"AI is the main reason for our layoffs." The article says that line has become a standard explanation across corporate cuts. In the first half of 2026, the U.S. tech sector shed nearly 140,000 jobs, according to the source text. Amazon cut 9% of its workforce, and Meta cut 10%, both citing AI-driven change and a need to streamline.
The same article adds another layer: more than 56% of layoff events in 2026 explicitly listed AI, automation, or machine learning as a reason, and AI has been the top stated driver of U.S. corporate layoffs for four straight months. At the same time, nearly 60% of companies acknowledged that they framed layoffs or hiring slowdowns as AI-driven when the underlying reason was financial pressure.
That shock has reached beyond Silicon Valley. The report argues that Web 3.0, sitting between technology and finance, has been hit especially hard. Large-scale layoffs in the sector have lasted for more than half a year and have unfolded with unusual speed. Since the start of this year, especially in recent months, news of staff cuts, team reshuffles, and talent movement around major trading platforms has circulated heavily on X, Reddit, Xiaohongshu, Maimai, and in private coffee chats among workers. BitMEX, once highly visible, has nearly faded from the mainstream view, while smaller platforms are exiting or trimming business lines.
Layoff notices came late, decisions were made much earlier
Kevin said he had only three days between receiving notice and his last day. He had spent several years at a large internet company before moving into Web3, drawn by higher pay and the sector’s narrative. Only later did he learn that his departure had effectively been decided more than a month earlier.
He said there had been almost no clear warning signs. Work continued as normal, meetings stayed on the calendar, and messages still got replies. Then HR contacted him. There was no convincing explanation, no case built around poor performance, but the layoff still landed.
Looking back, Kevin said the only real hint may have been that two people had already left his 10-person team before he did. At the time, the internal explanation was that they were “not a fit” or had gone to find easier jobs. He told Zhangsheng BeatZ that, in retrospect, management may already have been pushing them out then.
Richard described a more severe process at a smaller crypto exchange. After nine years as a full-time stay-at-home father, he returned to the workforce and found a job at a not particularly large crypto platform. He and many colleagues were cut soon after.
One morning, he opened his laptop and found that his system access had been shut off. At first he thought it was a technical problem. Then he opened the group chat and saw about 40 colleagues asking the same question: why they could no longer log in. No one knew what had happened. Hours later, they received cold termination emails in their personal inboxes, effective immediately.
What hit him harder, he said, was what happened shortly before the layoff. A manager had hinted that one developer on his team “might need adjustment.” Richard tried to help that colleague stay, even rearranging assignments to show the person was not replaceable. Before he could submit the plan, both of them were gone.
Another former exchange employee, identified as Xiaoyu, described a similar scene to Zhangsheng BeatZ. In her previous company, she said, the first step in a layoff was often to disable Slack accounts in batches and cut email access. Whenever someone suddenly disappeared from Slack, people would rush into private chats and send over phone numbers and LinkedIn profiles because no one knew who would be next.
Xiaoyu said that when it was finally her turn, her manager sent a Slack message asking if she had time for a call. Before she could answer, all her access was gone.
Quarterly cuts, locked systems, and severance disputes
Kevin said layoffs kept hitting his old team for months after he left. It is now down to two people. He estimated that the exchange cut around 10% of staff every quarter, reaching a cumulative 40% over a year.
The article says Coinbase announced in May that it would cut about 700 jobs globally, defining the move as an “AI-native reorganization,” or roughly 14% of staff. But Zhangsheng BeatZ, citing a person familiar with the matter, said the impact on Coinbase’s India office was much deeper. One former employee said about 90% of staff there were gone, affecting every business unit rather than sales alone. Only a very small number of engineers seen as top-tier were invited to relocate to Canada and continue working.
According to the report, a main reason for the heavy cuts in India was cost, combined with the time difference with the U.S. Coinbase was said to pay around 7.5 million rupees to an SDE2 engineer in India, equivalent to about C$110,000, roughly in line with mid-level engineering pay in Canada. The article also says architects in India at many high-paying product companies can earn more than peers in the European Union.
The report says workers at several exchanges were unable to reach agreement with HR on severance terms and were then informed that the same day would be their last working day, with system access shut off at once. BitMart, which the article describes as a recently closed trading platform, had entire departments cut starting in May.
Zhangsheng BeatZ also says layoffs often clustered around June 30. The reason, according to the report, was simple: July brings a new set of financial statements that need to be shown to investors. Cutting staff and removing expenses can make the profit-and-loss picture look better immediately. In that framing, layoffs are not just a cost move but part of financial narrative management.
Several interviewees said the process looked much the same across firms. Contact channels and internal permissions were cut so quickly that workers lost even the basic ability to keep in touch with colleagues or organize around their rights. A person cited in the article said operations and product staff, because they were more likely to work in offline or overseas offices, sometimes got a normal handover period and compensation. Many technical workers, by contrast, were remote and could be dismissed quickly with little practical friction for the company.
One problem is jurisdiction. Many IT employees are based in mainland China while the exchanges are registered offshore. The article says that makes enforcement expensive and difficult for individuals. For many people, the amount at stake is not large enough to justify a long fight.
Even where there was a short buffer period, workers did not necessarily fare well. Kevin said HR asked him to select a reason for departure in an internal system and urged him not to choose “dismissed by company.” He said they warned him that if he picked that option, future background checks might fail and negative comments could follow. If he chose to leave for personal reasons instead, the company would not need to pay extra compensation. In the end, Kevin received no severance, only salary and overtime due up to his final day.
Performance systems and monitoring turned layoffs into “compliant” exits
The article argues that, during the wave of cuts, exchanges also looked for ways to make layoffs appear to be something else.
Multiple interviewees told Zhangsheng BeatZ that major exchanges shipped company laptops to employees before onboarding. Those machines, they said, included strict monitoring software that tracked keyboard activity and mouse clicks, with the data feeding into performance reviews.
The report says one exchange employee was allegedly fired the day after using a company laptop to watch a drama on iQIYI for a while.
Another common practice, according to the article, was to assign KPIs that were nearly impossible to complete. Once the review period ended, employees could then be dismissed under labels such as “poor performance” or “not meeting company standards.” In that setup, a layoff was recast as a formal performance elimination, allowing the firm to avoid additional compensation.
The article also cites a former exchange employee on X who said that during one layoff period the company ran recurring tests on “Web 3.0 industry knowledge” and forcibly tied them to KPI reviews. Failing the test could expose workers to direct termination as well.
The piece describes the layoff wave as a tornado: fast, violent, and difficult to discuss openly inside a workplace shaped by long-term monitoring and pressure.
Those who stayed were left in a state of fear
Workers who remained were not necessarily the lucky ones. Xiaoyu said that after each layoff round, survivors sometimes envied colleagues who had already left because at least the uncertainty was over for them. Those who stayed lived in constant anxiety, unsure whether they would be next. She said morale became sharply negative once the cuts started.
Richard described a similar shift. Before the layoffs, work had still been intense and fast, but most of the effort went into actual product updates and feature releases. Later, he said, the workload changed character. More time was spent satisfying management-driven formalities. Attendance controls tightened, and meetings became more frequent.
In his exchange, the daily stand-up culture was pushed to an extreme. The original purpose of stand-ups is to keep meetings short and efficient, but he said the tool turned into a drain. The team held two stand-ups a day, yet no one could clearly say where the product was headed. Three project managers came and went within half a year, and the product management function was eventually hollowed out. Projects were underway, but if a key person was suddenly dismissed in the middle of a day, sometimes minutes before a meeting, work could stop on the spot.
Richard also said there were outsourced teams whose pay was far higher than that of full-time staff. Only later, after speaking face-to-face with two colleagues, did he learn that an executive had withheld employees’ raises for two years. His conclusion was that management did not really care about costs running out of control; it cared about power and control.
Kevin described a similar atmosphere. He increasingly felt his exchange resembled an aging state-owned enterprise. At a time when the broader exchange sector was being hit by repeated security incidents, technical teams did not receive more resources. Instead, they slipped into a mode focused on avoiding mistakes at all costs.
In his words, no one dared take risks anymore. People just wanted to get through their own tasks without something going wrong.
Timezone strain and language barriers compounded the pressure
John, who grew up overseas, said he had already run out of patience with that working environment before being laid off. From the time he joined, he felt the company had a heavily Chinese internal culture. Chat logs, JIRA tickets, meeting notes, and other documentation were almost entirely in Chinese. Foreign employees with weaker Chinese skills could feel pushed to the edge.
He also described a harsh work culture, fast pacing, and quarterly performance reviews. Because teams were spread across time zones, being online at unusual hours was treated as normal. John said his team’s weekly stand-up was scheduled for Sunday night, forcing him to end weekend plans early. A QA colleague in the U.S. time zone often sent messages at 11 p.m.
“We were always 24 on call,” he said, adding that he regularly saw coworkers submit code at 2 a.m. on Saturdays. Work-life balance, in his account, barely existed.
Power struggles at the top spilled down to everyone else
Richard joined his company at its peak and watched it decline. For him, the most striking part of that decline was not just business contraction but a raw and chaotic fight for power inside the leadership group.
The company’s partners, he said, ran into a severe trust crisis tied to government investigations and potential lawsuits. On one side, the CTO and CFO believed they had been misled by another partner, or at least had not received proper support when facing regulatory problems. The split became final.
One faction then took a core team and a senior employee to form a “board,” founded a new company, and became the actual developer behind the old product. Former friends effectively became clients within a month. By February, the new company was pushing forward at a rate of two new products a week. This was happening around the time Richard left.
Rank-and-file employees had neither a right to know nor a real choice in that conflict. The article presents them as casualties of internal struggle and instability.
It goes further, saying that in Web 3.0 many project founders and even exchange CEOs are only the public-facing figures. Real decision-makers often remain behind the curtain. Within that system, loyalty may matter more than innovation or technical strength.
Kevin said the toxic culture flows downward. In his view, people who survive and move up in that system are often those who are good at office tactics, skilled at managing upward, and forceful toward subordinates. Those outside the core circle can be eased out in stages: first they stop getting invited to meetings, then key decisions bypass them, then they are moved to marginal roles, then weekly reports and new assignments dry up. By the time a replacement is already in place, they realize they have been hollowed out.
He said Glassdoor comments often suggest coworkers themselves are decent and supportive, but the broader structure feels like a palace court where saying the wrong thing in front of a superior can carry consequences.
The exchange model is under pressure from several sides
Kevin said bluntly that he believes the business model across crypto has broken down. Historically, exchanges relied on two main revenue sources: trading fees and listing fees. When the market was hot, new projects flooded in, retail traders were active, and both fee streams rose, so teams expanded.
That setup has weakened, in his view. Many listed projects have since shown that they came in mainly to make money and then leave. Listing fees are now a bigger issue too. Kevin said exchanges charge extremely high amounts, with even a small project paying hundreds of thousands of dollars just to get listed, while its post-listing market capitalization may be only in the tens of millions.
He described a downward spiral: lower project quality, more tokens breaking issue price, retail investors leaving, trading volumes shrinking, fee income falling, and listing fees being pushed higher to compensate. In his account, exchanges have eaten too much of the ecosystem’s economics. The cost of building in crypto has become too high, while retail users are no longer willing to absorb it.
The rise of on-chain derivatives venues, represented in the article by Hyperliquid, has added to the pressure. The most profitable derivatives trading business for centralized exchanges no longer has to happen inside centralized systems alone.
Liquidity dried up, firms shut down, and VC money pulled back
Several interviewees pointed to October 10 last year. The report says that broad liquidation event left deep scars across the sector and badly damaged confidence. All open positions with leverage above 2x were forcibly liquidated that day, and retail investors were hit hard. The article says they have still not recovered.
The pressure on exchanges is now spreading outward. John told Zhangsheng BeatZ that many mid-sized Web 3.0 institutions managing between $100 million and $500 million are shutting down. Older financing strategies and DeFi yield strategies are getting harder to maintain. Since last summer, he said, crypto liquidity has been “extremely severe” in its deterioration.
According to his account, almost all altcoins launched in early 2025 are drifting toward zero, with very low book value. OTC volume is weak. In market making, apart from RWA-related business, there is little worth doing. A friend of his who ran crypto-neutral strategies at a market maker told him that even after improving strategy, gaining market share, and raising profit per trade, the company’s total profit still fell sharply. Profitability had broadly shrunk to about 30% of prior levels, and that person was eventually laid off as part of cost cuts.
Kevin said Web 3.0 venture capital is now very cautious on both ticket size and number of investments, to the point of almost not investing at all. Even when checks are written, they are far smaller than before. He put the contraction at 80% this cycle. In his view, that means fewer credible projects will even be available for listing if another bullish period arrives.
Project teams are under pressure too. Kevin said that apart from a handful of businesses with Web2-style B2B revenue, most projects have neither real B2B income nor real consumer-side income.
Many workers moved to AI, but bias followed them
For people laid off from crypto exchanges, the question is not just whether to switch sectors. It is whether there are enough places left to go. Kevin joined an AI startup after leaving his exchange job. Zhangsheng BeatZ says that is common: most people exiting Web 3.0 are heading into AI.
The logic is straightforward. AI is the hottest sector for funding and hiring, and crypto and AI share some overlap in growth work, user acquisition, and global operations, making parts of the skill set portable.
Richard’s break with Web 3.0 was more complete. He said the exchange where he worked was full of incompetence from top to bottom, from infighting among partners to apathy at lower levels. He also moved into AI and left crypto behind.
Only a small number managed to enter traditional industries. The article says some highly capable people from trading systems and risk control moved into traditional market makers and quant firms. Some operations, BD, and compliance workers also shifted into conventional brokerages as Hong Kong and U.S. equities markets stayed active. But these were exceptions. More often, laid-off exchange workers ended up at smaller exchanges lower down the ladder.
Bias is one reason. Zhangsheng BeatZ says some traditional finance HR departments will directly reject candidates if their resumes show they are still employed by a Web 3.0 company. In the stereotype held by part of the traditional finance world, crypto signals regulatory gray zones, speculation, and results that are hard to verify. People coming out of that environment carry that stigma with them.
The article says similar reservations exist inside AI as well. Some firms focused on large models and infrastructure remain wary of candidates from Web 3.0. In their view, growth in Web 3.0 has been built more on speculation and narrative than on durable technical barriers. In that frame, an operations employee from a crypto exchange and an operations employee from ByteDance are not viewed as comparable.
This downturn may be more than another cycle
The article closes by arguing that this Web 3.0 winter may differ from earlier ones. The competitive map has changed. Prediction markets such as Polymarket and Kalshi, along with retail brokerage trading, are now competing for the same retail users and the same pool of money. U.S. retail capital is flowing into AI stocks and prediction markets instead of returning to crypto.
Some industry participants cited in the piece believe conditions are even worse than in the 2022 crypto downturn. In 2022, retail users were still in the market. After the October 10 liquidation event, the article says, even the last leveraged retail participants were largely washed out.
Yet even in a shrinking market, exchanges have not stopped fighting one another. Zhangsheng BeatZ cites a person familiar with the matter saying that some HR departments at exchanges have even made “poaching talent from competitors with high pay” a KPI. The practice, according to the report, is to hire the person, let a few months pass, then fire them for one reason or another, disrupting a rival team while extracting intelligence and customer resources. The person hired becomes a disposable tool.
The article compares that behavior to China’s old food-delivery subsidy wars, where internet giants spent massive sums wearing one another down. It says crypto exchanges are replaying a similar script. The pie is shrinking, retail traders are leaving, and volume is falling, yet platforms still poach staff from each other, trip up rivals, and trade public insults.
The report argues that large-scale layoffs at exchanges are often explained away as a normal Web 3.0 cycle plus the pull of AI. But, as it noted at the beginning, more than half of tech layoff events in 2026 cited AI, while nearly 60% of companies admitted the real issue was financial pressure. Web 3.0, the article says, is no exception.
It then lays out the sector’s own problems: listing fees running into hundreds of thousands of dollars, large numbers of weak tokens being brought to market, retail users repeatedly losing money after price breaks, opaque performance systems and monitoring tools eroding trust and creativity, and firms spending scarce resources poaching from rivals instead of rethinking the business model. If this winter is not simply fate or a market cycle, the article asks, then who is responsible for the industry’s decline?
The original report says it interviewed multiple sources including Kevin, Richard, Xiaoyu, and John. Their real names, employers, roles, and tenure details were withheld for anonymity.