The explosive growth of the artificial intelligence (AI) market, which had largely been driven by the feverish expansion of generative AI platforms like OpenAI's ChatGPT, lit a fire under many tech stocks over the past year.

Some of those rallies were justified. Nvidia's stock deserved to soar because its chips were required to process high-end AI tasks. Microsoft's massive investment in OpenAI and the integration of its AI tools into its cloud-based services also made it a great play on that secular trend.

Androids working in an office environment.

Image source: Getty Images.

But beyond Nvidia and Microsoft, a lot of tech stocks shouldn't have been lifted with the rising AI tide. I believe C3.ai (AI 3.02%), Palantir (PLTR 3.73%), and SoundHound AI (SOUN 5.77%) are three such stocks that could stumble in 2024.

What's wrong with C3.ai?

C3.ai develops AI algorithms that can be plugged into a company's existing software to accelerate and automate certain tasks. That strategy sounds promising, but C3 still generates about 30% of its revenue from a joint venture with Baker Hughes that is set to expire in fiscal 2025 (which ends in April 2025).

If C3 doesn't renew that deal, its revenue will drop off a cliff. It still faces stiff competition from similar AI services that are directly integrated into Amazon Web Services (AWS), Microsoft's Azure, and other leading cloud platforms, while robotic process automation (RPA) platforms like UiPath and generative AI services like ChatGPT could disrupt its long-term growth.

C3's revenue rose a mere 6% in fiscal 2023, compared to its 38% growth in fiscal 2022. It's started to offer usage-based plans (as opposed to its original subscription plans) to gain more customers in this tougher macro environment, and it's rolling out new tools for generative AI platforms to stay relevant in the AI race. But those strategies will keep its bottom line in the red for the foreseeable future -- and its stock doesn't look cheap at 12 times this year's sales.

What's wrong with Palantir?

Palantir's data mining and analytics platform gathers information from disparate sources to help its clients make data-driven decisions. Its Gotham platform serves government customers, while its Foundry platform provides similar tools for large commercial customers. The U.S. military and most U.S. government agencies currently use Gotham to gather intelligence.

After its public debut in 2020, Palantir claimed it could grow its annual revenue by at least 30% through 2025. Its revenue rose 47% in 2020 and 41% in 2021, but only grew 24% in 2022. For 2023, it expects just 16% revenue growth.

Palantir blames that slowdown on the macro headwinds for Foundry and the uneven timing of Gotham's government contracts. However, it's also facing competition from similar data mining platforms across the commercial market as well as internally developed platforms within the U.S. government.

On the bright side, Palantir has stayed profitable over the past year as it reined in its spending, and it initiated a $1 billion buyback plan in August. Those strategies are responsible, but they also suggest Palantir's business is maturing -- and its stock still looks pricey at 60 times forward earnings and 14 times next year's sales.

What's wrong with SoundHound AI?

SoundHound AI provides audio and speech recognition services and apps for a wide range of industries. Its flexible services are appealing to companies that want to add audio recognition services to their products without tethering themselves to a tech giant like Microsoft or Alphabet's Google.

SoundHound is still growing like a weed. Its revenue rose 47% in 2022 and it expects 44-57% growth in 2023. But before it went public by merging with a special purpose acquisition company (SPAC) in 2022, it told investors it could grow its revenue by 41% in 2022 and a whopping 245% in 2023 as it scaled up its business.

SoundHound blamed that slowdown on the tough macro environment, but it's clearly struggling to expand its business in the shadow of larger speech recognition platforms like Microsoft's Nuance and Google Assistant. It also has customer concentration problems: More than two-thirds of its revenue came from just three customers in 2022.

SoundHound might look reasonably valued at eight times next year's sales, but it's still deeply unprofitable and it ended its latest quarter with a high debt-to-equity ratio of 4.6. Those glaring flaws could all set it up for a steep drop in 2024.