The contest for AI supremacy is shifting away from foundation models and toward the data infrastructure sitting beneath them, with enterprises that have spent decades accumulating proprietary data now positioned to extract the greatest value from the technology, according to Wedbush Securities.
The bank argues that as frontier models converge in capability, the trusted layer between enterprise data and external AI systems becomes the primary monetisation opportunity, with software companies able to leverage deep, embedded data environments set to be key beneficiaries.
Wedbush points to a wave of strong results across the software sector as evidence that the AI revolution has now reached the software layer in earnest, citing Salesforce, Palantir, Datadog, Innodata, Snowflake and MongoDB as companies already capitalising on the trend.
On the hardware side, the bank highlights Dell Technologies as a standout beneficiary of accelerating AI infrastructure investment, with the company reporting first-quarter AI server revenues of $16.1 billion, up 757% year-on-year.
Dell Technologies Inc (NASDAQ:DELL) recorded $24 billion in orders and more than $50 billion in backlog during the period, and now expects $60 billion in AI-optimised server revenue for its 2027 financial year, up 144% year-on-year and ahead of its previous guidance of $50 billion.
Wedbush notes that the AI buildout is expanding beyond graphics processing units (GPUs), the specialist chips that power most AI training, with demand now accelerating into central processing units (CPUs) and traditional servers as enterprises move to refresh compute environments capable of handling growing AI workloads.
Supply constraints are expected to persist through the second half of 2027, giving Dell pricing power that the bank believes will support continued outperformance.
On cybersecurity, Wedbush warns that the rapid proliferation of agentic AI, where AI systems act autonomously on behalf of users, has significantly expanded the corporate attack surface, creating new threat vectors including model poisoning, prompt injection and data exfiltration.
The bank argues that cybersecurity has migrated from discretionary technology spending to mission-critical risk management, and could represent as much as 10% of enterprise IT budgets as companies seek to protect their data environments.
Palo Alto Networks and CrowdStrike Holdings remain Wedbush’s top picks in the cybersecurity space, with both rated outperform and carrying price targets of $300 and $700, respectively.
The bank carries outperform ratings across its full coverage universe referenced in the note, with price targets of $230 for Palantir Technologies, $280 for Snowflake, $380 for MongoDB, $220 for Datadog and $100 for Innodata.
Wedbush describes the setup for upcoming earnings reports for its covered names as favourable, with significant growth runway remaining across the AI software and infrastructure landscape.