Although software stocks have gotten a little reprieve on Wall Street recently, the iShares Expanded Tech-Software Sector ETF (IGV 0.27%) is still down by more than 20% since the beginning of the year as of the close of trading on Tuesday. Investors have been lowering their terminal multiples on these stocks, and selling them due to concerns that artificial intelligence (AI) will be able to build similar software products and services much more quickly and efficiently.
Models from Claude to Perplexity to ChatGPT are certainly impressive, and they can complete a range of different tasks well. So it's understandable that investors and consumers are unsure about what the tech world is going to look like in the future. However, in a recent interview on CNBC, Nvidia CEO Jensen Huang said "the markets got it wrong" on enterprise software. Here's his argument.
Image source: Getty Images.
Replacements or assistants?
Concerns about where the trend is heading have mounted as AI companies have released new agentic AI tools and chatbots that not only can answer queries and create certain types of content when prompted, but can also carry out tasks that might have otherwise required more time.
For instance, Claude Cowork can conduct a range of tasks, such as organizing files, summarizing Slack or email conversations, and automating workflows.
Claude creator Anthropic also released new agentic AI tools that could help in specific job areas such as wealth management, investment banking, and human resources. For instance, these tools could review deals, conduct portfolio analysis, or produce guidance materials for a company's recent hires.
Perplexity CEO Aravind Srinivas recently retweeted a user who claimed to have used Perplexity Computer's $200-per-month system to build a terminal that analyzed real-time data for Nvidia stock, mimicking the services provided by Bloomberg terminals -- dedicated machines that cost users $30,000 per year.
NYSEMKT: IGV
Key Data Points
It's easy to see why investors and consumers are worried. However, Huang argues that agentic AI will not replace current enterprise software products and services; it will use them on people's behalf.
It's very likely these [software] companies that we're talking about are going to introduce agents that run on their platforms. You know, these agents, of course, they have to be experts at what they do, and nobody is going to understand customer service better than ServiceNow, and they are going to come up with agents that are really fine-tuned and optimized for the type of work that uses the tools that they have.
Huang went on to suggest that people should think about agentic AI in a similar way to how they would imagine physical robots behaving. Eventually, he suggests, when people have robots in their homes, it's unlikely that those robots will reinvent the products and tools they require to perform tasks. For instance, he said, if a household robot is tasked with heating up food, it's unlikely that it will try to come up with another version of the microwave; it will just use the microwave that is already in the kitchen.
Ultimately, Huang expects that agentic AI tools will assist humans and increase their productivity. If software engineers don't have to spend as much time writing actual code, but can direct agents to write code the way they want, they will be able to complete their work much more efficiently than before.
Is Huang's argument logical?
I think Huang's argument that AI will prove to be more of an assistant than a product replacement makes some sense. Think about how the internet enabled people to complete tasks remotely that they normally might have had to travel for, giving them more time to do other things. The internet also spawned many new jobs and companies.
However, I think it's not clear right now where these new jobs will come from. If AI can write computer code on command, presumably, fewer software engineers will be needed.
Citadel Securities market strategist Frank Flight recently wrote in a blog post that job postings for software engineers are on the rise, as is new business formation. However, the question becomes whether those new jobs can truly offset the level of job losses from the automation that is on track to occur. I haven't seen many experts really answer this question with anything more than a vague assertion that there will be new opportunities as a result of AI and that humans won't completely innovate themselves out of the labor market.
Until there are more specifics, I think many consumers will remain worried, and many software stocks will remain under pressure if investors believe their moats are being destroyed by fast-emerging AI technology.
