I listened to the entirety of NIO's second-quarter earnings call on the evening of September 1st. William Li spoke for nearly an hour, and the central theme was growth quality; he and CFO Stanley Qu addressed 11 questions in total—from ES8 orders to next year's new Onvo models—yet every answer circled back to those two words.

The numbers presented were impressive on the surface. First-half sales climbed 67%, revenue jumped 86%, and gross profit surged 282%, with each layer outpacing the last. Second-quarter revenue hit RMB 32.1 billion, up 70% year-over-year. Losses narrowed by 90%, and gross margin reached 18.4%, compared to just 10% at the same time last year. It was an encouraging presentation, but the market response told a completely different story.

Checking the trading data, Hong Kong-listed shares of NIO-SW (HKEX: 09866) fell 6% during the day to close at HKD 31, hitting an intraday low of HKD 29.44—a 52-week trough. That decline came before the earnings release, reflecting August delivery figures. Once the results hit after the close, US-listed NIO Inc. (NYSE: NIO) shares dropped another 4% to settle just above USD 4. What looked like the company's best report card still drew two punches, and neither was spared.

The first blow traces back to March 10th—specifically March 10th, 2026—when NIO released its Q4 2025 results and posted its first-ever quarterly profit since its founding. Net profit attributable to shareholders was RMB 120 million—modest in absolute terms, but monumental in significance. That evening, Li's tone was noticeably lighter as he declared NIO had officially entered the third phase of its development, ushering in a new cycle of rapid growth with a full-year profitability target for 2026. The market had waited over a decade for those words; US shares surged 15% that day, adding USD 12.6 billion in market value overnight, and Hong Kong shares followed with a 14% gain the next day. By late March, the company's market capitalization had reclaimed the HKD 100 billion mark.

That single profitable quarter pulled NIO out of its "loss narrative" entirely. From that day forward, the market's lens shifted—previously, narrowing losses from RMB 4.9 billion to RMB 500 million qualified as good behavior. Now that profitability had been achieved, investors demanded quarter-over-quarter thickening of earnings. The scoring system had been completely rewritten.

The reasoning is straightforward: a company that has bled money for over a decade gets to sell stories. Battery swapping, R&D, multi-brand strategy, premiumization—each quarterly loss of billions bought a bit of patience for each narrative. But once profitability arrives, the storytelling ends. From that moment, NIO lost its "loss-making company" protective cover, and all that remained were hard numbers. And when the market crunches numbers, it looks at one thing: acceleration.

Applying these new rules to the Q2 report reveals three red flags. Net loss was RMB 528 million, expanding 59% quarter-over-quarter; adjusted net profit was RMB 26.1 million, down 40% sequentially—selling more while earning less per unit. The parent-company attributable figure looked even bleaker at RMB 722 million, up 45% quarter-over-quarter. To put RMB 26.1 million in perspective: with over RMB 30 billion in quarterly revenue, the profit attributable to ordinary shareholders shriveled to just over RMB 20 million. Deutsche Bank had forecast RMB 180 million pre-report—the actual figure didn't even reach the "zero" mark.

This wasn't unprecedented. On May 21st, NIO released Q1 results showing its second consecutive quarter of adjusted profitability. US shares rose 5% in pre-market trading, only to close flat at +0.18%, then fell 7% the next day. During that period, Hong Kong short interest spiked to 56% as bears placed their bets with real money. Today's sell-off was essentially rehearsed back in May.

Even more telling is the institutional response. Bank of America raised its full-year adjusted net profit forecast for NIO by 593%, yet maintained a Neutral rating—numbers going up, stance unchanged. Analysts themselves are speaking in two tongues. In March, they believed. By May, they were skeptical. By September, they stopped listening altogether. The market's trust in the phrase "NIO profitable" has thinned with each successive report.

The share price has faithfully recorded this wavering sentiment—peaking at USD 6.87 in April before closing at USD 4.06 on September 1st, a 40% decline over five months. Li's presentation focused on year-over-year growth; the market's mind was stuck on quarter-over-quarter comparisons. Neither side is wrong—the layered growth rates are genuine, and the sequential softening is equally real. Together, they produced yesterday's market action.

This is the crux of NIO's current dilemma: revenue up 70%, momentum still intact, but the market demands acceleration. Can the next quarter deliver even more?

Why is the acceleration stalling? Breaking it down comes down to two factors: volumes are flat, and costs are rising. NIO's Q3 delivery guidance sits at 108,000 to 111,000 vehicles, representing quarter-over-quarter growth of just 0.28% to 3%—essentially no growth. Crunching the September numbers clarifies the situation: July delivered 35,900 units, August 35,800, and to hit the lower end of guidance, September needs 36,200—just 400 more than August. That sounds achievable, except August already marked the second consecutive monthly decline, making even stabilization a meaningful feat.

Looking at August's new-energy vehicle rankings, Leapmotor sold 103,000 units, XPeng 39,000, Li Auto 37,000, and NIO 35,000—dead last among the four. Leapmotor's monthly volume nearly matched NIO's quarterly output. Within NIO's own sales mix, there's an even thornier issue: Onvo. At this time last year, Onvo accounted for half of NIO's deliveries. In August, Onvo sold just 8,810 units—down nearly 50% year-over-year—now representing only a quarter of total sales. The sub-brand had been growing in previous months, making this recent drop-off particularly concerning.

Li attributes Onvo's weakness to brand awareness, a problem that defies quick fixes. He compares Onvo to NIO five or six years ago, arguing that conversion rates strengthen once consumers understand the product. His implication is clear: Onvo's issue isn't the product—it's time. But time is precisely what the market lacks patience for. Onvo's strategic new model doesn't arrive until next year, leaving the remaining three months of this year with no help from that direction.

Then there's the cost side. In the earnings call, Qu reported that per-vehicle costs in Q2 were RMB 14,000 higher than at the end of last year, with another RMB 2,000-3,000 expected in the second half. For the full year, each vehicle carries an additional RMB 16,000-17,000 in costs. Where's the money going? Automotive-grade memory chips—whose production capacity is being diverted to higher-margin AI chips—are inflating prices. Lithium prices are rebounding, and industrial metals like copper and aluminum are also climbing. It's all raw materials, and it's all in the suppliers' hands.

The chip situation is the most frustrating. Memory manufacturers are shifting production lines to AI chips with better margins, leaving automakers holding cash and waiting in line. The hotter AI gets, the more expensive cars become—there's no way around this equation for NIO. The company can raise prices for customers—its average selling price exceeds even BBA—and customers accept it. But it cannot pressure suppliers; when chip prices rise, NIO must absorb them. With pricing power on the customer side but none on the supplier side, NIO holds only one half of the leverage.

The cost pressure was telegraphed back in March. Qu mentioned in the Q1 call that per-vehicle costs would rise by at least RMB 10,000, potentially reaching RMB 15,000. The September-reported RMB 14,000 lands squarely in that range. This isn't a black swan—the bad news was already on the table.

Facing these twin headwinds, NIO deployed three countermeasures. First, product mix optimization. The NIO brand's average selling price stands at RMB 406,000, rising above RMB 430,000 in July. Three-quarters of ES9 buyers come from outside the NIO ecosystem, and the ES8 has surpassed 140,000 cumulative units, targeting 150,000 by September. Using higher-priced models to support gross margins is working for now—but cycles have their limits. Li noted at a forum in April that new models sell well at launch, but by the time production capacity ramps up, demand has already cooled—he calls this the "new car death valley effect." The ES8 has been carrying the load for nearly a year, and the next ES8-level hit depends on upcoming 5-series and 6-series models. The ceiling is visible: product mix optimization requires volume growth, and Q3 volume is essentially flat, thinning the marginal benefit of this lever.

The second countermeasure involves metrics presentation. The official claim of three consecutive profitable quarters uses adjusted operating profit, which was RMB 207 million in Q2. After deducting RMB 554 million in share-based compensation, what remains is that thin RMB 20-something million in net profit. All the numbers are technically accurate—it's just that fewer people are buying the narrative.

The third lever is cost reduction. R&D expenses fell nearly 30% year-over-year. Li defends this by claiming that quarterly R&D of RMB 2-2.5 billion achieves efficiency comparable to others spending RMB 3.5 billion. Savings on costs translate to savings on the bottom line, though the elasticity is limited. Li has calculated that reducing per-vehicle cost by RMB 10,000 requires selling three to four times more vehicles to make up the profit difference—which is why he refuses to cut prices. The math is sound, provided volumes hold up.

What would convince the market? The answer is written in the share price. It only believes in one thing: cash. First, look at the market cap versus cash equation. NIO's market cap was HKD 76.9 billion at the September 1st close—roughly RMB 70 billion—of which RMB 56.7 billion is cash. That leaves the brand, battery-swapping network, three product lines, and the autonomous driving team valued at just RMB 14 billion. To put that in perspective: NIO generates RMB 32.1 billion in a single quarter. The market prices NIO's entire non-cash asset base at less than half of one quarter's revenue.

Second, examine operating cash flow. Over the last four quarters, operating cash flow turned positive with RMB 3 billion in net inflows—that's real progress. But the ugly truth sits right next to it: capital expenditures run at RMB 6 billion annually, primarily for battery-swapping stations. NIO has invested RMB 20 billion cumulatively in charging and swapping infrastructure, operating 4,123 stations nationwide, still adding roughly 1,000 per year. With RMB 3 billion coming in and RMB 6 billion going out, the annual gap is RMB 3 billion—plugged through financing. Over the past year alone, equity offerings raised over RMB 10 billion. Cash reserves grew from RMB 45.9 billion at end-2025 to RMB 56.7 billion, but that growth actually underscores that operations alone can't close the funding hole.

From another angle, the network does work operationally. Cumulative swaps exceed 120 million, and fifth-generation stations have cut construction costs to RMB 1.4 million each. "Working" means operations run smoothly—it doesn't mean the money has been recouped. Bulls see it differently: the swapping network is NIO's unique asset that competitors can't replicate, and if electricity trading is deregulated, those 4,000 stations become 4,000 energy-storage nodes. But there's a gap between "makes sense" and "is profitable"—it's called cash flow. A moat and a noose are the same thing: in growth phases it's a moat, in market downturns it's a money pit. Right now, the market prices it as the latter.

Third, examine the balance sheet. NIO's shareholders' equity has dwindled to just RMB 4 billion. Sitting above it are redeemable non-controlling interests of RMB 10.2 billion—a mezzanine layer two and a half times thicker than common equity. Where did this layer come from? Two financing rounds for the Shenji chip in June and August, at a post-money valuation of RMB 12.25 billion, with NIO's subsidiary holding 59.95%. External investors came in holding redeemable preferred shares, with redemption obligations ranked ahead of ordinary shareholders. These mezzanine interests also require periodic amortization—RMB 192 million in Q2, charged directly against common shareholders. The RMB 722 million attributable loss exceeds the total loss by RMB 200 million; that's the difference. Equity issuance fills the cash-flow hole, mezzanine eats at the income statement—same financing, two ledgers. Common shareholders sit at the bottom, money flows to mezzanine first, and whatever remains is theirs.

Finally, there's the next number the market is waiting for. NIO isn't without ammunition: four consecutive quarters of positive operating cash flow, fifth-generation swapping stations cutting costs by RMB 100,000 each—these are real achievements. Li has guided for monthly deliveries to push toward 40,000 in Q4. August delivered 35,800. That 40,000 target requires 4,200 more units than August—higher than the top end of Q3 guidance. Digging deeper into this math: Li's full-year target set in January was 40-50% growth. Based on last year's 326,000 units, that means 456,000-489,000 vehicles for the year. The first eight months delivered 262,900. The remaining four months need 190,000-220,000 units—a monthly average of at least 48,000. Even Li's own Q4 guidance of 40,000 monthly falls short of the full-year target's lower bound. The target and the guidance sit at different levels, with a notable gap in between.

On March 10th, the market awarded NIO a premium—one profitable quarter bought a 15% single-day surge, and by late March, market cap reclaimed HKD 100 billion. Back then, investors genuinely believed in the second half of the story: full-year profitability, a new growth cycle. By September 1st, with market cap at HKD 76.9 billion, the market would only value NIO by its cash. The same company, five months apart, carrying two different prices—the distance between HKD 100 billion and HKD 76.9 billion. That gap is the depreciation of trust.

Looking ahead, there's one variable that matters: the market is waiting for the next number. That number won't reveal itself until next quarter's results.