As another earnings season approaches, Wall Street is entering a familiar but increasingly complicated phase: investors must decide whether lofty expectations for corporate growth are actually supported by earnings.
While simultaneously trying to understand where the next generation of artificial intelligence businesses will generate durable revenue. Morgan Stanley has highlighted more than a dozen stocks it is watching closely heading into earnings season.
Reflecting a market where expectations have become almost as important as the reported numbers themselves. For investors, the question is no longer simply whether companies can beat quarterly estimates.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
It is whether management teams can demonstrate that current valuations are supported by sustainable demand, improving margins and credible growth.
That distinction matters particularly for technology companies. The artificial-intelligence boom has already produced enormous gains for chipmakers, cloud providers and software companies supplying the infrastructure behind increasingly sophisticated models.
But earnings season can expose the difference between capital expenditure and actual economic returns. Companies may be spending heavily on AI because they fear falling behind competitors.
Yet investors need evidence that these investments can translate into higher productivity, stronger pricing power, new customers or entirely new revenue streams.
That leads directly to another question highlighted by Goldman Sachs: How will AI agents actually make money? AI agents represent a potential shift from today’s chatbot economy.
Instead of simply answering questions, agents are designed to perform tasks: conducting research, managing workflows, writing software, arranging transactions or interacting with other digital systems.
If that vision becomes commercially viable, the economic model could move beyond subscriptions and toward transaction-based or outcome-based payments. An agent that completes a business process could potentially be paid according to the value or volume of the work it performs.
A sales agent might generate revenue through completed transactions. A coding agent could be priced according to software development output. A financial agent might eventually execute permitted services and receive fees associated with those activities.
But monetization remains unresolved. The challenge is that AI agents could also place enormous pressure on existing software businesses. If one agent can replace several software interfaces, customers may become less interested in paying for dozens of separate applications.
Instead, they could pay for an intelligent system capable of completing the underlying task. That would fundamentally change the economics of software. For public markets, earnings season therefore becomes a test of two related narratives.
The first concerns whether today’s AI spending is producing measurable financial returns. The second concerns whether tomorrow’s AI products can establish business models capable of capturing those returns.
Morgan Stanley’s stock watchlist and Goldman’s focus on agent monetization point toward the same underlying issue: AI enthusiasm eventually has to become an income statement reality. Investors will be watching revenue growth, margins, capital expenditure, guidance and customer demand for clues.
Meanwhile, the AI industry is still experimenting with how autonomous systems should be priced and what customers will actually pay for. The market may have already priced in much of the technological promise.
The harder task now is identifying the economic architecture that converts that promise into recurring cash flow. This earnings season, therefore, is not merely about who beats estimates. It is increasingly about whether the companies building the AI economy can prove that intelligence itself can become a scalable business.