Jensen Huang's $78 billion April guidance number and Marc Benioff's AI monetisation roadmap have given tech bulls their first clean arguments against the doomsday narrative that has hammered software stocks all year
It has been a rough start to 2026 for tech investors. A creeping fear that AI models from Anthropic, OpenAI and others would hollow out existing enterprise software has weighed on names from Salesforce to Microsoft to CrowdStrike, creating what Wedbush analysts call the "AI Ghost Trade," a wave of selling built on the premise that the software layer is about to be disintermediated by foundation models. This week, the broker argues that the narrative took two serious hits.
What Nvidia's numbers actually said
The first came from Nvidia Corp's (NASDAQ:NVDA, XETRA:NVD) earnings, which arrived well above even the most optimistic whisper numbers. Blackwell and Rubin demand is now forecast to exceed $500 billion for 2026, and Jensen Huang guided to $78 billion in the April-quarter revenue.
For Wedbush, the significance goes beyond Nvidia's own stock. The numbers confirm that AI infrastructure spending is accelerating, not plateauing, which matters for the entire data centre supply chain, from memory to components to Asia-based chip plays.
Equally important was what Huang said about software. The Nvidia chief told investors the market has "got it wrong" on software's role in the AI revolution, arguing the sector will be a major beneficiary rather than a casualty. Coming from someone with arguably the best vantage point in the industry, Wedbush views that as a meaningful signal.
Salesforce takes a step forward
The second catalyst came from Salesforce Inc (NYSE:CRM, XETRA:FOO). Marc Benioff has been the most visible target of the software fear trade, with investors questioning whether Agentforce can generate real revenue before AI agents erode the company's core CRM franchise. This week's results and guidance gave the bulls something to work with. Agentforce monetisation is starting to show up in the numbers, and Wedbush believes it could represent roughly 25% of Salesforce's revenue within a year, up from negligible levels today.
Why the doomsday trade looks mispriced
Wedbush's core argument is structural. Anthropic and OpenAI may have cutting-edge models, but they do not have 20 years of enterprise distribution, deep CIO relationships, or embedded vertical workflows. Salesforce, ServiceNow and Microsoft sit at the application layer, where actual business logic. AI tools are only as useful as the data they can reach, and the companies that control that data are the incumbents. The model layer, Wedbush argues, will commoditise faster than the workflow layer, and historically, platform owners tend to capture value when new infrastructure emerges.
The firm sees Salesforce, ServiceNow, Microsoft, IBM, CrowdStrike and Palo Alto Networks as the names most likely to lead a recovery in the software space over the coming months. The thesis is straightforward: the selloff has priced in disruption that, on current evidence, is not coming in the form the market fears.
This week did not resolve every question. But it was, in Wedbush's view, the beginning of a bottoming process.