AI revives moribund software stocks

Artificial Intelligence has finally given moribund software stocks a jolt of life. Wall Street has been searching all year for the biggest beneficiaries of the AI wave sparked by the launch of ChatGPT: now it is the turn of a group of companies that were laid low after the pandemic tech boom.

Software companies should be well-placed both to supply the tools companies need to build generative AI into their business processes, and to embed it into the applications that millions of workers use in their everyday lives. But it’s not at all clear yet which will find the best uses for the technology, or how they will get customers to pay up.

A spike in the shares of two companies that have struggled to grow consistently since their recent stock market listings highlights both the hope and the uncertainty. The stocks of Palantir and C3.ai have both about doubled since early May, as each has touted itself as a supplier of the tech platforms needed to make use of generative AI.

However, companies like these will be competing with giants such as Google and Microsoft, and the revenue impacts are entirely opaque. As Palantir chief executive Alex Karp told his investors last month: “We have no pricing strategy” for generative AI. The theory: if the new AI services are as good as the company says, customers will be happy to pay one way or another.

There will be plenty of competition. The plug-and-play nature of generative AI — anyone can tap into the large language models created by companies such as OpenAI — has made the technology instantly available to every software company.

There is an obvious risk that suppliers will race to add AI bells and whistles to their existing products without thinking through what real benefit the technology adds. Also, if every email provider offers automated text suggestions when you write a message, the feature will quickly come to be seen as commonplace, making it hard to persuade customers to pay a premium.

There is an added risk that if AI actually makes workers more productive, it could reduce the amount of software customers buy. This is the question facing companies such as GitLab, which is used to create and deploy software. Like many software companies, GitLab charges by the seat, or the number of people who use its service. If AI makes developers more productive, will customers need fewer of them — and pay for fewer seats?

GitLab chief executive Sid Sijbrandij tried to brush that concern aside this week, arguing that if AI reduced the cost of producing software, more software will be created. Wall Street liked what it heard. GitLab’s shares jumped by a third after it announced good results and outlined its plan to implant generative AI into every facet of its service.

The threat to pricing based on the number of users and the potential difficulty of persuading customers to pay a premium has led many software companies to explore the idea of charging based on consumption: the more that customers use new AI features, the more they will have to pay. That also has the merit of tying revenue directly to the usage of a service that has a high computing cost.

In the short term, however, this will bring the kind of uncertainty investors usually hate. C3, for instance, has blamed a slump in the remaining revenue due from its existing contracts — usually an important indicator — on the fact that it is switching to usage-based pricing. The decline is clear, the impact of a future revenue uplift uncertain. 

Adding to the uncertainty will be a short-term dip in profit margins. Most software companies are starting out cautiously, offering new AI features free of charge while they work out which will catch on and how best to charge.

In an interview with the FT’s Cristina Criddle this week, Adobe chief Shantanu Narayen compared this with previous technology platform shifts. He predicted an eventual shake-out of the many venture capital-backed AI companies that have sprung up and that lack an obvious business model. Previous platform shifts, however, brought protracted uncertainty before the winners emerged. 

Investors are already betting that incumbents such as Adobe will be in a strong position to ride the AI wave with its shares rising 30 per cent this year. Likewise, shares in ServiceNow, another established cloud software company that has talked of adding AI to many of its services, have risen about 40 per cent this year. But companies like these they still need to show they produce real value, and not just act as resellers of the generative AI produced by companies such as OpenAI. 

richard.waters@ft.com