Bilibili: a real profit turnaround isn't enough while games shrink

Bilibili just put up one of its best quarters ever — and the stock sat near its 52-week low. On August 27 the Chinese video-and-games company reported second-quarter net profit of RMB339 million, up 55% from a year earlier, adjusted net profit of RMB704 million, up 25%, and its 16th straight quarter of gross-margin expansion. Two weeks later the American depositary shares traded around $15, down roughly 37% on the year and about 57% below their 52-week high near $36. When an income statement improves and the price keeps falling, either the market is wrong or the profits don't mean what they appear to. Here it is partly both — and sorting out which half you are betting on is the whole question.

A turnaround that's real, but narrow

Bilibili is a video-community platform for China's younger users that also publishes mobile games. For most of its life as a public company it lost money, spending heavily to grow its community and content. That has changed. The business is now profitable and free-cash-flow positive, holding roughly $3.6 billion in cash against more cash than debt, with about $760 million of trailing-year free cash flow.

What finally flipped it to profit tells you how durable that is. Advertising is now the largest revenue line, about RMB3.1 billion in the quarter, up 28% on AI-assisted targeting and a 14th straight quarter of more than 20% growth. Mobile games — the segment that powered Bilibili's profit surge in 2025 — fell 14% to RMB1.4 billion on tough comparisons. Value-added services (subscriptions and live streaming) grew a slow 5%. Total revenue rose just 8%. So the profitability is real, but it rests on one high-margin engine doing nearly all the lifting while the previous one shrinks.

Why the market won't pay for it

That narrowness is what the falling stock is actually saying. Growth has dropped to single digits, China's consumer environment is soft, and in early September management added a concrete reason to sell. Bilibili priced a US$700 million convertible notes offering in which Tencent, a major holder, subscribed US$200 million while selling roughly US$400 million of its existing Bilibili shares, alongside up to US$300 million of buybacks, with proceeds earmarked for AI-driven growth. Convertible notes become new shares when converted, and a large shareholder selling on top of that reads as dilution and overhang. The ADS fell about 18% over the 30 days through early September.

On paper the reset has been severe. The market capitalization is around $6.4 billion and enterprise value about $4.2 billion — roughly 11 times EBITDA and under one times sales — for a business that generates cash and holds a net-cash balance sheet. Profits went up while the multiple compressed, which is exactly the "valuation fell faster than the business deteriorated" setup that usually rewards patience.

The proof point that decides it

But the cheap-enough test requires those profits to be durable, and there the evidence is genuinely mixed. The strongest bear fact is that Bilibili's prior profit engine — games — is shrinking and overall growth is single-digit, meaning recent adjusted net income depends on ads carrying a bigger share of a smaller growth base. If ads slow and games stay soft, the earnings gain stalls, and the market's forward multiple suggests it is not assuming these profits are the new run-rate.

The decision over the next two to four quarters comes down to one proof point: whether the game engine returns. New titles are slated for late 2026 and into 2027, which analysts expect to inflect game revenue upward and bring it back to year-over-year growth in the fourth quarter. If games recover while advertising keeps its 20%-plus streak, the reset multiple starts to look undemanding against a company that both grows and prints cash. If games stay weak and ads decelerate with the consumer, the $15 price is finding fair value, not a bargain.

My read: the turnaround headline is real and cash-backed — this is a materially better company than its 52-week chart shows. But a genuine turnaround is not automatically a buy, and the market's skepticism has a name, a shrinking profit engine and a fresh dilution overhang. The valuation has reset enough to be interesting, but not yet enough to be too cheap to ignore, because the next leg of earnings depends on a games recovery that isn't here yet. That makes this a watch-and-verify setup: it becomes investable when game revenue turns up or the overhang clears. Until then, the fall looks less like a mispricing of current results and more like a market that sees the profit gain but refuses to pay up for an unproven next phase.