It's not today... we have been challenging our teams on how you can deliver twice the revenue with half the people. We need to dramatically change from input base to more output and outcome based new ways. So if you look at AI, there's a massive productivity benefit. We're living in an era of uncertainty and revolutionary technology changes - AI, Quantum Computing, and so on.
CVK, let me start by talking to you. When I spoke with both of you in Davos in January, the mood was one of confidence and optimism. You said we are sitting on a time of demand explosion for Indian IT services. Would you continue that conversation? More importantly, the questions being asked today about the mix for Indian IT services - whether this business model will continue to be relevant - yes or no?
As I mentioned in Davos, I strongly believe the business model is ripe for disruption. In the last 30 years, we saw linear scaling of IT services with people and revenues. I think the time is already out for that model. In fact, in the last couple of years, we've been challenging our teams on how to deliver twice the revenue with half the people. That's the kind of call I've made. It's a big moment of disruption. All of us are seeing good demand, but that's a linear extrapolation of the current business model. What is fundamentally changing is something we need to recognize and react to.
That's a very important statement. I don't know how many caught that you said time is out for this business model. What are we talking about in terms of disruption and adaptability? Where does the industry find itself if you believe time is out for this business model?
Should I continue? Yes please, and then I'll get to Salil. Okay, so I think a large part of the industry is driven by an input-based model - time and material or number of people. We need to dramatically change from input-based to output and outcome based. Services need to become platform based. We've heard of software as a service, but service as a software is taking place dramatically. We need to be more proactive, not complacent about growth. The changes AI is ushering in are very different. We need to be proactive, even cannibalize our revenues to create new businesses. Having said that, I remain optimistic about how we can reinvent as long as we recognize the change is not the same as what we've seen before. I think we are setting ourselves on a good basis to reinvent our industry.
I'll talk about reinvention in just a second, but let me get Salil in on the conversation.
Salil, before we came in you said 'what's the problem? There is no problem.' If I look at what CVK is saying, he's clearly not suggesting we don't have a problem. In fact, I fully agree with CVK. I think the view we see is a huge shift in technology - generative AI, quantum, etc. We within Infosys and the Indian IT sector are ready to face that challenge. It's exciting because new work is coming from India. For example, in quantum we have work in drug discovery. We see examples in telco network optimization. This is all being done because of massive scale in India. I agree we have to be ready and not complacent. But do you believe with what CVK said that the time is up on the current business model?
No, what I feel is what CVK is saying is a call to not remain complacent. There are aspects of what we are doing today that are growing well and will continue, but we need to find new ways. If you look at AI, there's a massive productivity benefit - in software development between 7% and 15%. We use that in our Finacle product. We also see new work in credit analysis that we would never have done without generative AI. So there's a balance: new work expands the market, and productivity gives efficiency. Both have to be taken into account. That's how I understood CVK - make sure we're not complacent.
Let's talk about the reinvention you mentioned, CVK. For the benefit of everyone, let's start by talking about what you've already started to work on - the measures and steps you've taken as part of this reinvention and transition from software as a service to service as a software. How ready are you today and what will you need to do to make that transition?
The first and foremost change we are driving is the mindset change. We're all used to a certain mindset of delivering value to clients, but that paradigm is changing. You could deliver the same value with half the people. Then you have to force yourself to think: how can you use the unlocked human potential to create more growth opportunities? That's the mindset shift. Obviously, we have growing businesses and people with aspirations. How do you take care of people's aspirations while being proactive about adopting these technologies and bringing productivity? The productivity unlocked can be used to create more value and growth. For example, for a large financial services firm, we were delivering a transformation program. The client is spending maybe a billion dollars a year on a 5-year program. We believe it can be done in three and a half years. The value is not just the 18 months saved, but the time to value - whatever policy modernization they wanted to roll out. That's where the biggest value creation is possible.
Speaking of value creation, let me ask you this. We're talking about human genius and ingenuity, and moving from people to platform. What actually happens to the workforce? This sector has been driving employment for a young country like India. What happens to the people part of this business? What happens to entry-level jobs in the next 5 years?
First, for existing employees, there's huge reskilling. With generative AI and through digital, cloud, cyber, we have reskilled. At Infosys, over 270,000 people are more ready for AI. Second, in terms of new roles, we've shared that for the next financial year starting April, we will recruit 15,000 to 20,000 new campus hires. That shows the direction: there is growth, productivity, and new work from these technologies. All combined gives us confidence to have new people joining from campus.
Speaking of taking risks and bets, I want both of you to comment. We're talking about big numbers - $500 billion over the next four years. You're trapped in the tyranny of quarter-to-quarter existence. When you talk about big bets, you're not talking about incremental changes but big bold decisions. Do you have the capacity to bet on them with the kind of bold investments required? Will the markets allow you to do that?
It's a very difficult choice. At HCL Tech, we have taken bold bets. Six or seven years ago, we invested $4 billion in building a software business. As long as you can take bold bets, walk the journey, make them work, and mitigate risks, we can navigate it. Investors and the market should not be looking at how many freshers you are hiring. That's the wrong question. The right question is: how can you create more value out of maybe 200,000 people? Can you create twice the value with the same number of people? If you challenge yourself in that direction, the business will unlock a completely new growth trajectory. I'm betting on unlocking more human potential in different ways. Today India is the services capital of the world. Why can't India become the research capital of the world? Those are the levers that will change people, employment, and training dynamics. It needs different thinking.
We are fortunate as CVK mentioned that there is ability to take bets within Infosys and the IT sector. For example, we built a product like Finacle a few years ago, invested in an insurance product called Mish, and are working on an e-commerce product Equinox. In generative AI, we built four small language models - one for banking, one for IT operations. This is a material change in how clients will work with generative AI. We made a large investment, not publicly disclosed, but within the construct of a P&L. It's creating huge impacts. The small language model for banking already has clients using it with their existing large language models.
You talked about small language models. Let's get into the debate of whether we should attempt large language models. Salil, let me start by asking you: should India make the choice of going the large language model way, or should we build on our strengths on the application side? Here I'll leave it to CVK to answer.
No, no, no, you can't get away with that. You've got to say something at least. I think without a doubt, the cost of building models is falling. When the time is right, we should make the right investments and build the right small language model.
Okay, you said nothing through that. Yes or no?
I think from an India perspective, we absolutely have to make sure we are the leading light of generative AI. There's no other way. As costs keep falling, we saw investments closer to $5 million than hundreds of millions. That will give us a benefit. We need to make those sorts of investments to create the foundation model.
OK, CVK, straight answer: yes to large language models or no?
From a country perspective, I would strongly argue that we will need to build our own models. The rationale is that there are many open source models available, but we should not assume they will continue to be open source. These are going to be the coins on which geopolitics is played. DeepSeek is available today, but tomorrow it may not be. Similarly, OpenAI, Llama - there could be countries that limit their use beyond their boundaries. To have long-term competitive advantage, it makes sense to build. Costs are coming down, so we need to find ways to economically create the training infrastructure to train models.
Since you spoke about geopolitics, let me get both of you to comment on what's happening in the US, the largest market for IT services. Salil, let me start by asking you: we wake up every morning to what President Trump tweets or what Elon Musk says on Twitter. That dictates the news cycle. Is it dictating the business cycle as well? Has it started to dictate the business cycle?
What we are seeing so far - last week or before last, I was in the US meeting with several clients. There's a view that what the government is doing is benefiting US businesses through less regulation and policies that will help the economy. If that plays out, as the US and Western Europe economies become stronger, we will see more business for tech. There are two types: tech and ops on efficiency is going extremely well, and tech to build for growth - cloud, AI, etc. As economy numbers improve, we'll see more of that. That gives us a second leg of growth. We have one leg giving decent growth now. When the second leg comes, we'll see more growth. But I link this back to the need to adapt. In a Trump era with H1B issues, what could it mean for Indian IT in the next few years? We've built a more resilient business for the US. We've built local delivery and digital centers, over 60% of our US teams are local, and we have strong nearshore capability. We feel better prepared. So far, we've not seen any changes.
I agree with Salil. The business environment is becoming more vibrant. In the last two or three quarters, we've seen increasing demand despite all the conversations due to executive orders. On H1B, it's a much smaller issue than two decades ago. All of us have built more local capabilities and are less dependent on H1B.
Since we're talking about capabilities, I want to get both of you to talk about acquisitions. Both HCLTech and Infosys have done acquisitions. Do you believe that will have to continue to be the route, especially as we enter this turbocharged AI and quantum era? Salil, let's start: capability building inorganically - is that going to have to be the route for Indian IT?
I'm not sure it has to be, but it's definitely a huge attention for us at Infosys. Our balance sheet is strong, cash flow is strong, so we have huge opportunity to buy larger things. We did an engineering services acquisition a couple of quarters ago. Our pipeline is broad across industries, capabilities, and geographies. We feel pretty good. There's no timeline, but if you look at the trend over the last four years, each acquisition has been larger than the previous. We see more larger ones. Which one will happen when is not clear, but there's appetite.
I don't know about larger, but we see a need to look at capability-led acquisitions and create more partnerships with startups and ecosystem providers. That's how we add to overall capability.
Okay, let's talk about the next steps in this journey. It's not the strongest or biggest that will survive, but the most adaptable. What have you been able to do, what are you focusing on, what are you prioritizing to make sure you're adaptable and agile? Salil?
