Written by: Cosmic Wave Naruto, Deep Tide TechFlow

In February 2026, the technology stock market is experiencing a systemic crash referred to by some media as the “SaaSpocalypse.”

Salesforce's stock price has dropped nearly 40% from its 2025 peak; ServiceNow fell over 11% in a single day after its quarterly earnings report, simply because management mentioned in a conference call that “AI agents are complicating the visibility of seat growth”; Workday dropped over 22%; the entire S&P 500 Software and Services index evaporated nearly $1 trillion in market value within the first six weeks of 2026.

The logic of the market is straightforward: AI agents are now able to replace a large number of manual operations, and enterprises using AI have completed tasks that previously required 100 people, thus naturally no longer needing 100 software seats. The seat-based charging SaaS business model is considered to have reached the end of its historical trajectory.

At the height of this panic trading sweeping the industry, HSBC's Head of U.S. Technology Research, Stephen Bersey, released a provocatively titled research report:“Software Will Eat AI”.

His core argument can be summarized in one sentence:The market's panic is a misjudgment.

A Contrarian Report

“The market worries that AI will replace enterprise software, and this worry is misplaced.”

He wrote at the beginning of the report. In his view, AI will not eliminate software, but will be absorbed by it, becoming an embedded capability within enterprise software platforms.Software is not an adversary to AI; software is the vehicle through which AI reaches the real world.

This logic flips the entire narrative framework of the current market. The market's fear is “AI replacing software,” while Bersey's judgment is “software will tame AI.”

He drew a historical analogy from the Internet era: when the Internet exploded, the initial value accumulation focused on physical infrastructure—servers, fiber optic cables, data centers. A lot of capital flooded into hardware infrastructure, while those struggling early Internet companies turned out to be the ones that ultimately gained long-term value.Software is the terminal point of Internet value.

Bersey believes that the evolution of AI is repeating the same script. The years 2024 and 2025 are the infrastructure-building period, where computational power, models, and code integration all pave the way for the explosion of the software layer. And 2026 is the year when the engine truly ignites.

“Software will be the main mechanism for AI to spread in the largest global enterprises. We believe 2026 is the kickoff year for software monetization.”

Why Can't Foundation Models Replace Enterprise Software?

The most substantial argument in the report is the step-by-step deconstruction of the logic that “AI will directly disrupt software.”

The critics’ views seem persuasive: large language models can already write code, vibe coding (generating usable software directly from natural language descriptions) is on the rise, and AI model companies are making more attempts at the application layer, so why do enterprises still need traditional software systems like Oracle, SAP, and Salesforce that are costly?

Bersey's answer unfolds from three levels.

First, foundation models have “inherent flaws.”

The report clearly states that foundation models “have inherent shortcomings” and cannot perform the task of “completely replacing” core platforms for large enterprises. They perform well in narrow scenarios like image generation, small application development, and text processing, but this is “not realistic” for high fidelity and enterprise-grade core platforms.

The root cause is the limitation of training data. LLMs are trained on publicly available internet data, while the proprietary architectural knowledge, business logic, and operational norms accumulated over decades in enterprise software systems—these core intellectual properties are not available on the public web, making it impossible for AI to learn from or replicate them. The moat of Oracle and SAP systems cannot be caught up with by just writing code; it has accumulated over time and through business scenarios.

Second, the capabilities of Vibe Coding are severely overestimated.

The report directly points out the fatal weakness of Vibe Coding: it puts the responsibility and burden of design entirely on developers. You tell AI “I want a system that can handle global supply chains,” AI can generate code, but “how to define the architecture of this system, how to deal with unexpected situations, how to ensure it doesn’t crash under extreme pressure”—these judgments still require human intervention.

More critically, Bersey points out that these major AI model companies “have almost no experience in creating enterprise-grade software.” They are entering an extremely complex environment from scratch. Enterprise software has evolved over decades of iteration to achieve levels of “nearly zero error, high throughput, and high reliability”—a benchmark that AI newcomers cannot reach in the short term.

Third, the cost of switching for enterprises is a real high wall.

Even assuming AI can indeed write code of equivalent quality, the cost of enterprises replacing core systems remains extremely high: income disruption risks, productivity losses, system compatibility issues across IT environments, trust accumulation in vendor brands and service capabilities… these are all real switching costs that will not disappear just because AI can write code.

What enterprise software requires is years of verified 99.999% uptime and error-free operation in various complex IT environments. This trust is earned over time, not created by code.

Who Will Be the Real Beneficiary of AI Monetization?

If the first half of the report is a defensive argument, the second half is an offensive layout.

Bersey's core judgment is: the largest share of the AI value chain will ultimately flow to the software layer, not to the hardware and chip layer.

“We believe AI is the primary source of value creation within the software stack, and the largest share of long-term value will belong to software rather than hardware.”

He also pointed out that scarcity in hardware, GPU shortages, power limitations, and data center bottlenecks will continue to exist in the coming years. This scarcity reinforces the strategic position of software platforms:Only software platforms can translate AI capabilities into scalable, repeatable business value.

And the specific vehicle for monetization pointed to by the report is AI agents.

Bersey predicts that in 2026, there will be large-scale deployment of task-oriented, workflow-embedded AI agents in Fortune 2000 companies and small to medium enterprises. However, his qualitative assessment of agents is distinctly different from the mainstream narrative in the market; he does not believe that agents are disruptors replacing software but rather thinks they must operate within the parameters and permissions defined by software, and it is this “bounded agent” that can meet enterprises' needs for AI risk control.

In other words, enterprises do not need an omnipotent, freely running AI; they need an AI that can be governed, audited, and operate within a compliance framework. And this can only be achieved by agents deeply embedded in enterprise software systems.

“Software is the key pathway for enterprises to use AI in a controllable manner.” This is the central judgment of the entire report.

At the same time, the report also predicts that inference demand will gradually exceed training demand, becoming the primary driver of computational consumption growth, which means that as agents become more popular, computational consumption will not shrink but will continue to grow, further supporting the entire software and infrastructure ecosystem.

Opportunity or Trap?

When the report was released, the overall valuation of the software sector had already dropped to historical lows. Bersey's judgment is:The combination of low valuations and the upcoming monetization year presents an entry opportunity, rather than a signal to exit.

“Software valuations are at historical lows, even though the industry is on the verge of massive expansion.”

Regarding specific recommendations, HSBC's logic is clear: those software companies that have already established deep data moats, possess the capability to embed AI agents, and do not rely on pure headcount billing models will be the biggest beneficiaries of this wave of AI monetization.The buy rating list includes Oracle, Microsoft, Salesforce, ServiceNow, Palantir, CrowdStrike, Alphabet, etc., nearly covering all core players in enterprise software.

It is worth noting that HSBC also downgraded the ratings of IBM and Asana, while listing Palo Alto Networks as “underweight”; not all software companies can safely survive the upheaval, and the key lies in whether they can become the infrastructure for AI agents or be bypassed by agents as a human interface.

Bersey's report logic is rigorous, its timing precise, and its contrarian stance itself possesses a strong communication effect.

However, one question the report does not directly answer is: if AI agents can indeed operate efficiently within the framework of enterprise software, will the enterprise demand for software “seats” still quietly shrink? The value of software as a carrier for AI may stand, but whether the “headcount billing” business model can sustain the current valuations remains an open question.

Will software swallow AI, or will AI swallow software? This debate will have new evidence in every financial report of 2026.