WeRide's H1 Results Show the Business Growing Even If the Stock Isn't

WeRide's stock has fallen roughly 63 percent from its October 2024 IPO price of $15.50 to about $5.70 today. The market concluded something obvious: another autonomous-driving company that hasn't yet proven it can print money. But the revenue and margin numbers from the first half of 2026 don't read like a stalled startup. They read like a company whose operating trajectory is getting cleaner even as the share price has not caught up.

The question for a reader who doesn't own the stock and is watching from the sidelines is whether that divergence reflects a real inflection or just a longer burn path dressed up in growth. The answer lives in one business line that most investors haven't priced in yet — and in one number that could prove it wrong.

The old story is becoming stale

WeRide's public narrative was always about robotaxis. Fully driverless vehicles in Abu Dhabi, Dubai, Beijing, and Guangzhou. A global footprint. A technology that could one day compete with Waymo and Baidu's Apollo Go. That story never changed. What changed underneath is that robotaxis are no longer the only business.

In the first half of 2026, total revenue grew 73 percent year over year to RMB 346 million (about $48 million), and the second quarter alone grew 82 percent year over year and 103 percent quarter over quarter to RMB 232 million. More importantly, gross margin hit a record 37.5 percent in Q2, up 9.4 percentage points from the prior-year quarter. Revenue is accelerating and the company is keeping a larger share of each dollar.

But here's the number that matters most: the L2++ and L3 advanced driver-assistance system business — essentially selling autonomous-driving software to car manufacturers — surged approximately 2,600 percent year over year in Q2. WeRide delivered roughly 30,000 software units in the quarter and has won production commitments for more than 30 vehicle models. Management targets over 100,000 units by year-end 2026 and cumulative deliveries past 500,000 in 2027.

That is not the robotaxi story. It's an ADAS licensing business. Car makers pay to install the software in production vehicles. The margins should be better than hardware, the scale comes from existing manufacturing capacity, and the revenue is recurring per unit sold. If the delivery numbers are anywhere close to the guidance, this business could become the cash-flow engine that makes the rest of the company affordable.

The cash that funds the dream

WeRide lost RMB 790 million in the first half of 2026 — about $109 million. That's the headline that keeps the stock suppressed. But losses without context are just noise.

The company held approximately RMB 5.4 billion in cash and liquid financial resources as of June 30, 2026. That cash pile is larger than the entire enterprise value of roughly $1.16 billion. The balance sheet carries only about $200 million in total debt. In plain terms: the market is pricing the entire company at less than what's already sitting in the bank.

At the current trailing twelve-month free cash flow burn of roughly $225 million, that cash provides approximately 3.3 years of runway. That's not generous, but it's not desperate. And WeRide has been actively repurchasing shares — HK$543 million in Hong Kong shares and $30 million in U.S. ADSs — which suggests management believes the runway is sufficient and the stock is cheap.

The company has also come back to the market for more capital, raising HK$2.39 billion in a Hong Kong IPO in November 2025. That was dilution, yes, but it extended the runway and signals that capital markets still have access to the company.

Management's stated timeline: positive cash flow in a single quarter by 2028, full-year profitability by 2029. That's two to three years from now. The bridge gets shorter only if the L2++ business scales as planned and the overseas robotaxi operations generate the recurring technology fees that management describes — estimated at over $50,000 per vehicle annually in steady state.

The asset-light pivot

WeRide is also changing how it operates internationally. Instead of buying and running its own robotaxi fleets overseas, it licenses its autonomous-driving technology to local partners — Uber in Europe and the Middle East, Grab in Southeast Asia, Geely Farizon in the Middle East. The company calls this an "asset-light model."

The accounting implication is direct: no vehicle purchase costs, no depreciation, lower capital intensity. Overseas revenue grew 164 percent year over year in H1 2026 and accounts for nearly 40 percent of group revenue. The Middle East fleet alone roughly doubled quarter over quarter to approximately 400 vehicles. WeRide now holds autonomous-driving regulatory licenses in eight countries and operates in twelve.

This matters because the asset-light model could make the international business genuinely profitable before the domestic one. It's also the part of the story that's hardest to verify — recurring revenue from Uber, Grab, and others depends on those partnerships holding and on regulatory approvals staying intact. The margin profile should be attractive if the model works.

What would prove this wrong

The bear case is straightforward and deserves its full weight. WeRide has not been profitable, has never been profitable, and may not be for years. Two thousand six hundred percent growth on the L2++ ADAS business looks impressive because it started from essentially zero last year. The question is whether it stays large or collapses back as car makers find cheaper alternatives, integrate their own solutions, or choose Tesla's FSD or a different Chinese competitor.

The 2028 positive-cash-flow target is management guidance, not a contract. R&D expenses grew 36 percent year over year in Q2 to RMB 434 million, driven by investments in AI infrastructure and foundation models. If R&D spending accelerates further — which it may, given how fast autonomous-driving technology is evolving — the path to profitability moves further away.

There's also competitive risk. Baidu's Apollo Go operates in 22 Chinese cities and has surpassed 22 million cumulative rides. Pony.ai is a direct competitor in both robotaxi and ADAS. Waymo leads in the United States. OEMs like Mercedes-Benz, with whom WeRide has a proof-of-concept L3 program, could develop their own solutions or pivot to a different software provider. WeRide's competitive moat is its regulatory approvals and safety record, but neither is permanent.

The single number to watch that tests this whole thesis: L2++/L3 unit deliveries. If the company hits 100,000 by year-end 2026 and keeps moving toward 500,000 in 2027, the ADAS business becomes a real revenue engine. If deliveries stall or the guidance gets revised downward, the profitability timeline stretches out and the market's skepticism returns with justification.

Where the numbers leave us

The market is still pricing WeRide as a pre-profitability robotaxi startup burning through its IPO proceeds. The operating setup is already showing revenue acceleration, margin expansion, a rapidly scaling ADAS business, an asset-light international strategy, and a balance sheet with more cash than enterprise value. The stock has been punished for the losses, which are real and material — but they exist alongside a trajectory that is getting cleaner, not worse.

This is not a call for conviction. WeRide is a young company in a young industry, and the path from 37 percent gross margins to positive free cash flow is still several quarters away at best. The $740 million cash pile extends that runway, but it won't last forever. What makes this interesting is the gap between the tape and the numbers: the revenue growth, the margin expansion, the L2++ delivery pipeline. If those continue, the stock may look like it's pricing a company that no longer exists. If they don't, the losses are what they are.