Kodiak AI heads into its third-quarter report with an awkward split screen. On the operating side, management spent the last call describing a company moving at high velocity toward a year-end launch of driverless long-haul trucking. On the market side, the stock has lost roughly three quarters of its value since that call, sliding to $1.12, far below its 200-day moving average of $6.12 and only modestly above its post-earnings low of $0.96. When a narrative improves while the stock collapses, the report that follows has to prove that the story is still worth financing.

The Street is not asking for much on the top line. Consensus calls for revenue of about $2.5 million, more than triple the $0.8 million of a year ago but below the $3.5 million reported last quarter. That sequential dip is less alarming than it looks. Roughly $1 million of last quarter's revenue came from a one-time DriveOhio demonstration, so the estimate effectively assumes the underlying business holds steady rather than shrinks. Analysts expect a loss of $0.25 per share, a sharp narrowing from the $0.80 loss last quarter and the $0.49 loss a year earlier. A swing that large deserves scrutiny, and investors should check whether any improvement comes from genuine cost discipline or simply from changes in share count and accounting items.

This quarter was always going to be a pause. Management guided to only three or four net truck additions, taking the deployed fleet into the high 30s, down from a recent pace of seven or eight per quarter, as Kodiak validates its new Gen 7 platform on Daimler's Western Star 49X. That transition is the most important catalyst introduced last quarter. Gen 7 promises about 50% more compute, a longer operating life, a low-double-digit cut in hardware cost and the first autonomous package compact enough for cheaper day cabs. A clean report would confirm the slowdown stayed within guidance, show Western Star validation on schedule, and restate that deployments reaccelerate in the fourth quarter. Any hint that the pause is stretching into a delay would undercut the back-half-weighted ramp the entire plan depends on.

The bigger milestone is the long-haul safety case. Kodiak's readiness measure climbed steadily from 78% to 91% over four calls, with management saying the hardest, longest-lead safety claims are already closed. Closing most of the remaining nine points this quarter would make the year-end launch look credible. Stalling near 91% would raise questions that no revenue figure can answer. Secondary checkpoints include whether the California heavy-duty permit arrived, whether the West Fraser logging pilot in Canada launched, and whether the Atlas delivery timetable and dual-depot expansion in the Permian are proceeding.

Cash may matter most to the stock. Management narrowed full-year free cash flow guidance to a burn of $155 million to $162 million, helped by a better-than-expected $38 million burn in the second quarter. With $151 million in liquidity that management said would last into the second quarter of fiscal 2027, the steep share price decline suggests the market is already bracing for a dilutive raise. Holding or improving that cash guidance, along with clarity on funding plans, would go a long way toward easing that fear.

Sentiment has softened from clearly bearish last quarter to only mildly bearish now, suggesting some investors view the selloff as overdone. That leaves a low bar for a relief move, but it does not guarantee one. In the end, the report comes down to a single question: can Kodiak show that the Gen 7 transition and safety case remain on track for a year-end driverless launch, while proving it has the cash runway to get there without punishing shareholders further?