The State of Software Stocks

Software has taken a relative beating this year: while the software ETF, IGV, is up 8%, many stalwart software names like Adobe, ServiceNow, Salesforce, Asana, Workday, and Atlassian are down or treading water at best.

My take is there's opportunity here, as a bunch of these names will return to the premium valuations they used to command. NOW, CRM, INTU, VEEV, CDNS, and SNPS are some ones that look interesting.

The narrative that has punished these stocks is that AI will "eat" software. But not all of the threats that have been articulated withstand scrutiny:

1. AI "Eats" Software: Overhyped. The idea is companies will use AI to build their software in-house and stop buying from software companies. However, defining and testing software still takes work. There's a reason companies outsource or buy what isn't their core competency, and this should continue in the future.
2. More competition: Slightly overhyped. The idea is it's cheaper to create software, so there will be more competition. This is likely true, and especially true in niche markets that were too small to build for in the past. However, the history of the software industry is that the cost to create it has steadily decreased, allowing better and better software to be created. Given this, with increasing supply of software it seems like competitive advantages like distribution channels, brand, and lock-in will continue to be important. So many current leaders probably continue to win here.
3. Seat model risk: Real but not catastrophic. The idea here is that SaaS companies have overoptimized on charging per seat, but with potential reductions in headcount and generative AI computing economics, a model of charging for consumption may make sense in many software categories (think Adobe charging per use of Firefly rather than monthly fee per employee). Feels like this will require a transition much like the the move from packaged software to SaaS.
4. Relegation risk: Real. The idea here is that someone could become the "front end" for all enterprise software, interfacing with the employee and directing requests to other services or incorporating those services. This feels real, even if you're a system of record: imagine Microsoft dis-intermediating Workday, for example, by having copilot take care of HR related requests.
5. Content generation risk: Real. The idea here is that products that primarily output or help employees output content are at risk from the models themselves. I think this one is real too; a company like Adobe is at risk here.

Given all of this, I think it's possible to take a software name and evaluate its relative risk. I've done it for a bunch of names here.

Which software companies do you think are undervalued?