Why are software stocks struggling while AI megacaps soar to new highs?
Despite historically cheap valuations, cloud and software companies haven't benefited from the AI craze.
Do they offer the next big upside, or is there a structural issue with traditional software?
Some thoughts ๐๏ธ:
Why this has happened? What triggered the underperformance?
What caused the software sector to so drastically underperform S&P500 and NASDAQ? Well, there has been a ton of recent decline in forward-looking guidance from major software companies. Cutting guidance and getting on the low end of revenue growth is never good for growth stocks... This on the surface seems very unusual despite significant R&D investment from megacaps like Google, Apple, Microsoft, Amazon, Tesla, etc. Every megacap that is flush with cash has poured on average tens of billions into AI capex. The divergence between AI software spend acceleration and deceleration in traditional enterprise software spend is remarkable, with hyperscalers investing heavily in Gen-AI revenues while enterprise software faces more scrutiny and higher internal rate of return barriers.
Cyclical or Structural Weakness in Software Demand?
On one hand, this might just be purely cyclical because guidance/earnings, and therefore stock prices/multiples of software companies were looking pretty good back in Q4'23 and Q1'24 when we had cooler inflation data that led to lowering in rate expectations. Once we had higher rates again when inflation roared back up, multiples contracted. There is actually a very tight correlation between multiples and interest rates. On the other hand, this may be purely structural because while high cost of capital may make companies want to spend less on software bills like CRMs, ERP, etc. and really has to justify ROI, AI has a blank check and anything that can potentially be transformative gets sign-off.
Tech Cycle Patterns and Generative AI
Completing just writing off software as a defunct category makes no sense because tech cycles usually follow the logical progression of the I-->P-->A pattern: infrastructure, platforms, and applications. For a cycle to succeed, a killer application is essential. In the case of generative AI, we are still waiting for such applications that deliver significant results across various domains. The industry might have overestimated the initial infrastructure investments as the peak of innovation. Historically, value shifts from infrastructure to platforms and applications, as seen with the rise of cloud computing, which didn't diminish enterprise software as feared.
Traditional Software incumbents actually have an important moat
Data is crucial for the success of AI companies. Companies like Salesforce, Adobe, Intuit, ServiceNow, and Snowflake have a significant advantage due to their data incumbency. However, the necessary data to enhance LLMs and agents is still being developed. Combining older structured data with unstructured data and capturing new machine intelligence data is essential. Incumbent software companies need to adapt their architectures to make data more accessible, highlighting the importance of data platform companies like Snowflake, MongoDB, and Databricks.
Investment Outlook and Software Valuations
In the short term, easing high rates could improve buying cycles and demand for software. Companies that embrace generative AI and re-architect their platforms are likely to succeed. While software multiples are compressed due to plateauing growth expectations, hitting lowered sales estimates and raising guidance could lead to a re-rating. Although interest rate headwinds continue to affect valuations, anticipated rate cuts could provide a boost. Higher-margin software still trades at a premium, but higher-growth companies may see significant upside if the market shifts back to favoring growth over profitability.
tl;dr
I think overall the software sector is a fertile ground for the next wave of gains in the stock market.
1) The sector is cheap from a valuation perspective
2) Continued positive Macro data and interest rate cuts should provide a key fundamental and technical boon / re-rating to the industry
3) Many software incumbents being left for dead is stupid because they have the most data. And once GenAI LLM algos are commoditized (because many players are going to have their own ChatGPT or Claude or Gemini or Perplexity, etc.) long-standing longitudinal data in a particular niche is going to give the edge for specific verticals
4) Higher margin, profitable compounders that either are monopoly players or those that incorporate AI will be able to win and withstand any macro
5) Historically we have seen companies able to innovate their business model or company structure to tap into new tailwinds. As software shifted from on-prem to cloud, we saw companies like Adobe and others go to a Saas / recurring model and created tremendous value for shareholders.
Itโs done!! ๐๐๐๐
Introducing Addy - Your AI email assistant ๐ค
Addy is powered by ChatGPT and can write emails for you 10x faster in your preferred style and tone.
If you want early access, please retweet and comment below and Iโll send you an invite code!