Have you ever witnessed a magician do a trick only to ask, "how did they do that?" Magic is interesting because even though you know what you're witnessing is merely an illusion, oftentimes you find yourself actually believing it.
Over the last year, the world has been captivated by the rise of artificial intelligence (AI). With chatbots, self-driving cars, and humanoid robots, AI almost seems magical -- except this time it's real.
Investors have been eager to buy into just about anything powered by AI, but not all opportunities are created equal.
Earlier this year a small developer of voice-recognition software called SoundHound AI (SOUN 5.00%) became an overnight sensation following news of an earlier investment from Nvidia. As of market close on March 28, shares in SoundHound AI had surged 178% so far in 2024. Given the hype, it's natural for investors to question whether they've missed the show.
One research firm is not impressed with SoundHound AI, and I think it has raised some legitimate concerns. Let's dig into a freshly published short report by Capybara Research, and assess why an investment in SoundHound AI carries outsize risk compared to other opportunities in the artificial intelligence (AI) realm.
SoundHound AI is not unique
AI-powered speech recognition is an interesting component of the overall AI narrative. Whether you realize it or not, you actually interact with this technology fairly often.
For instance, Apple integrates voice-powered AI into its hardware devices through the Siri platform. Both Amazon and Alphabet also leverage the tech in their Alexa devices and Google Home products.
Additionally, some companies that provide voice-recognition software services focus on specific markets and use cases. Two examples are Cerence and Genius, which primarily service the automotive and music streaming landscapes, respectively -- two areas in which SoundHound AI also operates.
Given the level of competition, it's not surprising to learn that short-seller Capybara Research labeled SoundHound AI's speech recognition tech as a "commodity service." Shorts make money when a stock's price falls, so it's important to understand that Capybara has a vested interest here. Nevertheless, I still think their analysts bring up a good point as it relates to the competitive landscape.
Image source: Getty Images.
Numbers worth noting
When I reviewed SoundHound's fourth-quarter and full-year 2023 earnings report a few weeks ago, I noticed something.
Management used to include a key performance metric called "cumulative bookings backlog" -- a figure that was meant to provide investors with a glimpse of future revenue commitments. However, in the fourth-quarter report this metric was updated to be called "cumulative subscription and bookings backlog."
While this subtle change in verbiage may seem harmless, management's commentary is what raised my eyebrows. CFO Nitesh Sharan said that this metric represents bookings backlog plus "new subscription revenue streams that we are focused on."
This is vague, at best. Sharan went on to say that "subscription backlog takes into account customers where we are the leading or exclusive provider and assumes a four-year ramp to fully scale with a total five-year duration." So management is forecasting future revenue that may not be fully committed yet.
My trepidation with SoundHound AI is that investors are pouring into the stock without really understanding the entire picture. Sure, it's exciting that Nvidia is working with the company. However, Nvidia's investment in SoundHound AI was in 2017. It's not new, and SoundHound AI is far from the only company Nvidia is invested in or partnering with in voice-powered AI.
SoundHound AI has reported some encouraging financial results in recent quarters. In fact, the company reported record revenue in Q4. However, Capybara's research report included an interesting detail surrounding the company's financials.
Specifically, the report leveraged footnotes from SoundHound AI's regulatory filings to uncover that the company is using one-time payments from terminated contracts to inflate revenue and margins in any given quarter. In other words, if a customer churns, SoundHound AI pulls termination fees forward. These are non-recurring sources of sales, but they appear as if revenue is experiencing upward momentum.
This accounting maneuver combined with the vagueness of the new cumulative subscription and bookings backlog metric makes me wary.
Where is SoundHound AI stock headed?
Capybara ended its report by putting a $1 price target on SoundHound AI stock -- it closed March 28 at $5.89.
I cannot say with any real certainty if that is where SoundHound AI stock is headed, but what I can say is that the company will need more than new metrics and it will need to prove that it can organically grow revenue as opposed to using an accounting device to accelerate its top-line profile.
If SoundHound AI fails to execute, liquidity will become an even bigger risk factor considering the company is still losing money and has only recently managed to reduce cash burn by laying off employees.
For now, I'd steer clear of SoundHound AI stock. The company represents an interesting opportunity, but with shares soaring to levels disconnected from the fundamentals of the business, investors are better off pursuing proven, established players in the AI industry.
SOUN PS Ratio data by YCharts
