapse of FTX, tightening regulations, and several rounds of cyclical reshuffling, the crypto giants that have truly survived can hardly explain their predicaments with "lack of funds"—they have a large user base, mature infrastructure, and sufficient cash reserves to continue acquiring and investing.
On October 6, Binance co-founder He Yi reiterated in a Chinese AMA during a smart product launch that: "Binance's biggest challenge is talent; although not everyone will be hired, and some may not pass the interview, we are indeed eager for talent."
This statement is not new. In December 2025, He Yi, who had just taken on the role of co-CEO, gave almost the same answer: the biggest challenge Binance faces is still "talent density." What it needs are people who understand financial products, know crypto, and believe in this industry. Ten months later, the same issue still ranks first at a top global crypto company.
This can hardly be understood as a recruitment issue for a single company; it resembles a structural blood loss that the entire industry is experiencing.
1. What Binance lacks is not people, but "talent density"
Talent density is never just about the number of employees. A company can have thousands of employees but still lack those who can truly define products and solve complex problems. Especially in an industry like crypto, what is scarce is not "people who can write code," but those who can judge "what should be done next" when rules have not yet formed and demands have not yet been validated. He Yi's example is very specific: when encountering a good candidate, the first reaction is to ask, "Do you want to move to the UAE?" The candidate might respond, "What do you want to bring me to the desert for?"—even if the salary can cover market prices, geographical location and industry stigma remain additional hurdles.
This is precisely the most noteworthy aspect of Binance's expansion: its strongest competitive advantages have always concentrated on centralized trading, matching, liquidity, user distribution, and risk control, and these capabilities can continuously bring new assets into the existing pipeline.
Tokenized stocks are a typical example—bStocks launched in June 2026, with AUM growing from $5.6 million on the first day to over $500 million by the end of July; as of July, 41.5% of bStocks users had never used Binance's stocks or related derivatives before, while 58.5% of early users were also using Perps or direct stock products. Much of the growth is built upon Binance's existing account system, users, stablecoin balances, and trading infrastructure.
When products are built on Binance's strongest capabilities, scale advantages can quickly translate into growth advantages; however, when the business enters areas like chains, wallets, DEX, and AI that require redefining product forms, the existing scale does not automatically translate into product advantages.
Wallets are the most direct example. By the end of 2024, He Yi publicly admitted that Binance's wallet "did not perform particularly well" over the past year and decided to rebuild from the ground up; in 2025, she acknowledged that Binance's wallet still lags behind industry leaders. Binance does not lack resources—hundreds of millions of users, traffic entry points, and the wallet could easily be integrated into the app, but the core of a centralized exchange is matching, liquidity, risk control, and scaled operations, while the core of a wallet is self-custody experience, on-chain interaction, and developer ecosystem; the former excels at maximizing mature demand, while the latter requires the team to continuously discover immature demands.
Similarly, Binance's layout in the Layer 1 track has repeatedly faced setbacks, and its investments and layouts in perpetual contracts and prediction markets have significantly lagged behind competitors, suffering major strategic setbacks.
This gives "talent density," which He Yi repeatedly emphasizes, a more specific meaning.
2. AI is redefining talent pricing
Why is it increasingly difficult to recruit talent? Because one of the biggest competitors facing crypto today has become AI.
Tiger Research's statistics on crypto industry recruitment data for the first half of 2026 show that the proportion of positions requiring AI skills has risen from 23% at the beginning of 2025 to 53.1% by March 2026; PwC's AI Jobs Barometer indicates that workers with AI skills earn an average salary 56% higher than their peers; data from CryptoJobsList shows that the average annual salary for mid-level AI-related positions is $115,000, compared to $95,000 for non-AI positions, a premium of 21.1%. In just 2025, global corporate investment in AI reached approximately $581.7 billion, with private investment in generative AI growing over 200%—AI is no longer a trend waiting to be validated but is becoming the new infrastructure of the entire tech industry.
This has changed the career choices of top talent. In the past, crypto could compensate for industry risks with high salaries, tokens, and entrepreneurial opportunities; today, AI also offers a vast wealth of imaginative possibilities but possesses stronger industry narratives and social prestige. For a young engineer, the question becomes: if both are entrepreneurial, why create a new DeFi protocol instead of an AI agent? Choosing AI is not just because the salary may be higher, but because it could give rise to the next Google or OpenAI; even if it fails, that experience is still an asset for the next job. What crypto needs to answer is another question: "Why should you bet your next five years here?"
This migration has already occurred among mature practitioners. The founding team of CoreWeave started with Ethereum mining and transitioned to AI computing power after shutting down their mining business in 2022; Circle co-founder Sean Neville founded Catena Labs, bringing stablecoin experience into AI finance; Sam Blackshear, creator of the Move language and co-founder of Mysten Labs, left the Sui ecosystem to join Anthropic in August 2026; and NEAR founder Illia Polosukhin is one of the co-authors of the Transformer paper. What these individuals take away is not just positions but also protocol design, system architecture, financial products, and entrepreneurial experience.
Jocy Lin, founder of IOSG Ventures, warned as early as 2025 that crypto's focus on short-term memes and speculation is causing core developers to shift toward AI; more critically, once these individuals achieve success in the AI industry, the likelihood of them returning to crypto will decrease. AI is not just taking away crypto's workforce; it is taking away the next generation of high-level talent. This has even formed a one-way flow: people nurtured by crypto are helping AI solve its next stage problems, while AI rarely reciprocates by sending back equally scaled new high-level talent to crypto.
3. What is truly lost is not developers, but innovators
Data from Electric Capital reveals a more subtle change. The number of active developers in crypto has fallen from a peak of about 45,000 in 2022 to about 23,000, but the loss is primarily concentrated among newcomers: the turnover rate for developers with less than a year in the industry is as high as 52%, while the number of established developers with over two years in the field has reached a new high, contributing about 70% of the code. Veterans remain, but newcomers are becoming increasingly scarce, while new high-level talent is concentrating on AI and a few mature ecosystems—Bitcoin developers have grown by 64.3% over two years, Solana by 11.1%, while Cosmos and Polkadot have decreased by 51.1% and 46.9%, respectively, with talent flowing from narrative-dependent ecosystems to those with real users and revenue.
The job market is also contracting in sync, and the contraction points in the same direction. Dragonfly's 2026 talent report shows that the crypto industry added about 3,700 new positions in 2025, but approximately 4,100 were canceled, resulting in a net decrease of 472 for the year; companies are no longer engaging in widespread expansion typical of bull markets but are focusing on hiring for positions that truly impact the business—compliance positions grew by 340%, data science by 74%, while marketing, design, and customer service saw significant contractions, with application volumes dropping by about 26% in the second half of the year.
In other words, companies are selecting people, and people are selecting companies. In the past, a hot project, a high salary, or a token incentive could complete recruitment; now, top candidates are more concerned with three questions: What problem is this company really trying to solve? What can I do after I join? Will it still exist in three years? Dragonfly's recruitment observations also found that the main factors influencing candidates to decline opportunities are not salary, company size, or funding stage, but the "uncertainty" of the position and the company—if a company cannot clearly articulate why this job is important and why it can be sustained, the talent conversion rate will significantly decline.
This is especially dangerous for the industry. In past cycles, bear markets would eliminate a large number of practitioners, but new entrepreneurs would bring products like Uniswap, Aave, and Hyperliquid in the next round; this round is different in that the departing newcomers may not just be temporarily exiting; they are likely entering AI directly, completing entrepreneurship, accumulating capital, and gaining social prestige in another industry.
So the question becomes: who will create the next generation of crypto products when the next cycle arrives? If the industry can only rely on existing projects for iteration, lacking new entrepreneurs and technological paradigms, crypto will transform from an industry that continuously creates new categories into one that competes for existing users and traffic.
This is also a layer of the talent issue that is most easily underestimated: the lack of ordinary employees can be solved with higher salaries, the lack of skilled engineers can be supplemented through recruitment, outsourcing, or AI tools; but the lack of those who can pose new questions, establish new products, and redefine user needs is something that capital finds hard to fill. Innovation does not require more people, but rather the judgment of a few individuals to continuously break through existing paths.
4. The decline in talent density will ultimately lead to path dependence
The talent issue will ultimately feedback into products. Hyperliquid is an extreme positive case: it established a perpetual contract market capable of competing with large trading platforms with a core team of just over a dozen people—monthly transactions exceeded $180 billion at one point in 2025, with annual fee revenue of about $961 million. This indicates that in a complex and rapidly changing industry like crypto, team size is not the most critical metric; rather, the product complexity and organizational efficiency that each unit of talent can create are what matter.
Conversely, looking at Binance presents another path. This does not mean that Binance "cannot create on-chain products." What is more worth discussing is that an organization centered on CEX, when entering the on-chain world, requires a talent structure that is precisely different from its existing advantages: centralized exchanges pursue stable systems, mature risk control, and certain business outcomes; on-chain products, however, require teams to accept greater uncertainty in the early stages, understand the real behaviors of developers, communities, and traders, and be willing to experiment over a long period.
When a company achieves great success through centralized trading, liquidity, and global distribution, the easiest thing for it to do is to replicate these advantages in new businesses—connecting to exchange traffic where possible, relying on scale where user numbers can solve issues, and rarely reinventing from the ground up when distribution can solve problems. Tokenized stocks are particularly suited to this capability structure, and thus Binance quickly found its growth path, with Binance Research attributing this to: "Distribution channels are replacing issuance as the growth driver."
However, the truly difficult issues—encouraging users to form new trading habits on-chain, redesigning wallet experiences, and establishing the next generation of trading mechanisms—require a different density of talent.
The industry has not been unresponsive. On the supply side, Binance launched the Pioneer Talent Program and Accelerator Program aimed at young talent; on the demand side, another pricing logic has emerged.
Morpho CEO Paul Frambot candidly mentioned during discussions about mergers and acquisitions that the team has received numerous opportunities to acquire crypto companies over the past six months, but Morpho **"rarely acquires companies for intellectual property or technology," preferring to recruit the right people to develop end-to-end solutions themselves. He established a valuation principle for such acquisitions: "Valuation will be purely based on the comparison of the talent acquired with the cost of hiring an equivalent team."—treating "talent acquisition" as a substitute for recruitment rather than dreaming of it as a strategic transaction.
Both point to the same issue: what is truly scarce is no longer code, but the people who can decide where the code should be written. Whether these efforts will be effective depends on a more fundamental change: Can Crypto become an industry worth investing a decade of one's career for excellent young talent?
Otherwise, the industry will enter a rather pessimistic cycle: decrease in senior talent → decrease in new products → slowdown in user growth → decline in industry attractiveness → fewer senior talents willing to enter. Jocy Lin's concern about the "negative cycle" is essentially this process, and the data on the user side is already reflecting the latter half of it—in August 2026, the number of global monthly active on-chain addresses decreased by 18% year-on-year, while passive holders increased by 16%, indicating that users are still present, but they are no longer active.
In the past, Crypto competed with the internet and finance for talent; now, it must compete with AI for the most capable, ambitious individuals in the entire tech industry who also believe they can change the world. This is no longer just a matter of "is the salary high enough?"
Because money can only compensate for risk, but cannot create meaning.
A truly outstanding talent is often willing to join an industry not because the salary is 20% higher, but because they believe there are still significant problems, enough new opportunities, and worthwhile challenges that merit investing a few years to solve.
This may be what Crypto currently lacks the most. There is already plenty of money, but the number of people willing to use that money to create the next generation of Crypto is dwindling.
And a true crisis in an industry is never about a lack of money or users, but rather an increasing scarcity of outstanding talent who believe they can still create the future here.