&format=webp&quality=medium)
A fresh round of caution is emerging in technology stocks after a short-lived recovery, as rapid advances in artificial intelligence continue to raise questions about the future of IT services.
A recent report by Motilal Oswal Financial Services highlights how Anthropic’s new model, Mythos, could be the latest trigger behind this renewed nervousness.
The brokerage notes that while markets had stabilised after earlier volatility linked to AI developments, the pace and direction of innovation are again prompting investors to reassess the long-term outlook for traditional IT services.
The Nifty IT index has seen a recent recovery after a sharp decline, rising 3.38 per cent over the past one week and 4.34 per cent in the last one month.
However, the broader trend remains weak. The index is still down 17.24 per cent over three months, 11.67 per cent over six months, and 17.55 per cent so far this year.
At its current level of 31,484.95, the index is down about 21.9 per cent from its 52-week high of 40,301.40.
Anthropic’s Mythos is a next-generation AI model focused on deep code understanding and cybersecurity. It is being rolled out in a controlled manner through “Project Glasswing”, with access limited to a select group of large enterprises, including Amazon Web Services, Apple, Microsoft, Google, NVIDIA and JPMorgan Chase, among others.
According to the report, Mythos outperforms earlier models such as Claude Opus across both coding and cybersecurity benchmarks. It scored 93.9 per cent on coding tests versus 80.8 per cent for Opus, and 83.1 per cent on cybersecurity benchmarks compared with 66.6 per cent earlier.
More notably, the model has demonstrated the ability to identify vulnerabilities that have gone undetected for decades. In one instance, it reportedly found a 27-year-old bug in OpenBSD, a security-focused operating system, which had not been flagged by human experts or existing tools.
The report notes that in certain cases, Mythos is “superior to most human cybersecurity engineers,” particularly in identifying and fixing vulnerabilities.
The concern for IT stocks stems from the expanding scope of AI capabilities. Earlier, AI disruption fears were largely limited to coding and basic automation. However, Mythos extends this into cybersecurity—an area that has traditionally been considered high-value, human-intensive work.
The report highlights that AI is now moving beyond horizontal use cases into domain-specific applications. Mythos represents a shift towards specialised models designed for enterprise workflows, particularly in areas such as vulnerability detection and exploit simulation.
This development “expands the list of things AI can do better than humans – coding, ERP, and now cybersecurity,” the report said.
For IT services companies, this raises questions about effort intensity and billing models. A large portion of revenue in areas like testing, maintenance, and security assessments depends on manpower and time-based billing. If AI can compress the time required for these tasks, it could impact revenue growth in the medium term.
The report identifies legacy enterprise systems as a key inefficiency that Mythos aims to address. Large organisations typically operate complex codebases built over 15–30 years, with multiple layers spanning applications, middleware, and infrastructure.
Security processes in such environments are often periodic, rule-based, and heavily reliant on manual effort. This leads to gaps, where vulnerabilities can remain undetected for years and security work becomes reactive rather than preventive.
Mythos appears to directly target these gaps. It can scan large, complex codebases, identify deep vulnerabilities, and significantly reduce detection timelines—from years to potentially overnight.
The report adds that this could reduce effort in traditional security services, particularly in testing and audit layers, though it does not expect an immediate disruption across the entire IT services stack.
Despite the technological leap, the report maintains that enterprise adoption of such advanced AI models is likely to be gradual. One key reason is the complexity of deploying AI in legacy, or “brownfield”, environments.
Unlike new-age companies that operate cloud-native systems, large enterprises require significant integration, data cleanup, and governance alignment before deploying AI at scale.
The report notes that 90 per cent of the top token users for OpenAI are new-age firms, indicating that adoption remains skewed towards companies without legacy constraints.
This suggests that while AI capabilities are advancing rapidly, their real-world deployment in large enterprises—and therefore their full impact on IT services—may take time.
The current decline in the Nifty IT index reflects a combination of these concerns. The index has fallen 17.76 per cent over three months and 12.23 per cent over six months, indicating sustained pressure.
The report also draws parallels with the earlier release of Claude Opus in February 2026, which had triggered a sharp correction in global tech and SaaS stocks. While Mythos may not have the same immediate market impact, it reinforces the direction of AI progress.
In the near term, the report suggests that this is more of a signal than a disruption. However, it flags cybersecurity and testing services as key areas to monitor, as AI-driven efficiency gains could gradually reshape demand