Anthropic's AI Agent Just Crashed Indian IT Stocks 6%: What Cowork Can Do, What It Can't & Why TCS, Infosys Are Panicking

Today Was Brutal for Indian IT Stocks

February 4, 2026, 10:30 AM: Nifty IT was down 5.9%.

By noon: Down 6.3%.

The damage:

  • TCS: Down 6.5% (₹3,014 from ₹3,225)
  • Infosys: Down 8.2% - steepest fall in 2.5 years
  • Wipro: Down 4-5%
  • HCL Tech: Down 5.1%
  • Persistent Systems: Down 7.5%

Total market cap wiped out: ₹2 lakh crore. In one day.

This is the worst single-day crash for Nifty IT since March 2020 (COVID crash).

What happened? One word: Anthropic.

More specifically: An AI tool called Cowork with 11 new plugins that can now do the work Indian IT companies charge billions of dollars for.

Let me explain what this tool actually is, what it can do, what it can’t do, and whether you should be worried about your IT job or career.

What is Anthropic Cowork?

Short answer: It’s an AI agent that can actually DO work, not just chat about it.

Longer answer:

Anthropic (the company behind Claude AI) launched Cowork on January 12, 2026. Then on January 30, they dropped 11 specialized plugins that turned Cowork from a helpful assistant into something closer to an autonomous worker.

The Key Difference: AI That Works vs AI That Talks

Traditional AI (ChatGPT, old Claude):

  • You ask questions
  • It gives answers
  • You copy-paste the answer
  • You do the actual work yourself

Cowork:

  • You give it a task
  • It does the task across multiple apps
  • It makes decisions
  • It delivers completed work

Example:

Old way:
You: “Write me a contract review summary”
AI: Gives you text
You: Copy, paste, format, send

Cowork way:
You: “Review these 50 contracts and flag the ones expiring next month”
Cowork: Opens files, reads them, analyzes, creates spreadsheet, sends report
Done. Without you touching anything.

That’s the shift. From AI as helper to AI as coworker.

And that’s what crashed the market today.

The 11 Plugins That Triggered the Panic

On January 30, Anthropic released 11 open-source plugins for Cowork. Each one automates an entire job function.

Here’s what they do:

What it automates:

  • Contract reviews
  • Compliance checks
  • NDA screening
  • Risk flagging
  • Document analysis

Who it replaces:

  • Junior lawyers
  • Paralegal teams
  • Legal process outsourcing (LPO) firms in India

Market reaction:
Legal tech stocks (Thomson Reuters, LegalZoom) fell 15-20% in the US.

2. Sales Plugin

What it automates:

  • Prospect research
  • CRM data entry
  • Call preparation
  • Follow-up emails
  • Lead qualification

Who it replaces:

  • Sales development reps (SDRs)
  • Inside sales teams
  • BPO sales support

3. Finance Plugin

What it automates:

  • Journal entries
  • Account reconciliation
  • Financial statement preparation
  • Expense tracking
  • Invoice processing

Who it replaces:

  • Junior accountants
  • Bookkeeping teams
  • Finance process outsourcing

4. Marketing Plugin

What it automates:

  • Content creation
  • Campaign analysis
  • Social media scheduling
  • SEO optimization
  • Performance reporting

Who it replaces:

  • Content writers
  • Digital marketing coordinators
  • SEO specialists (entry-level)

5. Data Analysis Plugin

What it automates:

  • Dashboard creation
  • Trend analysis
  • Report generation
  • Data visualization
  • SQL query writing

Who it replaces:

  • Data analysts (junior)
  • Business intelligence teams
  • Reporting specialists

6. Customer Support Plugin

What it automates:

  • Ticket resolution
  • Knowledge base queries
  • Escalation routing
  • Response drafting
  • FAQ management

Who it replaces:

  • Level 1 support agents
  • Customer service BPOs
  • Help desk teams

7-11: Productivity, Enterprise Search, Product Management, Biology Research, Custom Plugins

The pattern is clear: Any task that involves reading, analyzing, and producing documents can now be automated.

How It Actually Works (Technical Details)

Architecture

Cowork is built on four components:

1. Skills
Pre-programmed capabilities for specific tasks (e.g., “review contract for termination clauses”)

2. Connectors
Integrations with business tools (Salesforce, Google Workspace, Slack, Asana, etc.)

3. Slash Commands
Quick shortcuts for frequent operations (e.g., /analyze, /summarize, /compare)

4. Sub-Agents
Smaller AI agents that handle portions of complex workflows

How a Task Gets Executed

Let’s say you ask: “Analyze last quarter’s sales data and email the report to the leadership team.”

Step 1: Cowork breaks this into sub-tasks:

  • Access sales database
  • Pull Q4 data
  • Create analysis (trends, winners, losers)
  • Generate charts
  • Write summary
  • Draft email
  • Send to specified recipients

Step 2: Sub-agents execute each step autonomously

Step 3: Cowork monitors, adjusts if errors occur, delivers final output

You don’t write a single line of code. You don’t open any apps. It just… happens.

What Cowork Can Do (Capabilities)

Based on Anthropic’s documentation and early user reports:

✅ Can Do:

1. Multi-step workflows
Connect 5-10 different apps in sequence without human intervention

2. Read and analyze documents
Process hundreds of PDFs, contracts, reports in minutes

3. Write production-ready code
Generate, test, debug code across multiple languages

4. Manage complex projects
Track tasks, update stakeholders, adjust timelines

5. Handle data operations
Query databases, clean data, create visualizations

6. Perform compliance checks
Flag regulatory issues, ensure policy adherence

7. Learn company-specific processes
Plugins can be customized to your organization’s workflows

8. Work 24/7 without breaks
No sick days, no vacations, no overtime costs

What Cowork CANNOT Do (Limitations)

This is critical. Despite the panic, Cowork is not magic. Here are its hard limits:

❌ Cannot Do:

1. Make strategic business decisions
It can analyze data and present options, but it can’t decide “should we enter the Japanese market?” based on gut feel, relationships, politics.

2. Handle ambiguous human situations
Client is upset but won’t say why? Customer wants “something different”? Cowork struggles here.

3. Build deep client relationships
It can send follow-up emails. It can’t have coffee, build trust, understand unspoken needs.

4. Navigate office politics
Getting buy-in from 5 departments for a project? Understanding who has real power vs formal authority? Human skill.

5. Creative innovation
It can execute known patterns. It can’t invent the next iPhone or create a brand-new business model.

6. Physical presence
Can’t attend in-person meetings, shake hands, read body language, adjust pitch based on room energy.

7. Judgment in grey areas
“This contract clause seems odd but I can’t pinpoint why” – human intuition beats AI here.

8. Handle novel, unprecedented situations
COVID-like black swan events? AI follows patterns. Humans adapt.

9. Ethical reasoning in complex scenarios
“Should we launch this product even though it might have unintended consequences?” AI can’t navigate moral complexity.

10. Deep domain expertise built over decades
A 20-year veteran surgeon, lawyer, or engineer has tacit knowledge AI can’t replicate (yet).

Why Indian IT Stocks Crashed Today

Now let’s connect this to Indian IT companies.

The Indian IT Business Model

TCS, Infosys, Wipro, HCL – they all follow a labor-arbitrage model:

Step 1: Win contract from US/EU company
Step 2: Deploy large teams in India (cheaper labor)
Step 3: Charge by time & materials (billable hours)
Step 4: Profit from wage difference

Example:

  • US developer: $100,000/year
  • Indian developer: ₹12 lakh/year ($15,000)
  • IT company charges client: $40-50/hour
  • Pays Indian developer: ~$10-15/hour equivalent
  • Profit margin: 15-25%

The Cowork Threat

Old scenario:
Client needs 50 developers for a 12-month project.
Cost: $2-3 million to Indian IT company.

New scenario with Cowork:
Client needs 10 developers + Cowork doing 80% of routine work.
Cost: $500K-800K

Revenue impact: 60-70% reduction

Which Services Are Most At Risk?

High Risk (80-90% automation possible):

  • Application maintenance
  • Testing & QA
  • Data entry & processing
  • Level 1/2 support
  • Document processing
  • Routine coding tasks

Medium Risk (40-60% automation):

  • Application development
  • System integration
  • Migration projects
  • Analytics & reporting

Low Risk (10-30% automation):

  • Architecture & design
  • Client relationship management
  • Strategic consulting
  • Innovation & R&D

The Numbers Tell the Story

Indian IT industry size: $283 billion
Portion at risk from AI automation: 40-60% (~$110-170 billion)

Employment:

  • TCS: 6 lakh+ employees
  • Infosys: 3.5 lakh employees
  • Wipro: 2.5 lakh employees
  • HCL: 2.2 lakh employees

If 30-40% of work gets automated, that’s 4-6 lakh jobs potentially at risk over 3-5 years.

That’s why markets panicked today.

The Global Reaction

India wasn’t alone. This was a global selloff.

US Markets (February 3)

  • Goldman Sachs software basket: Down 6% (worst day since April)
  • Nasdaq: Down 1.43%
  • Software companies hit hard:
    • Atlassian: Down 8%
    • ServiceNow: Down 7%
    • Salesforce: Down 6%
    • Gartner: Down 9%

European Markets

  • Legal tech companies: Down 15-20%
  • Data analytics firms: Down 8-12%
  • Professional services: Down 5-8%

Indian Market (February 4 - Today)

  • Nifty IT: Down 6.3% (worst since March 2020)
  • ₹2 lakh crore market cap evaporated
  • Every single IT stock in red

Total Global Impact

$285 billion wiped out across software, IT services, professional services sectors.

What Experts Are Saying

The Bulls (AI Won’t Kill IT Services)

G Chokkalingam (Equinomics Research):

“Concerns are overstated. BFSI companies cannot replace core IT products with AI due to security concerns and regulatory risks. Indian IT companies are themselves integrating AI into their offerings.”

Key arguments:

  • Regulatory barriers in banking/finance
  • Security concerns prevent wholesale adoption
  • Indian IT firms will use AI too (level playing field)
  • US-India trade deal offsets some concerns
  • Rupee depreciation helps exporters

The Bears (This Is Just The Beginning)

Ambrish Shah (Systematix Group):

“As Indian enterprises integrate Claude into critical coding workflows, dependency on large vendor teams may decline, squeezing billable hours and margins.”

Jefferies (Brokerage):

“The evidence continues to grow that OpenAI is losing the corporate market to Anthropic’s Claude. Anthropic reached $1 billion in ARR just 6 months after launch.”

Key concerns:

  • Labor-intensive model fundamentally at risk
  • Entry-level talent pipeline threatened
  • Margins will compress as clients demand AI-driven pricing
  • Anthropic’s valuation reportedly hit $350 billion (up from $183B in 6 months)

The Realists (It’s Complicated)

VK Vijayakumar (Geojit Investments):

“The rally from US-India trade deal will face hurdles. Since valuations continue to be high, there is no fundamental support for a sustained rally.”

The middle ground:

  • Short-term panic overdone
  • Long-term concerns are real
  • Won’t happen overnight
  • Adaptation is possible but painful

Impact on SaaS Companies

Cowork doesn’t just threaten IT services. It threatens the entire SaaS (Software-as-a-Service) ecosystem.

SaaS Categories At Risk

High Risk:

1. Legal Tech

  • Contract management (Ironclad, DocuSign CLM)
  • E-discovery (Relativity, Logikcull)
  • Legal research (Westlaw, LexisNexis)

Why: Cowork’s legal plugin can do most of this work

2. Sales Tools

  • CRM add-ons (Outreach, SalesLoft)
  • Prospecting tools (ZoomInfo, Apollo)
  • Email automation (Yesware, Mixmax)

Why: Sales plugin handles prospecting, follow-ups, data entry

3. Marketing Automation

  • Content creation (Jasper, Copy.ai) – already disrupted
  • SEO tools (Ahrefs, SEMrush) – analysis automated
  • Social media management (Hootsuite, Buffer)

Why: Marketing plugin + AI creates and schedules content

4. Data Analytics

  • BI tools (Tableau, Looker) – threatened by data plugin
  • Reporting software (Domo, Sisense)

Why: Cowork can query databases and create dashboards

Medium Risk:

5. Project Management

  • Task tracking (Asana, Monday.com)
  • Collaboration (Slack, Microsoft Teams)

Why: Cowork can manage tasks, but human coordination still needed

6. HR Tech

  • Recruiting (Greenhouse, Lever)
  • Onboarding (BambooHR, Workday)

Why: Some automation possible, but people decisions require human judgment

Low Risk:

7. Deep Infrastructure

  • Cloud providers (AWS, Azure, GCP)
  • Databases (Snowflake, Databricks)
  • Security (CrowdStrike, Palo Alto)

Why: Cowork runs ON infrastructure, doesn’t replace it

The Microsoft Connection

Here’s an interesting subplot: Microsoft CEO Satya Nadella personally tested Cowork and reportedly urged his team to accelerate Microsoft’s own AI agent development.

Why this matters:

Microsoft owns:

  • GitHub Copilot (coding AI)
  • Microsoft 365 Copilot (productivity AI)
  • Dynamics 365 Copilot (business AI)

If Microsoft integrates Cowork-like capabilities into these products, every company using Microsoft tools gets AI agents by default.

That’s 345 million Microsoft 365 users potentially getting access to automated workflows.

Indian IT companies sell a LOT of Microsoft-related services (implementation, customization, support). If Microsoft’s AI can do this itself, the services revenue evaporates.

Should You Be Worried About Your Job?

I’m an engineering professor. My students aspire to work at TCS, Infosys, or Wipro.

Here’s my honest take on what this means for careers:

If You’re a Student (Not Yet Working)

Bad news:
Entry-level jobs will shrink. The “learn on the job” positions that Indian IT was famous for? Many will disappear.

Good news:
You have time to adapt. Learn what AI can’t do:

  • System design & architecture
  • Client relationship management
  • Domain expertise (deep knowledge of banking, healthcare, etc.)
  • Innovation & problem-solving
  • Leadership & people management

What to focus on in college:

  • Build projects that show strategic thinking
  • Learn AI tools (don’t fight them, use them)
  • Develop soft skills (communication, negotiation)
  • Get internships focused on high-value work

If You’re Working (0-3 Years Experience)

Highest risk group. Your work is most automatable.

Survival strategy:

  • Specialize quickly (don’t be a generalist coder)
  • Move up to senior roles faster
  • Switch to product, consulting, or client-facing roles
  • Learn AI tools and become the “person who manages AI”

Reality check:
Some layoffs are likely. Not immediate, but over 2-3 years, attrition won’t be replaced with new hires.

If You’re Mid-Level (4-8 Years)

Medium risk. You have experience but are still doing execution.

Strategy:

  • Transition to architecture, design, or consulting
  • Become a domain expert (fintech, health tech, etc.)
  • Move into management or client relationship roles
  • Consider startups (they still need hands-on builders)

If You’re Senior (8+ Years)

Lower risk. Your value is judgment, relationships, and strategy.

But stay sharp:

  • Keep learning (AI will change your tools)
  • Focus on areas requiring human wisdom
  • Mentor others on adaptation
  • Build your personal brand (LinkedIn, networking)

What Indian IT Companies Will Do

They’re not sitting idle. Here’s the likely response:

1. Embrace AI (Can’t Beat It, Join It)

  • Deploy Cowork and similar tools internally
  • Reduce team sizes, improve margins
  • Pitch to clients: “We use AI, so you get better/faster/cheaper”

Example:
Instead of 50-person team, offer 15-person team + AI agents at 60% of the cost. Still profitable.

2. Move Up the Value Chain

  • Focus on strategic consulting
  • Industry-specific solutions (can’t be easily automated)
  • Innovation labs and R&D
  • Product development

3. Acquisitions & Partnerships

  • Buy AI startups
  • Partner with Anthropic, OpenAI
  • Become AI service integrators

4. Cost Cutting

  • Hiring freezes (already happening)
  • Reduced campus placements
  • Attrition not replaced
  • Salary hikes slow down

5. Geographic Diversification

  • Expand in Japan, Europe
  • Reduce US dependency
  • Focus on markets slower to adopt AI

The Bigger Picture: Is This the AI Disruption We’ve Been Warned About?

Yes and no.

Yes, because:

  • This is real disruption, not hype
  • Jobs will be lost (probably 20-40% over 5 years)
  • Business models will change fundamentally
  • It’s happening faster than most predicted

No, because:

  • It’s not overnight extinction
  • Humans still needed for high-value work
  • New jobs will be created (AI managers, AI trainers)
  • Adaptation is possible

Historical parallel:

When ATMs were introduced (1980s), people said: “Bank tellers are finished!”

What actually happened:

  • Teller jobs fell 30% over 20 years (not overnight)
  • But banks opened MORE branches (ATMs reduced cost)
  • Tellers shifted to sales, customer service, problem-solving
  • Total banking employment actually GREW

The lesson:
Technology displaces tasks, not entire jobs. Humans adapt by doing different (often higher-value) work.

But here’s the catch:
That adaptation took 20 years. AI is moving faster. We might not have 20 years to adjust.

What This Means for You (Actionable Takeaways)

If You’re an IT Employee:

Short-term (Next 6 months):

  • Update resume and LinkedIn
  • Start learning AI tools (Claude, ChatGPT, GitHub Copilot)
  • Network aggressively
  • Save more money (emergency fund)

Medium-term (1-2 years):

  • Upskill into higher-value areas
  • Consider role change (technical → client-facing, execution → strategy)
  • Explore adjacent industries (fintech startups, SaaS companies)

Long-term (3-5 years):

  • Build career moat (what can’t AI do?)
  • Become irreplaceable through relationships, judgment, expertise
  • Consider entrepreneurship or consulting

If You’re a Student:

  • Choose specializations AI struggles with (system design, security, architecture)
  • Build soft skills (AI can’t negotiate, persuade, or lead)
  • Start using AI tools NOW (they’ll be everywhere when you graduate)
  • Consider non-IT careers if you’re only choosing IT for stability

If You’re an Investor:

Avoid (or underweight):

  • IT services companies (TCS, Infosys, Wipro) – structurally challenged
  • Legal tech SaaS
  • Basic automation tools

Favor (or overweight):

  • AI infrastructure (Nvidia, cloud providers)
  • AI-native companies (Anthropic, OpenAI via private funds)
  • Indian companies using AI to disrupt incumbents
  • Deep-tech businesses AI can’t replace easily

If You’re a Company Leader:

  • Experiment with AI agents (Cowork is free for paid Claude users)
  • Identify which processes can be automated
  • Invest in reskilling employees
  • Don’t panic-layoff (you’ll lose institutional knowledge)
  • Focus on areas requiring human judgment

My Personal Take (As an Engineering Professor)

I’ve been teaching for 14 years. I’ve seen tech waves come and go:

  • Dot-com boom/bust
  • Outsourcing wave
  • Cloud computing
  • Mobile revolution

This feels different.

AI agents like Cowork aren’t incrementally better tools. They’re qualitatively different. They don’t just help humans work faster. They work instead of humans.

For Indian IT specifically:

The business model that built the industry – cheap labor, large teams, time & materials billing – is now vulnerable.

But I don’t think it’s the end. I think it’s a forced evolution.

Indian IT companies have survived disruption before:

  • Y2K ended → they pivoted
  • Dot-com bust → they adapted
  • 2008 crisis → they endured

They’ll survive this too. But it won’t be painless.

For students and workers:

You’re entering a job market that’s changing faster than any generation before you faced.

The good news? You’re digital natives. You adapt quickly. You can learn AI tools faster than older generations.

The bad news? The safety net of “I’ll just join TCS/Infosys” is shakier than it was for your seniors.

My advice:
Don’t fear AI. Don’t ignore it. Master it.

The people who thrive in the next decade won’t be those who compete with AI. They’ll be those who use AI to do things that were previously impossible.

What to Watch Going Forward

This week:

  • More IT stock volatility likely
  • Analyst downgrades coming
  • Company management statements

Next month:

  • Q4 earnings season (TCS, Infosys report in mid-April)
  • Management commentary on AI impact
  • Hiring plans (or lack thereof)

Next quarter:

  • Campus placement numbers (big indicator)
  • Client contract renewals (are sizes shrinking?)
  • Margin pressure visible in financials

Next year:

  • Workforce restructuring announcements
  • AI integration by IT companies themselves
  • New business models emerging

The Bottom Line

Today’s crash happened because:

  • Anthropic’s Cowork demonstrated that AI can now do real work, not just assist
  • 11 plugins automate tasks Indian IT companies charge billions for
  • The threat is real, immediate, and growing

What Cowork can do:

  • Multi-step workflows across apps
  • Document analysis at scale
  • Code generation and testing
  • Data analysis and reporting
  • Routine professional tasks

What it can’t do:

  • Strategic decision-making
  • Deep client relationships
  • Creative innovation
  • Navigate ambiguity and politics
  • Handle unprecedented situations

For Indian IT:

  • 40-60% of current work is at risk of automation
  • Entry-level jobs most vulnerable
  • Business model needs fundamental change
  • Adaptation possible but painful

For SaaS:

  • Point solutions are vulnerable
  • Platform plays are safer
  • Infrastructure remains necessary
  • Value shifts from software to services + AI

For your career:

  • Automate-able work will shrink
  • Human judgment, relationships, creativity remain valuable
  • Master AI tools rather than fear them
  • Move up the value chain quickly

Final Thought

The stock market crashed 6% today. That’s real money, real pain.

But remember: Markets price in the worst case scenario first, then reality emerges slowly.

Indian IT isn’t dying. It’s being forced to evolve. Companies that adapt will survive. Those that don’t, won’t.

Same goes for individual careers.

The question isn’t “Will AI take my job?”

The question is “How do I use AI to become 10x more valuable?”

Answer that, and you’ll be fine.

Disclaimer: I’m an engineering professor who follows tech and markets closely. I’m not a SEBI-registered investment advisor. This is analysis and commentary, not investment advice. Indian IT stocks could bounce back or fall further - nobody knows for sure. Do your own research before making any financial or career decisions.

Data Sources:

  • BSE, NSE (Indian stock prices)
  • Reuters, Bloomberg, Economic Times (market analysis)
  • Anthropic official documentation (Cowork capabilities)
  • TechCrunch, Fortune, CMSWire (Anthropic coverage)
  • Business Standard, India TV (Indian market coverage)

All market data as of February 4, 2026, 2:00 PM IST. Stock prices and indices may have changed by the time you read this.

About the Author: Chandra Shaker is an Assistant Professor teaching Electronics & Telematics Engineering, with 14 years of teaching experience. He follows technology trends, markets, and their impact on careers and investments. Views expressed are personal.