What earnings tell us about data, the workflows, and AI adoption
The Read-Through
We know that AI investment questions are not just about models. As important and top of mind, it is the economics of AI and who captures the economics or who captures the value.
Each earnings report gives us a bit more clue. This past week was no exception. It is clear that data and analytics companies are folding AI into products that their customers already use for their workflows.
Three things remain true as ever, and we see from the earnings last week:
- Proprietary data is valuable when it is important for decision support.
- Whoever controls the enterprise workflow has a moat that is hard to break.
- Infrastructure demand is running hot.
Let’s dive into each of the three observations.
Data is an asset. Decision support captures the value.
Let’s first start with the numbers from earnings.
Equifax grew revenue 10% to $1.69 billion. U.S. Information Solutions grew 13%, and Workforce Solutions grew 7%. Its new-product Vitality Index exceeded 17%, well above its 10% long-term target. The company also lowered the midpoint of its full-year revenue outlook to $6.66 billion.
MSCI grew operating revenue 12.4% to $892 million. Organic recurring subscription run-rate growth was 8.7%, retention was 95%. Index revenue grew 18.1%.
Moody’s grew 3% on an organic constant-currency basis while the company's Analytics business grew 8%. Its Decision Solutions grew 10%.
These are information services companies effectively selling highly specialized and curated data products. These data products help inform decisions around credit, investment allocation, risk, and financial markets. The data products are not valuable merely because they have proprietary or specialized data. They create value and are growth drivers because end customers use these data products to make high-stake decisions. With these data products, the customers are flying blind.
Said differently and zooming out, a dataset alone is rarely a durable moat. Data becomes more defensible when it is difficult to replicate, improves its value with use, and produces reliable or better decisions when data is being used inside a customer's workflow.
The investment implication is not that every data or information services company will win in AI. Rather, companies with proprietary data that directly have decision support use are attractive.
No tech can replace strong customer relationship management and nurture.
Let’s first look at SAP and IBM earnings. SAP reported a current cloud backlog of €24.1 billion, up 28%, and cloud revenue growth of 26%. Cloud ERP Suite revenue grew 29%. Total quarterly revenue exceeded €10 billion for the first time, while non-IFRS operating profit rose 18%. CEO Christian Klein said the company’s Autonomous Enterprise strategy “is not just resonating, it is being adopted.” SAP also announced an agreement to acquire TechWolf, which I covered last week.
IBM grew revenue 5% to $17.1 billion, or 6% at constant currency. Software revenue rose 9% at constant currency to $7.8 billion, and Infrastructure grew 5% at constant currency. Operating EPS increased 11% to $2.95. IBM also raised its full-year free cash flow outlook to an increase of more than $1 billion year over year.
Having run finance for IBM Cloud, I pay attention to what these numbers actually tell us. Backlog indicates customer commitments. Software growth shows how the existing business is performing. None of these metrics alone establishes strong market observations on AI adoption per se but these metrics speak to what makes these companies have staying power – their customer base and sales channels.
Of course, no AI startups can replicate these sales channels of incumbents in a very short period of time or if they want to try, they have to raise a huge amount of money but even then, nothing is guaranteed. We can’t buy trust from customers. We have to build trust from customers step by step.
Smaller companies must recognize this and not just throw pretty demos over the wall to customers. It is going to back fire. Build the customer relationships and trust by demonstrating deep understanding of the problem, thoughtful approach in building data-driven and AI-driven products, and nurture every customer relationship all the time.
AI infrastructure demand is strong
Western Digital reported fiscal first-quarter revenue of $4.38 billion, up 55% year over year. Non-GAAP gross margin was 58.1%. The company guided to approximately $5.1 billion in next-quarter revenue.
Bloom Energy grew revenue 138% to $1.24 billion, with product revenue up 179%. It raised its full-year revenue outlook significantly. Vicor grew revenue 78.5% to $197.1 million. Royalty revenue was $66.6 million. Backlog rose 166% year over year to $406 million. Lam Research reported $8.2 billion in revenue, up 22% from the June quarter and from $5.3 billion one year earlier. Texas Instruments grew revenue 28% year over year to $6 billion, observing growth across industrial, automotive, and data-center markets.
We are still in an extended spending cycle on AI infrastructure, and demand is driving it. Many of the workloads are coming from consumers and businesses alike, and I don’t think we are going backwards.
The read-through to private markets
Invest in specific areas of defensibility and spend the necessary time in due diligence
Some of the questions we need to consider:
Get deep on the data, the pipeline, the customers, and use cases. What are the real moats and market advantages?
Understand how critical the workflow is and lay out a clear argument in using AI and agents.
What key decisions will be made with data intelligence? Would the decision be better with data intelligence? How do we know?
This phase of AI investing requires more than identifying a large market. The underwriting requires understanding where the economics settle: who owns the customer, who controls the critical data, who captures the productivity gain, and who has the pricing power. And how do we know that we have the right insights?
