Cheche Group Inc. CCG

Key data

The business

Cheche Group runs China's largest independent digital platform for buying and selling auto insurance. It sits between the carriers who underwrite policies and the places drivers actually buy them: car-dealer systems, embedded checkout inside new-vehicle purchases, and phone apps. Cheche does not carry the risk; it routes the transaction and takes a slice of the premium. That is why revenue looks huge, RMB3.0 billion last year, while the company keeps only about a nickel of every dollar as gross profit. The pivot underway is deliberate: shed the low-value cooperative business, concentrate on insurance for new electric vehicles, where premiums are richer and Cheche's software, an AI agent that renews policies and a risk model, "Cheche Score," that prices each car, can do work people used to. That model now runs across more than 100 Chinese cities, and Cheche has agreements with more than 200 insurers. It is a thin-margin volume business trying to become a thinner-revenue, better-margin one.

The numbers

The story is two lines moving in opposite directions: revenue has fallen back to where it was three years ago, and gross margin has climbed to the highest Cheche has posted.

Revenue peaked at RMB3.47 billion in 2024, then fell to RMB3.01 billion in 2025 and to RMB885 million in the first half of 2026, down 34.4% year on year, all of it chosen. Gross margin went the other way as electric-vehicle premiums grew to 31% of the total from 22.5%. The catch is the bottom line: the loss narrowed hard in 2025, to RMB18 million from RMB61 million, then re-widened, with the first half of 2026 alone losing about RMB44 million. Restructuring is costing more than it is saving so far.

Here is the honest tension. Gross margin rose, but gross profit in money did not: about RMB58 million against roughly RMB66 million a year earlier, down 13%. The second half of 2026 is not yet reported, so the real test set last go-round, revenue down again but gross profit holding or rising against the first half, has not run. What the newest print shows is gross profit still following revenue down, just less steeply. The variant is narrow: if second-half gross profit turns up against H1's RMB58 million while revenue keeps falling, the leaner business is real. The single figure that settles it is second-half gross profit in money, not margin.

Management

There were no insider buys and no insider sales in the last twelve months, so nobody with the pen is voting either way with their own money; executive pay detail is not in the filings pulled here. Founder Lei Zhang still runs it. The record that matters is capital and survival. Cheche has never earned a full-year profit, has burned cash every year, RMB41 million from operations in 2025, and holds about RMB145 million, US$21 million, in cash. In January 2026 Nasdaq flagged the shares below the US$1 minimum bid; the company took a reverse split and won an extension to January 11, 2027, to comply. That is the clock management is actually running against.

How it fails or surprises you

The loss re-widens instead of healing. FY2025 lost RMB18 million; the first half of 2026 alone lost RMB44 million. The restructuring is subtracting revenue faster than it adds margin, and the read that this is a path to profit is what that number fits least well. If H2 2026 losses stay near the H1 pace, the story is broken, not maturing.

Delisting and dilution. The reverse split is spent; a second brush below US$1 by January 11, 2027 puts the listing at real risk. With US$21 million of cash against steady burn, a raise looks likely, and at eight-tenths of book it dilutes holders sharply. Watch the closing bid into year-end.

The lean business inflects (right tail). Electric-vehicle premiums are already 31% of the total and margin has reached 6.5%. If the AI renewal agent and per-car pricing lift high-margin revenue while costs stay flat, gross profit turns up on a company worth only US$38 million. The first proof is H2 2026 gross profit and NEV mix both rising.

Closing thoughts

This is a binary the next filing largely resolves, wrapped in a stock almost nobody can own at size. About 2.4 million shares exist after the split and roughly US$159,000 trades a day, so a real position moves the price by itself. The fatter tail is the left one: continued burn, a dilutive raise below book, and the delisting clock all point the same way, and only the tiny valuation, already under book, caps the damage. Thin-margin distributors that pivot to a "high-margin" mix are a familiar tape; the mix usually improves years before the profit does. What is at risk if the loss keeps running is most of the equity through dilution; what the right tail is worth is a re-rate off a sub-book, sub-US$40 million base if gross profit finally turns.

The bet is still that Cheche keeps selling car insurance to Chinese drivers, more of it on new electric cars, and makes each sale worth more as it drops the cheap business. What breaks it is simple: second-half gross profit below the first half's RMB58 million while the loss stays near RMB44 million, which would say the leaner business is just a smaller one. The claim is falsifiable on one line of the next 6-K: gross profit in money, up or down against RMB58 million.