iFlytek Q3 Revenue Up 8%, Non-GAAP Losses Widen
nashnova research
iFLYTEK (科大讯飞, 002230) posted Q3 revenue of RMB 6.56 bn, up 8% year-on-year, with net profit rising 31% — yet stripping out one-off items reveals a RMB 126 mn operating loss for the quarter and RMB 763 mn for the first nine months, meaning virtually all reported profit came from government grants and paper investment gains.
It reported a profit — so why is it still losing money?
Q3 net profit attributable to shareholders hit RMB 226 mn, up 31% YoY. But after stripping out non-recurring items, the company posted a RMB 126 mn loss. This means → the headline profit does not come from the core business's own earning power.
For the first nine months, reported net profit was just RMB 21.9 mn, while non-recurring gains totalled RMB 785 mn — of which RMB 471 mn was government subsidies and RMB 370 mn came from fair-value gains on stakes in Guoyi Quantum and UdiTech after their IPOs.
In plain terms = without subsidies and paper investment gains, iFLYTEK lost RMB 763 mn in the first three quarters — and that loss is still widening year-on-year.
Why does R&D spending exceed gross profit?
Q3 R&D spend topped RMB 1.7 bn, up nearly RMB 400 mn YoY, accounting for over 25% of revenue. Nine-month R&D expense reached RMB 3.85 bn, up roughly 20.8%.
Nine-month gross margin was about 39.8%, while the R&D expense ratio ran at ~21.2% and selling expense ratio at ~19.8% — combined ~41%, already above the gross margin.
This means → R&D and sales alone consume all gross profit before G&A and finance costs are even counted. A core-business loss is structurally inevitable at this spending level.
Where is the business mix shifting?
Government-sector revenue fell 2.39% YoY in Q3; management said it is actively shrinking smart-city projects. Smart education and smart healthcare both grew over 20%.
Contract liabilities — cash collected but not yet delivered — rose 54.83% from year-start to RMB 2.58 bn, mainly from prepayments. This means → the order backlog is healthy, but whether it converts to recognised revenue on schedule is the key variable.
In plain terms = iFLYTEK is pivoting from big government contracts to education and healthcare. The direction is clear, but the new segments are not yet large enough to close the profit gap.
Cash flow improved — but where did the cash come from?
Nine-month operating cash flow turned positive at RMB 264 mn, up 115% YoY. Q3 sales collections totalled RMB 7.79 bn, up over RMB 700 mn.
Yet period-end cash surged 83.5% to RMB 4.50 bn, mainly because a RMB 4.0 bn private placement completed in April 2026 — subscribed by 14 investors including the National AI Industry Investment Fund.
This reflects a cash-flow improvement largely funded by equity financing, not a fundamental shift in the core business's cash-generation ability.
Why is the balance sheet getting heavier?
Short-term borrowings rose 63% to RMB 1.89 bn; long-term borrowings rose 50% to RMB 4.06 bn; bonds payable nearly doubled, up 98% to RMB 1.60 bn — interest-bearing debt expanded across the board.
Inventory jumped 58% from year-start to RMB 4.79 bn, driven by raw materials and finished goods; management cited higher spending on large-model R&D and memory-chip stockpiling. Receivables stood at RMB 17.80 bn, roughly 97.9% of nine-month revenue.
In plain terms = borrowings are rising, warehouse stock is piling up, and customers owe nearly a full year's revenue — the balance sheet is under pressure on three fronts at once.
Can the core-business loss narrow?
iFLYTEK plans to unveil a new "cloud + edge" Spark large model on October 22, built on a fully domestic compute stack, upgrading its integrated cloud-edge solution.
This means → management is betting the turnaround on how fast Spark can monetise. Only if the model converts to paying revenue quickly will the heavy R&D spend pay off.
But with R&D plus sales expense ratios already exceeding gross margin, and debt and inventory both swelling, the window for commercialisation to deliver is not wide.
市场有风险，内容仅供研究参考，不构成投资建议。