Author: Joe Zhou, Foresight News
I’ve been a headhunter in Web3 for four years: I’ve witnessed the most insane talent rush ever
I've been a headhunter in Web3 for four years.
At its peak, a single Web3 company burned through over $20 million in our hiring budget in one quarter. Fresh graduates could earn a million-dollar salary after just six months in the industry. Project teams would raise funding in the morning and start hiring thirty people all at once in the afternoon.
At that time, the entire industry felt like a runaway printing press—salaries doubled, no office required, global remote work, token incentives... everyone believed Web3 money would keep flowing forever.
Even I’m starting to think this industry might just be疯狂 forever.
But it suddenly cooled down starting in the second half of 2025.
Funding news is becoming increasingly scarce, and project teams are gradually cutting back on hiring roles one round after another; many bosses who previously expanded aggressively are now slowly stopping job postings altogether.
Some people have started traveling. Some companies have even vanished entirely. Headhunters are the industry’s thermometer—every boom and bust across all sectors is first and most authentically reflected in hiring demand.
This is what I’ve witnessed over the past four years as a Web3 headhunter: the madness, the bubbles, the cooling down, and the restructuring.
At its wildest, a company burned through $20 million.
Four years ago, I started working as a Web3 headhunter.
Back then, I didn’t understand anything—not even what blockchain was. I just knew the company had suddenly partnered with a “tech-oriented project team,” who were so secretive they barely dared to mention their name. In the end, we helped them hire nearly 20 people and collected a headhunting fee of 3 million.
You read that right: a company hired two to thirty people in a single quarter, paying over 3 million in recruitment agency fees.
If calculated at a headhunter fee rate of 25%, the combined annual salaries of these individuals exceed 12 million. Adding benefits, bonuses, and other costs, the company spends at least 20 million annually just on the people hired through a single headhunter.
And that year, there were many similar stories.
At that time, in Web3, nearly every project was racing to expand. Many companies, after raising funds, didn’t focus on building their product first—they prioritized assembling their team. Whether the product could succeed wasn’t the immediate concern; securing talent was.
One instance that stands out to me was helping a top-tier crypto company place a recent graduate.
To be honest, I was hesitant at first—the client wanted someone younger, but not a recent graduate. However, this candidate’s background was truly impressive: graduated from a top-three university and had conducted blockchain-related research during college. After learning about the company, he showed great enthusiasm, believing the project had real value and offering unique insights into the industry—which is rare.
We usually don’t recommend fresh graduates since most lack practical experience. But I decided to give it a try and referred him to a leading crypto company. To my surprise, they offered him an annual salary close to a million.
He had other offers, but this one clearly offered more. That year, the market was generally strong, and by year-end, with performance bonuses, salary increases, and year-end bonuses, he had earned over a million yuan in less than a year—specifically, within six months.
Are you saying this is ability? Luck? Or a gift of the times? I think it’s all three.
Those who sense the temperature of the industry
For a long time, the first thing I did each day was to check funding news.
From 2022 to the first half of 2023, Web3 saw as many as dozens of funding announcements daily, often amounting to tens of millions of dollars.
It was impossible for us to contact every company, so our logic became simple: focus only on the projects with the highest funding, because they had the most money and were the most willing to expand.
During those years, the entire industry was in a state of extreme euphoria. As soon as a project raised funding, job openings would immediately appear. Some teams didn’t even have a complete product—just a PowerPoint presentation. Yet, that didn’t stop them from opening dozens of HC positions all at once.
That stage was also when headhunters made the most money, as many projects went from 0 to 1 and fiercely competed for talent in China.
Many people realized for the first time that it's possible to earn far more than they ever made at big tech companies without working a traditional 9-to-5 job.
Even today, top exchanges continue to hire steadily. But the logic has changed completely from before. Previously, it was about expansion. Now, it’s mostly about normal business iteration—hiring only occurs when there are new initiatives.
And the hiring requirements are clearly higher. They want new hires to be better than the existing employees. Overall, the number of staff at top-tier exchanges has already stabilized.
What’s truly interesting is another category of companies. Several top-tier exchanges have been consistently laying off staff, with very low employee retention—few people stay for more than a year.
Companies like this hire year-round. As one group leaves, another is brought on.
So often, the热度 of hiring itself doesn't actually indicate that the industry is truly thriving—sometimes it's just due to excessive liquidity.
We feel the industry warming up first. We’re also among the first to know when it cools down.
Starting in the second half of 2024, funding news began to decline noticeably, dropping from dozens of stories per day to just a handful. Many project founders no longer post updates about their business progress on social media—instead, they share vacation photos. Several founders I previously worked with raised funding in the tens of millions of dollars, but recently, they’ve all been traveling.
I could tell right away. The company has most likely given up on moving forward. Many initiatives seemed promising at first, with good stories to tell, but once they entered the compliance phase, it became clear nothing could actually be implemented.
For headhunters, this change will be more direct, because when companies stop expanding, hiring demands are the first to disappear.
Hong Kong Web3 Carnival: I've given up completely
In April 2026, I attended the Hong Kong Web3 Carnival, one of Asia’s largest Web3 conferences.
To some extent, it has always been a mirror of market sentiment.
Before going, I was still holding onto the mindset of “one more充值 of faith.” But on the first day inside, I started to lose it—everyone was running a pyramid scheme.
In previous years, attendees were mostly legitimate project teams, technical developers, and frontline investors, who came together to discuss decentralized innovation, global financial transformation, and industry trends. This time, many booths couldn’t even clearly explain what they were doing—some didn’t even have a product.
At that moment, I truly realized: this industry is going through a bear market.
Previously, people talked about transforming finance, decentralization, and globalization. Now, more people are discussing compliance, cash flow, and how to survive. Many project teams that once aggressively expanded have drastically reduced job postings. Some companies have even entered a semi-halted state.
I will still continue working in Web3. But at least this year, I won’t go all-in on Web3 like I used to. Because based on hiring demand, this industry has clearly entered a mature phase.
Of course, bear markets may also harbor significant opportunities.
There is still growth in the areas I’m following with hiring demand: top-tier exchanges, RWA tokenization, AI Crypto, DeFi infrastructure, and the stablecoin ecosystem.
Especially some medium and small exchanges. When the ecosystem was thriving, you barely noticed them. Now that the industry has cooled down,反复翻查，反而还是这几家公司一直在持续招聘。
But over time, we encountered fewer and fewer clients of this type, because many small exchanges are essentially startups—lacking standardized processes and poor in stability. Some even operate in gray areas. After all, we have a responsibility to our candidates. Recommending people to inherently unstable platforms carries significant risk.
More realistically, headhunters’ referrals have an expiration date. If a client’s process is disorganized and turnover is too high, you’re likely to end up doing all that work for nothing. I previously partnered with a well-ranked exchange, and their hiring process was extremely chaotic.
Since then, I’ve pretty much avoided companies like this.
Turn around and look at AI: The frenzy of a new opportunity, but with a changed logic
Web3 has cooled down, but I can't sit idle. The company has performance targets, and the year passes quickly. So starting this year, I've also begun looking into AI.
To be honest, I initially thought AI hiring was similar to Web3—both were startups going from 0 to 1, both were fast-paced, and both were competing for talent. But once I actually got into it, I realized the differences are quite significant.
First, the AI sector is highly homogeneous. Our team primarily works with startup projects, not AI divisions of large internet corporations. While AI startups available on the market appear diverse in direction, their underlying logic is surprisingly similar.
Over the past few weeks, I’ve closely followed seven or eight projects and noticed a clear trend: the direction of pure general-purpose large model agents is no longer viable. The companies that were meant to succeed in this space were already established last year, or even the year before. Now, if you’re launching a purely software-based, purely agent-driven project, investors are no longer interested.
So, the AI wave has quietly reached the hardware sector.
I currently have partnerships with three to four companies, all focused on AI-powered smart hardware—not traditional smartwatches or headphones, but devices that use hardware as an interaction gateway, with an Agent working behind the scenes. For example: you’re wearing a smartwatch, a piece of jewelry, or even headphones, and you say, “Book me a flight to Shanghai tomorrow,” and it automatically handles it; or you say, “I need to build something—set up the backend for me,” and it executes the task. The core innovation isn’t the hardware itself, but rather that the hardware becomes a “shell” for an Agent.
I’ve recently seen smart watches, smart earbuds, and devices designed to be worn like brooches. Their functions are largely the same—you interact via voice, and they leverage underlying models and APIs to complete tasks for you. Do they have fundamental differences? No. The level of homogenization remains high, but they’re all betting on one thing: whoever gets out first will claim the next entry point.
I asked a client, "Why did you stop focusing solely on agents?" He said, "With pure software, no matter how well you do it, you're ultimately just building something for the big players. They can easily wipe you out with a single API. But if you have hardware that users wear every day, that’s an entirely different story." Hardware means stickiness, means barriers, means it’s harder to replace.
Of course, AI hardware companies aren't entirely immune to software talent— their hiring needs still include roles like Agent development and backend development. But you'll notice these positions are no longer about "building an app," but rather about "supporting a hardware product." Moreover, the requirements these companies have for candidates are completely different from those in the early days of Web3. AI hardware is the opportunity I currently see that most resembles the early wave of Web3: it's highly homogeneous, intensely competitive, and difficult to execute—but precisely because of this, it indicates the space hasn't been finalized yet and still has room to grow.
Recruiting for startup teams taught me one thing: no matter how the industry changes, the logic of building a 0-to-1 startup team remains the same. If you can tell whether a project is “just funded and urgently needs to build a team,” you can make money. I’ve honed this skill over four years in Web3.
AI vs Web3: Which is better to work on?
You ask me, which is easier to do now: AI or Web3 back then? Let me give you two numbers first: 1 million and 3 million.
A few years ago, a top-tier Web3 undergraduate graduate could earn an annual salary of 1 million. At that time, we thought this was already a very high number—an undergraduate with no work experience, simply because of the blockchain boom, leading companies were willing to spend big. Now? A top-tier AI PhD graduate (and very few master’s graduates) can reach an annual salary of 3 million. But the requirements are much higher: young, high-potential, ACM awards, publications in top-tier journals and conferences—these are almost mandatory.
Back in 2021, when Web3 was just getting started, all you needed was someone who could use it—that was definitely a time when headhunters could close deals easily and make good money. But now, some Web3 startups have much higher requirements—there are educational thresholds, demands for technical depth, and many prefer candidates with backgrounds from major internet companies. The standards keep rising.
Let’s talk about recruitment fees again. In the early days of Web3 (such as 2021), the industry was not widely accepted, making it extremely difficult to hire talent. Candidates were skeptical, fearing that overseas opportunities were scams or posed safety risks. As a result, recruitment agency fees were typically at least 25%, and sometimes reached 28% overseas—for example, on a $1 million annual salary, the agency would take $250,000. Later, as demand surged, more agencies entered the market, leading to price competition. From 2022 to 2024, most projects offered fees in the range of 22–25%. After that, demand sharply declined, and top exchanges pushed rates even lower, with some accepting 20% or even 21%.
What about AI companies? The fees seem similar, but there’s a fundamental difference: AI companies screen the qualifications of recruiters. Major internet firms and reputable AI startups vet their vendors, giving established recruitment agencies an advantage. AI companies don’t indiscriminately onboard numerous recruiters, and fees aren’t driven down by cutthroat competition.
But Web3 is different. Many project teams don’t screen headhunters carefully, allowing solo headhunters to enter the space—even some who work for just 15%. Many Web3 companies don’t even have proper employment contracts, let alone headhunter agreements. This leads to cutthroat competition—low fees, poor service, and ultimately, no one makes much money.
In early Web3, high-paying jobs were much easier to close, with significantly higher conversion rates.
The AI market currently has more high-paying positions and larger deals, but competition is much fiercer and the order conversion rate is lower. On the Web3 side, deals may close a bit faster, but while AI offers higher salaries and commissions, there are simply too many people competing for them. The challenges in both industries are quite different and cannot be directly compared—it’s essential to consider the specific context.
Now I’m watching both tracks. Each track has its own madness— one is mad in plain sight, the other is mad in the shadows.
No regrets for entering the game—now I understand the cycles better.
You ask me: If I could go back four years, would I still enter Web3?
My answer is still the same: yes.
I’ve asked myself this question many times, and the answer has never changed. Web3 gave me my first real close-up look at how an industry moves from frenzy to calm. I learned how to tell whether a company’s needs are genuine or not, how to help a 0-to-1 team build up as quickly as possible, and how to pivot promptly when the industry heats up or cools down.
More importantly, this industry is always changing. NFTs, GameFi, Web3 social, RWA, AI Crypto, stablecoins... During those years, I was almost constantly in a state of rapid learning. Later, I realized that this ability is actually very valuable. Whether it’s AI, going global, or other entrepreneurial paths, many underlying principles are the same.
More realistically, I have indeed earned commissions from it, accumulated some general professional skills and industry intuition, and helped many candidates find jobs and lifestyles they truly enjoy.
I've seen many people whose lives have been transformed by Web3—some received their first salary in a currency other than RMB and moved abroad to pursue opportunities; others began working remotely long-term and embraced a digital nomad lifestyle.
So you ask me if I regret it? I don’t. Not everyone gets the chance to witness firsthand the most passionate and upward moment of an industry.
The hot trends always rotate, and the tide of commerce never stops. But I’ve gradually come to understand the cycles: all frenzies eventually cool down, and all bubbles eventually burst. The era of frantic talent hunting is over, but the long-term era for those who dig deep has just begun.
Twitter: https://twitter.com/BitpushNewsCN
BitPush Telegram community: https://t.me/BitPushCommunity
BitPush TG subscription: https://t.me/bitpush