This article comes from the WeChat public account China Entrepreneur Magazine. Author: Sun Xin; Editor: He Yifan; Trainee Editor: Li Yuan; Cover image generated by AI.
Over the past six months, a brand-new job concept from overseas—FDE—has ignited a talent war among tech giants, with million-yuan annual salaries merely the baseline.
ByteDance, Alibaba, Tencent, Kingsoft and other tech giants are all going full throttle to urgently recruit FDEs. China Entrepreneur found on platforms such as BOSS Zhipin and Liepin: Doubao is offering an annual package of 1.05 million yuan; Alibaba Cloud's large-model FDE role pays a monthly salary of 20,000 to 55,000 yuan with 16 months' pay per year; Moonshot AI's Kimi has posted an FDE position with an annual package of 1.09 million yuan. Industry insiders in Silicon Valley reveal that overseas FDE compensation is even uncapped.
A person close to ByteDance told China Entrepreneur that some ByteDance FDEs stationed at client companies can earn a daily rate of 10,000 yuan.
FDE, or Forward Deployed Engineer, refers to a Forward Deployed Engineer. It was popularized by the American company Palantir, and its core is "on-site delivery"—engineers go directly into the customer's site and embed standardized data analysis products into the customer's existing systems and workflows.
This playbook helped Palantir's stock price climb from under $6 at the end of 2022 to $200 in 2025, and it has also shown a remedy to domestic companies struggling with AI-to-B implementation.
The AI office war that began this summer has further intensified the giants' anxiety. Mid-year, ByteDance merged Feishu, TRAE and Coze into Doubao's workspace in just over twenty days; Alibaba combined QoderWork, some flank products of DingTalk, and MuleRun into Qianwen Office; Tencent's WorkBuddy requires thousands of service providers in its ecosystem to have their own FDE delivery capabilities.
Why are the giants collectively targeting FDEs? On one hand, the AI office track has seen price wars, channel wars, and product integration wars play out in turn. "What they are ultimately competing for is who can truly fulfill enterprise needs and standardize them into products—but there is a time window for this," said Liu Yuxuan of CDH Investments' Innovation and Growth Fund.
Amid the fierce battle, AI products also face a core contradiction in delivery that remains unsolved: "Sales can't articulate the requirements, and R&D can't reach the customers," a head of a leading cloud vendor summarized to China Entrepreneur.
"FDE was first proposed by Volcano Cloud, which is related to their to-C genes. In the past, sales didn't know where R&D was—they sold whatever they had. But with AI to B, especially to large B customers, you must discuss requirements, do planning and testing. If salespeople can't articulate it, you just put an FDE next to them—which is really just deploying a group of R&D people. So in the final analysis, FDE is still about improving the combat capability of frontline personnel," the above executive said.
FDE Is Just a New Name for Sales
Although FDE is a newly popular term, it is not a new concept. "The conditions for becoming a qualified FDE are extremely demanding, which is why they are expensive," said Li Feng, General Manager of Kingdee's Central SOE Digitalization Consulting Department.
In his view, FDE has emerged at this point in time because AI is no longer a tool but is beginning to enter organizations as a "work partner"—when productivity changes, production relations must change too.
In the industrial era, marketing, pre-sales, and delivery were typically handled by three separate lines, whereas an FDE must be able to win deals in the market, understand the business, translate requirements into technical language, and also deploy and deliver on their own. In short, it requires a versatile talent who can take responsibility for results and be paid based on results.
This is not a problem unique to China. "Hiring an FDE in the US is just as hard," Li Feng said.
In August 2026, Zhang Jian received FDE interview invitations from both Alibaba Cloud and Volcano Engine at the same time. He recalled the interview requirements: "A role similar to a TAM (Technical Account Manager), but requiring the candidate to cover a very long technology stack, from servers all the way to Agent platforms, with the position graded from level 1 to 10."
Zhang Jian told China Entrepreneur that FDE positions on the market actually fall into two categories: one is based on the company's own Agent products to build vertical industry capabilities, requiring understanding of industry workflows and boundaries, with deeper technical depth; the other has no in-house Agent product and uses external tools to custom-develop software from 0 to 1 according to customer needs.
The FDE on-site paid engagement model in China depends on complexity and is generally divided into two types: project-based, charging per person per day; and subscription products, with bundled service projects paid for.
For example, the ByteDance position Zhang Jian applied for does not require the FDE to be stationed on-site long-term; instead, it focuses mainly on improving their own Agent, with only occasional business trips. Alibaba Cloud also allows remote collaborative work, whereas some startups' FDE positions require long-term travel.
What an FDE specifically does is go deep into the customer's site, standardize customer requirements, and feed them back to product design. Zhang Lanyin, founder of Hangzhou Shikong Yongxian Technology, has been an entrepreneur in the to-B track for many years. She noted that an FDE must understand technology in order to spot where AI can be applied within the customer's processes.
While possessing full-stack technical capabilities, an FDE also needs a salesperson's keen instincts.
Li Ge, an OPC entrepreneur in AI to B, describes FDEs as "patching gaps." In 2022, a certain office software company widely recruited a batch of CSMs (Customer Success Managers) responsible for sales, channel training, and customer service. "Now they are mass-recruiting FDEs, which to some extent is making up for the CSMs' shortcomings in AI capabilities."
Guo Qizhi, Senior Partner at CDH, offered his judgment—consulting firms have always separated "hunters" and "implementers," whereas an FDE who can discuss business processes with clients "is best a former top salesperson, whose technical shortcomings can now be filled by AI Coding capabilities. Someone who can learn programming from zero is a natural FDE candidate."
From an applicant's perspective, Zhang Jian has no aversion to sales tasks. "If you only do technology, your boss most likely won't understand what you're building—you're just a cog."
This also leads to a situation where FDEs are often hired as technical roles, used as delivery workers, and ultimately evaluated as salespeople.
Why Do Customers Actually Pay?
In the giants' imagination, an FDE can single-handedly handle business analysis, AI technical closed loops, productization, and on-site communication. But after months of hype, social media platforms have become flooded with headlines like "The FDE Bubble Has Burst."
In fact, FDE is still far from the ideal state. Guo Qizhi said that what many FDEs currently do when they enter companies boils down to two words—training: running AI competitions for clients and teaching AI courses.
To understand why there is such a mismatch in delivery, one must first answer a question: What are customers actually paying for? And why haven't customers paid yet?
"In domestic to-B delivery, the moment there's a disagreement, the client makes you do customization, and if the platform ends up as a 'heavily modified' version, that's already pretty good."
Li Ge said that to understand why FDEs are necessary at delivery sites, one must first acknowledge one thing: most domestic enterprises do not trust AI, let alone heavily depend on any AI product on the market.
"Enterprise AI today has only achieved partial adoption." Li Feng of Kingdee divides AI implementation into four stages: individual application, partial enterprise application, enterprise-wide application, and societal management.
Although the market is flooded with claims like "we'll build you an ERP in three days," Li Feng's judgment is that the enterprise side has only reached the second stage. "End-to-end, full-domain—I haven't seen it yet to this day."
After all, from the enterprise's perspective, the B side has an extremely low error tolerance. "Customer needs are not standardized, and for AI agents in long-process business scenarios, the tolerance for error rates is very low. Even if AI has a 0.01% probability of mistakenly deleting enterprise data or getting an account entry wrong, enterprises will absolutely never accept it," said Liu Yuxuan.
Even overseas, customers are far from trusting AI. She cited the case of a semiconductor company served by OpenAI: the team produced a demo in just 4 weeks, but subsequently spent 4 months fully annotating every output of the system with its data sources. "This isn't an actual need; it's more of a psychological need."
Alongside the lack of trust, insufficient context and fragmented data are also becoming stumbling blocks for general AI entering large enterprises for implementation. Feng Luyao, founder of Yaoshi Technology, is currently an entrepreneur in the FDE track and was previously a frontline to-B salesperson at a major domestic tech company.
When working at a big tech company, Feng Luyao often encountered this same picture: hundreds of employees at a client company all used a certain to-B office product and built many AI agents, but the company's data didn't live in that software. Employees had to manually feed information into it, and every decision and review still had to go through a human pass.
"To customers, AI seems like icing on the cake, and a bit superfluous at that," Feng Luyao said bluntly.
She also encountered clients who purchased a leading AI platform, paid several hundred thousand yuan in the first year, and then completely refused to renew in the second year. That meant "from the big company's perspective, the deal became a money-losing business."
Many clients posed questions to Feng Luyao: Is there a way to let the boss intuitively see the results of using AI? "At least give me a platform that tells me how many tokens were used and how much efficiency improvement the company gained."
But today, the actual results of AI products are hard to display intuitively, and even the delivery process has yet to be standardized.
Feng Luyao revealed that after a salesperson at a leading AI platform closes a deal, the work is handed off to channel partners, who then try to bundle their own services into it. Although the big company's SOP (standard operating procedure) is standardized, the level of investment also depends on the client's brand strength.
"Reporting is chaotic, KPIs are chaotic, even work goals are chaotic," said Li Ge. "Sales of collaborative office tools can't take on the data integration module, because the revenue from selling data integration is credited to a different group of people."
In her view, big companies still take the product as the subject and work backward to B-side business processes, rather than working backward from B-side business processes to the product.
The size of an FDE team is the inverse of model capability
After more than half a year as an entrepreneurial FDE, Feng Luyao has truly felt how harsh reality is. Yaoshi Technology's clients fall into two categories: one explicitly requires getting everyone using AI first, in pursuit of cost reduction and efficiency gains; the other wants a complete platform, more like traditional software development work.
There is also a real temperature gap between the market's heat and enterprises' coldness. A manufacturing boss in Suzhou told Feng Luyao that smart manufacturing is still hard to achieve through AI: "Everyone says AI can bring earth-shaking transformation, but at least among the companies we've come into contact with, we haven't seen it. It's always the vendors trying to educate the market—customers simply have no corresponding demand."
Trickier than market education is the internal redistribution of interests that B-side deployment inevitably touches. After all, every deployment project cannot avoid the internal balance of interests within the client.
Which departments' data will be opened up, which positions' existing verification processes will be simplified, and who signs off on the final launch result. "You have to write these things clearly; then it's up to the boss to decide. If you can't write them clearly, even the strongest technology will be worn down to death in meeting rooms," said Li Feng.
Even in AI to-B delivery, there are natural differences between domestic and overseas models. Liu Yuxuan has a vivid sense of this: "Overseas, it's mostly standardized SaaS products; in China, project-based delivery is always needed."
Liu Yuxuan once spoke with the head of OpenAI's FDE team, whose team had only two people in the early days—"the company's attention was all on to-C products like ChatGPT"—and FDE was only formally elevated once to-B became a mandatory strategic question. She revealed that even though OpenAI's FDEs have achieved very high per-capita efficiency, as of the end of 2025 the entire team numbered only around 50 people.
By contrast, big Chinese companies are all building delivery and ecosystem service teams of several hundred people. Behind the stark gap in manpower investment—beyond corporate trust logic and private data management habits—the root cause also lies in the products and models themselves.
"Model capability still isn't good enough, and delivery is still being done with a SaaS to-B mindset. Model capability directly determines how many people are needed as a safety net after FDE products are standardized. As the long-horizon task accuracy of models keeps improving, FDE per-capita efficiency will keep rising."
According to her observation, Zhipu at this stage basically only outputs standardized APIs, backed by a strong model foundation. "They're being chased by money every day. The more people a company piles up on-site, the more it shows its products and models aren't good enough."
Liu Yuxuan emphasized: the true purpose of FDE is to "bring requirements back into the product in a standardized way." If it's merely on-site one-off project delivery, it's no different from outsourcing.
The next form of FDE: human + AI agent, one-person company
"From being deified, to coming down off the pedestal, to even being stigmatized—this reflects the market's own extreme irrationality," an industry insider believes. The core of FDE hasn't changed: going down to the business front line and then extracting common needs—that still holds true. "Human-machine combination is an inevitable trend."
"FDE itself is not a pseudo-concept, but if you treat it merely as a door-knocker for sales and customer acquisition, it becomes a pseudo-concept," said Guo Qizhi. Every company's business processes, management style, and data foundation vary enormously. "Big companies naturally pursue generalization—one product they'd love to sell to the whole world. To deeply solve real business problems, you often still need to find a third party to do it."
This has allowed a group of small players to grow in the gaps that big companies' delivery can't reach.
Feng Luyao's team initially had only three people: she was responsible for diagnosis, producing solutions, and "whipping up" demos; an engineer handled deployment and implementation; and one more person handled customer acquisition. The client profile is clear—companies of 50 to 300 people with no strong technical team, or where business and technology don't mesh well. Pricing starts at 58,000 yuan to transform one business process; for two processes plus consulting and training, they charge 129,000 to 159,000 yuan per year.
"The best people to do FDE are this wave of OPC (super individuals), because from sales to implementation and delivery, it's all one person," a netizen shared on a social media platform.
Zhang Jian has also noticed a trend: a group of programmers with more than ten years of development experience are transitioning into "super individuals." They personally complete only 30% of the architecture design and technology selection, handing the remaining 70% of execution work over to AI.
"A good FDE can be the CTO of a mid-sized company, and a qualified CTO can go start a company," Zhang Jian said.
Li Feng himself is the most concrete sample of a "one-person company." On physically isolated equipment, he has locally deployed four or five large models and 8 AI agents, which he calls his "8 mates." Around this computing setup, he has built supporting power supply, access control, isolation fencing, memory maintenance mechanisms, and an emergency task queue, defining it as "controlled industrial-grade."
However, the growth ceiling for small players has already appeared, and the door to big clients is slowly closing on them.
Feng Luyao observed that, for example, many of Feishu's original big clients have switched to purchasing Doubao Work. "Enterprise services—whether you call it consulting, on-site presence, or in-house engineers—is still a business of piling up manpower."
A model that relies on stacking manpower cannot escape the low-gross-margin dilemma. Some SME owners have already done the math: rather than paying a software service provider's million-yuan quote for an AI upgrade, it's more profitable to spend 200,000 yuan hiring a developer who understands AI agents and spend a year landing the project.
Feng Luyao has worked out two paths to break through. One is amplifying individual capability and accumulating human capital compounding: "For the same roles, I don't need to hire another Sales Engineer or a product manager—one FDE can solve it." In her view, the goal of big companies raising FDE salaries is to have a single person carry more functions.
The other path is to productize delivery results: distilling and abstracting customized solutions for individual clients into a reusable general-purpose foundation. The team has already accumulated two standardized solutions, for restaurant chains and cross-border e-commerce.
Whether compounding returns can be generated from a manpower business directly determines the quality of the FDE wave. As for the outlook, the industry remains deeply divided.
Li Feng believes: "Opportunity and bubble are separated by an instant—enter well and it's an opportunity; enter poorly and it's a disaster." His go-to-market strategy is to run dual S-curves in parallel, without rashly overturning the existing mature business. Kingdee has also already started related initiatives internally, pushing engineers to obtain both AI Engineer and FDE certifications.
"Everyone, please don't just go from concept to concept. Whether this role is called FDE doesn't matter—what matters is whether AI adoption at Chinese companies can truly be made to work."