freelance

The Chinese Social Algorithm

20 billion yuan vs 5,000 yuan a month. One man vs 320 million Chinese. The same mechanics. And probably the clearest glimpse of the future of work worldwide.

---

In June, China’s statistics bureau confirmed that 320 million Chinese now make their living from “灵活就业” (líng huó jiù yè, “flexible employment”), the official euphemism for freelance, gig and platform work. That is 44% of the country’s labour force. Weighted average income: 6,200 yuan a month. After mandatory individual social contributions: 5,000 yuan. That is €625.

The same week, 20 billion yuan from a single hand. That is the sum Liang Wenfeng, founder of DeepSeek, committed on 16 June 2026, in a personal capacity, to his company’s first external funding round. That is 2.8 billion dollars.

Everyone has seen both figures. No one reads them together.

That is precisely where the information lies. And probably a glimpse of the future of work across the entire developed world.

---

Liang Wenfeng, or the Absolute Concentration of Chinese Capital

DeepSeek’s 16 June funding round is anything but conventional. The structure is unprecedented in Chinese venture capital: capital channelled through a limited partnership controlled directly by Liang, a 5-year lockup for external investors, no voting rights granted to commercial investors. Only the state-backed Big Fund holds voting rights and is not locked up.

Tencent puts in 10 billion. CATL puts in 5 billion. JD, NetEase, IDG and Monolith make up the rest. The post-money valuation reaches between 52 and 59 billion dollars, exceeding MiniMax (17.7bn) and Moonshot (30bn) combined.

To the investors gathered in May, Liang gave a pitch no founder could ever deliver to Sequoia or Andreessen Horowitz: the priority is fundamental research towards AGI. Not commercialisation. Not an IPO. Not short-term shareholder returns. The technology itself, as the ultimate goal.

This deal says what it says. Liang Wenfeng has aligned his governance structure, his discourse and his capital with the priority Beijing has set for the decade: Chinese technological sovereignty.

It is an alliance between Liang and the Chinese state, backed by China’s big tech.

---

320 Million Chinese, or What Happens When the Sink Overflows from the Top

In 2015, 灵活就业 accounted for 120 million people, or 15.7% of the labour force. In 2025, it reached 280 million, or 38.6%. In 2026, it will exceed 320 million, or 44%. An absolute fourfold increase in 11 years.

Almost one in two working Chinese is a freelancer.

The raw figures do not tell the whole story. Three elements complete them.

Working hours. The average 灵活就业 worker puts in 49.6 hours a week. To earn 6,200 yuan gross. That is an hourly rate of 30 yuan (€3.75). For a municipal refuse collector or a delivery rider in the provinces, it falls below 20 yuan (€2.50). Beijing’s hourly minimum wage, the highest in the country, is 27.7 yuan (€3.50).

Social contributions. 灵活就业 status requires workers to pay 100% of their own pension and health contributions, i.e. 800 to 1,300 yuan a month. No employer contributes. No unemployment insurance. Real net income collapses to 4,900-5,400 yuan. The reality: many freelancers simply stop paying contributions.

The real composition of the flow. This is the most revealing element. Beneath the official figure of 320 million, the income distribution makes visible what the official report avoids naming.

China’s 灵活就业 is not just a new form of work. It is also, to a significant extent, a statistical receptacle absorbing what no one wants to call by its name: underemployment, disguised unemployment, post-layoff downward mobility, the reclassification of urbanised farmers without formal jobs.

---

The Chinese Social Algorithm: How Defensive Savings Finance Technological Sovereignty

The same China that produces 320 million freelancers on 5,000 yuan produces, in the same movement, the capital that makes it possible to concentrate 20 billion yuan in a single AI funding round.

The connection is not metaphorical. It is mechanical. But indirect.

Here is the chain. The Chinese household savings rate in 2025 is 36.2%. In the United States, it is 5.9%. Chinese savings exist because uncertainty exists: with no unemployment insurance, no guaranteed pension for freelancers and no robust safety net, ordinary Chinese people save out of defensive necessity. Out of 1.4 billion inhabitants, around 200 million retail investors have put part of these savings into China’s A-share markets. They make up the baseline liquidity of Shanghai and Shenzhen.

This massive, stable liquidity, fed by hundreds of millions of ordinary Chinese, is what makes high-performing quantitative funds possible in China. High-Flyer, the fund Liang Wenfeng founded in 2015 and which fully financed DeepSeek from its birth until the 2026 round, operates in this liquidity ecosystem. High-Flyer’s quantitative returns on A-shares allowed him to accumulate, over ten years, the personal capital that now enables Liang to write a 20-billion-yuan cheque.

Indirectly, it is China’s 200 million retail investors who made DeepSeek possible. A growing share of them are precisely 灵活就业 workers parking their defensive savings there.

Symmetrically, the technological fortification financed at the top will, over the next five years, accelerate the automation of white-collar jobs at the bottom. AI agents deployed at scale between 2026 and 2030 will replace junior analysts, copywriters, graphic designers, translators and accountants. Several tens of millions of university graduates will thus tip into 灵活就业.

The Chinese social algorithm runs in a loop. The defensive savings of a precarious population finance the technology that will widen that precarity.

---

The Future of Work: China Is Not Behind, It Is Ahead

The Western analytical error would be to read 320 million freelancers as a specifically Chinese crisis. It is a comfortable reading, but it is wrong.

Let us compare the ratios:

Note: definitions vary by country. Sources: Tencent x Ant Group (China), Upwork Freelance Forward (United States), IPSE (United Kingdom), INSEE (France). Orders of magnitude are comparable but not strictly homogeneous.

The direction is the same everywhere. China is simply further along the curve. The massive deployment of generative and agentic AI in 2026-2030 will mechanically bring Western ratios into line with China’s. In the United States as in Europe, the white-collar jobs that formed the backbone of the middle class (analysis, design, writing, legal support, accounting, junior project management) are precisely the ones AI agents can now perform.

China’s 灵活就业 is not an anomaly. It is a forerunner.

What is being built in China today:

a massive platform economy, destabilised social status, individualised social protection, contractual hyper-flexibility, income polarisation, capital withdrawing towards technological fortification rather than social security..

is probably the clearest glimpse we have of the model of work towards which developed economies are converging.

The difference is not in the destination. It is in the pace. And in the capacity for social absorption.

---

So, for International Brands in China: Choose, or Be Squeezed Out

China’s 320 million freelancers are not a peripheral segment. Together with the 500 million Chinese who share their disposable income bracket (3,000 to 7,000 yuan a month), they make up the real composition of China’s median market in 2026. That is five times the combined populations of France and the United Kingdom.

For international brands, three choices must be made. None is entirely comfortable.

Choice 1: Serve the Bottom, or Actively Retreat to the Top

Two legitimate options in the face of the freelance majority.

Option A: rebuild. Simultaneously rewrite the product formula (cheaper components, function preserved), the price (40 to 70% below the historical European positioning) and the channel (presence in Chinese hard discount is mandatory). This is not “downgrading”; it is a complete industrial overhaul, from sourcing to merchandising. Aldi has done it.

Option B: explicitly abandon the segment and focus the entire China strategy on the top of the pyramid. 80 to 100 million Chinese remain in the bracket above 12,000 yuan, plus the new “tech-rich” elite (AI engineers, EV unicorn founders, senior platform executives, A-share investors). A smaller, denser, more profitable market.

Both options are defensible. The third, the one most European headquarters still choose out of inertia, is not. That China no longer exists. Brands steering 2026 on the assumption of its return are managing an asset in silent decline.

Choice 2: Accept That the Middle Is Dead. Officially.

The Chinese market is not “polarising”. It has already polarised. Between ultra-luxury (which retains its value because it is decoupled from use) and industrial frugality (which is becoming mathematically unbeatable), the mid-market no longer has any economic space.

International brands positioned in this middle (mid-range fast fashion, FMCG without functional innovation, mass-market sportswear, general retail without local roots) are losing on average 2 to 3 points of market share per quarter. This is not a crisis. It is a continuous, slow, irreversible squeeze-out. No marketing campaign, no celebrity, no “brand refresh” will reverse this movement: it is not a brand problem, it is the structural erasure of the segment.

Staying in the middle is no longer a strategy. It is passively waiting to be squeezed out.

Choice 3 — Align Everything Else with the Real Market

Three operational adjustments, whichever segment is chosen.

Pace. The Chinese market runs on 7- to 14-day cycles, with things going viral within 72 hours. A brand steered from Europe with six-month sign-offs is operating out of time. Decentralisation to local teams becomes a condition of execution.

Channel. Hard discount now captures a dominant share of Chinese FMCG and snacking. Being absent from it amounts to withdrawing from the segment without explicitly deciding to.

Marketing. The median consumer no longer seeks to validate their status through the brand. They validate their intelligence through the quality-price trade-off. The emotional brand-building of the 2010s is spending in a mental geography that no longer exists.

The verdict is simple. Five years to pick a side. Serve the bottom or hold the top. The middle is dead.

---

And What Happens in China Will Come to Europe..

China’s 320 million freelancers are the forerunners of the European consumer of 2032-2040. When AI has automated 30 to 40% of European white-collar jobs, the median French, German and Italian consumer will also be in a squeezed-income bracket.

Brands that master this consumer base in China from 2026 will gain, over a ten-year horizon, a decisive advantage in their home market. Those that withdraw from the bottom of the Chinese pyramid will, by managerial design, withdraw from the segment that will dominate their national market in 2035.

China is not just one market among others. It is the operational laboratory for the future of global consumption.

---