The tech sector stocks have seen a sharp rebound in trade today. The Nifty IT Index is up over 2%, and the top 5 gainers on the Nifty are tech sector stocks. Infosys, TCS, Tech Mahindra and HCLTech are all up over 2% apiece.
In fact, the tech stocks have been battered significantly in February on concerns about AI-led disruption and the recent Jefferies downgrade on key stocks. Many market observers now believe that the sector might see some near-term rebound.
Meanwhile, the bigger concerns about AI continue to linger. In a recent report, HSBC Securities (USA) says that while artificial intelligence may be driving the current technology rally, the bigger and longer-term winner could be enterprise software companies.
HSBC says 2026 may be the real start of AI monetisation
In a February 24, 2026, report titled “Software Will Eat AI Equities”, HSBC argues that AI will not replace enterprise software platforms. Instead, AI will be embedded inside them, and 2026 could mark the beginning of serious monetisation for software vendors.
“We believe software will be the primary mechanism for the diffusion of AI across the world’s largest enterprises,” HSBC Securities (USA) said in the report.
The firm also states that foundation models and “vibe-coding” are “not replacements for software within an enterprise for their main IT platforms.”
At a time when investors are debating whether AI could disrupt traditional software vendors, HSBC’s analysis makes a direct counterargument: AI will sit inside enterprise systems, not displace them.
HSBC on why foundation models cannot replace enterprise systems
HSBC Securities (USA) says the idea that large AI models can replace core enterprise platforms is unrealistic.
“Foundation AI models are inherently flawed, technically, and are not suited for a ‘lift-and-replacement’ of major software platforms,” the report states.
The firm adds that while limited use cases such as image creation programs or small apps may be appropriate, this approach “is not realistic for the majority of high-fidelity enterprise class platforms.”
HSBC argues that enterprise-class software has been built and refined over decades to be “almost error-free with high throughput and reliability.” These systems are designed to process, store, check, format and execute operations in a deterministic manner, which is essential for global corporations accountable to shareholders and markets.
By contrast, HSBC describes AI as inherently non-deterministic.
“Within a full-blown enterprise application, we think AI is destined to be subordinate to the overall software platform,” the firm says.
It further notes: “The world is used to software platforms that are repeatable, auditable, and error-free in their day-to-day operations, and foundation models offer none of these attributes.”
According to HSBC, this makes wholesale replacement impractical.
HSBC on ‘Vibe-Coding’ and AI writing software
Another concern addressed in the report is whether AI tools could generate new software products that displace incumbents.
HSBC says writing code is only one part of building enterprise software.
“Vibe-coding shifts the burden of intelligent design to the coder,” the report states.
The firm argues that startups, foundation model vendors and consumer platform companies “have little to no experience creating ‘enterprise class’ software in the various silos within the software stack.”
Even if AI-generated code were available, HSBC says the hardest part of enterprise software is architecture and accumulated intellectual property. A “from scratch” effort would be “decades behind on the hardest part of the project (the overall architecture).”
HSBC also notes that displacing an incumbent vendor that runs the daily operations of a global enterprise is extremely difficult, even with a better or cheaper solution.
HSBC on why companies won’t build their own AI platforms
The report draws a historical comparison with the 1980s and 1990s, when companies wrote their own IT systems because robust enterprise platforms did not yet exist.
HSBC states: “Companies gave up writing their major IT systems decades ago.”
The reason, according to the firm, was economics. Developing, maintaining, and staffing massive internal systems was costly. Centralised enterprise software vendors could spread development costs across thousands of customers, making outsourcing more economical.
HSBC says that logic still applies. Companies focus on what they are most efficient at and outsource software to specialists. It argues there is no reason to expect a return to in-house development of complex AI-driven enterprise systems.
HSBC on who captures the AI value
HSBC Securities (USA) positions legacy enterprise software vendors as the key beneficiaries of AI adoption.
“We see the legacy enterprise software vendors as among the key beneficiaries and diffusion paths for unlocking the value-creation potential of AI within the production of the +USD100trn global GDP ecosystem,” the firm says.
According to HSBC, major enterprise vendors have already been designing, vibe-coding and beta testing embedded AI agents, with development beginning around 2024.
The firm states: “We see 2026 as the kick-off for monetization.”
These embedded agents, HSBC says, will handle domain-specific, high-value tasks inside larger software systems in a controlled manner. This approach mitigates the technical limitations of foundation models while integrating AI into production environments.
The report further argues that although hardware and semiconductor companies have benefited from AI so far, “we see the lion’s share of value being generated within the software sector.”
HSBC on valuations
In its comparative valuation table dated as of the close of February 20, 2026, HSBC covers major software companies and notes that sector valuations are at historic lows.
“Software valuation levels are at historic lows, even though the sector is poised to expand massively,” the firm says.
HSBC adds that it sees “strong demand momentum lasting for the foreseeable future.”
Conclusion
HSBC Securities (USA) concludes that AI will not replace enterprise software systems. Instead, AI will be embedded within existing platforms, where enterprise vendors control reliability, auditability and scalability.
The firm argues that foundation models are not suited to replace major IT systems, that vibe-coding does not eliminate the need for deep architecture expertise, and that companies are unlikely to return to building their own enterprise platforms.
