SoundHound Just Beat and Raised. The Stock Is Still Down Half. The Next Year Is a Conversion Test.

SoundHound AI reported its best quarter ever on August 5 — record revenue of $61.9 million, up 45% from a year ago and ahead of the roughly $52 million Wall Street expected, with gross margin up six points and full-year guidance raised to $230–260 million. A few months ago that would have been the setup for a victory lap. Instead, the shares are down about a third in 2026 and roughly half over the past year, sitting near $6.78 with a market value around $3 billion.

That gap is the story. Improve an operating business and watch its stock fall, and the natural read is that something is broken. The less dramatic explanation is that the price ran far ahead of the numbers and is being repriced back to them. The next twelve months decide which reading was right.

The hurdle was set at the top

Score the business and the market separately, because they traded different hurdles. The business genuinely improved: revenue grew 45% year over year to $61.9 million in the June quarter, GAAP gross margin rose to 45.1% from 39.0% a year earlier, and the adjusted EBITDA loss narrowed 33% to about $9.6 million. Management pointed to OASYS, the company's agentic-AI platform, for larger enterprise wins, including an eight-figure channel deal in Latin America. That is an 85-point quarter by any normal measure.

The market's hurdle was set much higher, at the top of a run that took the stock past $22 within the last year. A fast-growing but still-loss-making AI favorite priced at a premium multiple has to beat a demanding expectation every time. The price of admission is still steep today: near a $3 billion market cap, you are paying roughly 15 times trailing sales, or about 12 times the midpoint of this year's raised guidance, for a company that has not yet made money and burns cash while it scales.

The conversion is the whole ballgame

The bullish case has never been about this quarter's revenue. It is about whether the backlog, the bookings, and the new customer base convert into the scale the valuation is asking for. On that front the second quarter added a structural answer: the pending acquisition of LivePerson, the digital-messaging and customer-service platform, for an equity value of about $43 million (while picking up roughly $74 million of LivePerson's cash), expected to close in the second half of 2026. The deal would add a combined footprint serving 25 of the Fortune 100, and management's projection is that the combined company can reach at least $350–400 million of revenue in 2027, from a 2026 track pointing near $245 million at the guidance midpoint.

Run that through the calculator and the tension resolves cleanly. At $3 billion, a $350 million 2027 revenue number drops the multiple to the mid-single digits on forward sales — a level a high-growth software company can plausibly defend. That is the entire trade: the LivePerson customer base and OASYS upsell engine have to turn a ~$245 million business into a ~$350–400 million one without needing to go back to the market for more capital.

What breaks the story

Those 2027 figures are company targets, not delivered results, and the model is unprofitable enough that the financing matters as much as the growth. SoundHound held about $203 million in cash with essentially no debt at the end of June, which buys runway but not forever. Between cash burn and the cost of integrating LivePerson, two weak quarters would force a choice: sell stock and dilute holders, or let the multiple compress further to match slower conversion. The bull case needs the revenue ramp and a narrowing burn to arrive together. If either one stalls, the same stock that once priced effortless growth will reprice the other way.

The checkpoint, then, is concrete: Watch how Q3 and Q4 land against the $230–260 million track, whether the LivePerson deal closes without heavy new share issuance, and whether the operating losses keep narrowing toward break-even. The signal that breaks the current narrative is a full year that scrapes the low end of guidance paired with an equity raise to fund the burn. As long as the conversion stays on schedule, the recent sell-off looks like the market working off a peak-era price — not the end of the story. When it slips, the discount will keep coming regardless of whether the underlying business is genuinely improving.