Voice control expert SoundHound AI (SOUN 1.39%) has been a volatile investment recently.

The stock soared in February 2024 as Nvidia (NVDA 4.11%) bought a few shares. It cooled down a bit when the artificial intelligence (AI) chip giant didn't follow up with a larger investment or a tight partnership. Then the meme stock community stepped in, driving SoundHound AI's share price skyward in an attempt to cause a lucrative short squeeze.

But short-sellers largely held on to their negative bets on the stock, and the broader market's volatility also weighed on these high-priced shares. Today, SoundHound AI is trading 62% below last December's record price.

Where will SoundHound AI go from here? Is this a good time to buy into a hot growth story, or is the stock still overpriced? Let's see what the company might do over the next year, and how investors should treat this promising but unpredictable investment right now.

What SoundHound AI actually does

Let's start with the basics.

• SoundHound AI addresses a massive target market, providing high-quality voice interpretation services to several important industries. From in-car system controls and drive-through windows to phone-based menu systems and home electronics, the company can deliver game-changing human-to-machine interaction experiences.
• The company has been around for two decades, but wasn't focused on business growth for a long time. SoundHound AI raised money in a 2022 initial public offering to support its newfound verve for financial gains. It had just started to apply its music-based AI audio research to other industries, with early clients including Hyundai, Pandora, and Mercedes-Benz.
• The business opportunity is enormous, but investors need to be patient. In February's fourth-quarter 2024 report, $34.5 million in revenues resulted in a GAAP net loss of $258.6 million. Most of that pain sprung from soaring stock-based compensation expenses, but the direct cash costs were also substantial. The company consumed $108.9 million of operating cash flows last year.
• So SoundHound AI's stock valuation is not supported by profits so far, and even the revenue-based valuation is extremely lofty at 45 times trailing sales. That's a lot, even for a company that doubled its fourth-quarter revenues year over year and has a $1.2 billion backlog of unfilled long-term contracts.

The big backlog debate: How to value future contracts

The most optimistic SoundHound AI bulls argue that the stock price should be calculated from that beefy order backlog, resulting in a hypothetical price-to-sales ratio (P/S) of 3.2. However, the average deal in that order book is about six years, so the average annual value over that period works out to $200 million. Then I'm back to an adjusted P/S ratio of 19.1, and the revenue conversion will be a gradual increase rather than a crisp jump.

To be fair, the order backlog also keeps growing. It's up from about $1 billion in the previous quarter and $661 million in the year-ago report. So the annual revenue value of this growing contract list should account for some continued growth, too. But the whole discussion only gets more theoretical and less realistic with these extra adjustments.

At the end of the day, I think it's fair to call the stock "very expensive" and leave the exact sales-based calculations alone. Some investors will be ready to take the high-priced risk of buying into SoundHound AI's long-term growth story at a lofty price. Others will stay away until one of two things happens: The stock price could drop to a more reasonable level, or the underlying business performance could rise.

Both of these value-boosting trends could very well play out simultaneously, and that idea looks likely to me. If so, today's SoundHound AI buyers will be fine in the long run -- but those who wait for a better entry price will probably enjoy even stronger returns.

What to watch for in the year ahead

2025 could go down as a transformative year in SoundHound AI's history. Management expects full-year revenues to land near $167 million, about double the $84.7 million seen in 2024. Earnings before interest, taxes, depreciation, and amortization (EBITDA) is also supposed to turn positive this year, compared to a $61.9 million EBITDA loss last year.

Of course, these targets were set in February and the global economy looks very different two months later. Whether you plan to buy now or wait for a better entry point, you should probably wait for May 8's first-quarter report. Management will probably provide more accurate full-year targets, either underlining or undermining your investment thesis.