With the rampant spread of AI (and the hype), tech stocks were bound to crash at some point – a jolt to make the industry rethink and reshape itself. Recent releases from Anthropic simply became the inflection point, a worthy one nonetheless.
The unveiling of Claude Cowork, which extends the benefits of the hugely popular Claude Code to non-programmers, and Opus 4.6, which greatly enhances the capabilities of Code and Cowork, led to a slide in tech stock values worldwide, especially undermining large IT services companies. We break down this development into four parts for you to mull over.
1. Runaway success of Claude Code
This whole story starts with developers falling in love with Claude Code, which was launched in 2024. Surveys show that many developers these days prefer vibe-coding. That is, they focus on creative aspects such as design and prototyping, leaving coding to AI assistants like Anthropic’s Claude Code, OpenAI’s Codex, Microsoft’s GitHub Copilot, or Google’s Gemini Code Assist. They simply describe their needs using natural-language prompts, and the AI assistant generates the code. No more worrying about cumbersome syntax and murky code reviews. Although every tech major has its own AI coding assistant, it is well known among the developer community that Claude Code is a class apart from the rest. The buzz is that even developers at Google, Microsoft, and OpenAI use Claude Code to get their work done!
In a recent X post, Boris Cherny, who heads Anthropic’s Claude Code, said that 100% of his code is now written by Claude Code and Opus 4.5, and that he has not written any code in over two months. Speaking at the World Economic Forum in January this year, Anthropic CEO Dario Amodei predicted that in the next six to 12 months, advanced AI models will be able to handle most, if not all, software engineering work from end-to-end. Many other tech majors have also openly admitted to AI assistants writing a significant amount of code within their organisations. While some developers feel threatened by this trend, many see it as a boon. It helps them to get more done in less time—and to focus their skills on prototyping, editing, and so on.
2. Claude Cowork extends benefits to non-coders
While revelling in the runaway success of Claude Code, Anthropic soon realised that developers were using Claude Code not just for coding, but for a lot of general-purpose tasks too—for example, to browse through a large folder and find some information, or to study and interpret some figures, to respond to emails, and so on. Many, including Cherny, claim to use Claude Code to handle project management tasks, like automatically following up with team members for weekly updates.
This realisation led them to develop Claude Cowork, an agentic system for non-coding tasks, built on the same foundations of Claude Code. At its most basic, it is an intelligent file management and automation tool that can handle multi-step tasks autonomously, with minimal human hand holding. It can read, draft, and review documents, handle correspondence, and execute multi-step workflows across legal, finance, sales, and marketing. For example, you can ask it to review a set of screenshots or invoices and prepare your account statement, or to identify spending trends and generate a presentation based on the spending. You can make it respond to customer requests or complaints over email, and initiate required actions for each. Claude Cowork is a proper autonomous agent that not only provides answers but also helps you by taking action.
Released in phases across January and February 2026, Claude Cowork is now available for both Windows and macOS, and includes 11 open source plugins that cover various verticals and functions – productivity, enterprise search, sales, finance, data analytics, legal, marketing, customer support, product management, bio-research tools, and plugin create utility (which you can use to easily create a plugin for your own specific needs).
The legal plugin has sent waves through the industry, automating tasks such as non-disclosure agreement (NDA) triage, contract review, compliance tracking, and legal risk assessment, which companies often handle with large, expensive teams.
3. Opus 4.6 takes all this to the next level
On the heels of Cowork, Anthropic unveiled its latest model, Opus 4.6, which elevates the capabilities of Code and Cowork. As per early reviews, it handles complex multi-step coding jobs with ease, even those that require careful planning and tool calling. Opus 4.6 makes it possible to assemble and manage teams of AI agents to work synchronously on varied tasks. It also features ‘adaptive thinking’ or the ability to understand contextual clues on how much to use its extended thinking, and ‘effort controls’ that give developers more control over intelligence, speed, and cost.
4. “Claude is built with Claude”
What really spooks the industry—is it that Claude Code/Cowork and Opus 4.6 can do so much and potentially put many out of work, or is it the fact that Anthropic keeps reiterating that Claude is built with Claude? At every opportunity, the company’s senior spokespeople make it very clear that more than 90% of the code at Anthropic is written using Claude Code. The company also claims to have built Claude Cowork in just 10 days using Claude Code!
An AI agent with 11 plugins spanning various functions and verticals developed in 10 days—that is startling. It simply means that a lot can be done in less time and, obviously, with fewer people. Does this put all coders at risk of losing their jobs? No, companies still need engineers and developers to oversee and control the AI systems. Also, a human mind thinks in ways that no machine can (to date). However, entry-level coders and specialists, who have very niche skills, may face problems keeping their job. Companies, including Anthropic, have now revealed that they prefer hiring generalists over specialists. Developers need to step back, take a broader view, and focus their time and skills on the creative aspects of design and development, as AI can handle the routine tasks. So, engineers need to upskill in the skills required to work effectively with AI. The educational system also has to revisit its syllabi to empower future generations with what is required to survive in the AI-powered tech world.
This trend tends to threaten not only developers but also IT services firms, which have for a long time been billing based on man-hours. No wonder stocks of such firms across the world took a beating following Anthropic’s announcements in early February. In reality, this was waiting to happen. Claude was a tipping point. The rapid rise of AI over the past year or two made this outcome inevitable. But does this mean IT services companies, especially those in India, will no longer have a role? Certainly not. Did the rise of email and instant messengers make post offices obsolete? The number of post boxes in a neighbourhood may have decreased, but post offices continue to operate, offering new services and modified operations. Likewise, tech companies now need to reinvent themselves, right from the roles they play to the business models. But they will continue to exist and support companies’ operations worldwide.
In a note titled India IT Services: Looking through the AI fog 2 – Discounted for extinction? Div/FCF yields at crisis levels, JP Morgan’s Asia Pacific Equity Research Team argues that AI is unlikely to replace IT services companies in the near future. Instead, it will be another tool to help them deliver more within constrained budgets. The report says it is too simplistic to assume that AI can automatically create enterprise-grade software and replace the integration and customisation work done byIT services companies, which it describes as the “plumbers of the tech world,” recognising their role in making complex systems work smoothly in large organisations. The report also says that the growth of AI will create new opportunities for IT services and support, including modernising legacy systems, rewriting customised software-as-a-service (SaaS) applications where required, building AI agents for operations, ensuring trust and reliability in AI systems, and integrating physical AI solutions.
As of the time we write this article, we see the stocks of global and Indian IT firms fluctuating significantly. Those who take this as a wake-up call and shake up their businesses will be the ones who survive this storm.
Janani G. Vikram is a freelance writer based in Chennai, who loves to write on emerging technologies and Indian culture.
