AI is a disruption, a huge disruption. We saw this disruption in 2001, when the internet came. We had absurd valuations for new companies, and the earlier companies were beaten down. They said the world is going to change, and it took a lot of time for that to actually happen.
AI is different because machines have intelligence, and the speed of change could be faster. It is not that you can buy some AI like a cloud and now put it in the system, and automatically it starts working and doing things; it doesn't happen. All that AI has done is become a productivity tool.
And companies like Anthropic are using it cleverly because they are raising capital. And yesterday, they just raised $30 billion.
They released the latest Claude version a few days before the funding round closed. They want people to feel they are running the world, they are transforming it.
And some of these companies have been saying that it will destroy jobs for the last three years. Some of the jobs are going, but more jobs are being created; it is not happening at the speed that they are saying.
Even Palantir made a similar comment because Palantir has a rich valuation, and they need this kind of campaign to keep the price going. They are all playing to the market. And many people are also playing to the market, and the market is reacting to these claims.
We have seen this argument before, that software or AI will replace everything from brokers to doctors and lawyers. The whole thing with AI is nobody knows how it will run. They are trying to create this kind of hype.
AI is definitely a disruption. It will improve productivity. When you look at AI, there is accumulated software, and installed software is between $20 trillion and $25 trillion.
The world has been spending one and a half trillion dollars, maybe two trillion dollars every year for the last 15 years.
Second, large enterprises spend multi-billion-dollar budgets every year because they have complex legacy systems. Many global banks spend over $15 billion a year on such legacy systems.
Such large complex systems will have multiple databases, 12 to 13 databases for many of them. AI can assist with tasks or do discovery within one database, but it cannot work across multiple databases in large enterprises.
In core systems, when you write new code, AI can help in effort reduction by 30-40 percent. On the legacy code, AI cannot be of much help.
And the world and large enterprises run on legacy systems. And if somebody thinks that Claude will come and everybody will change SAP and their enterprise system and run, they are foolish.
The pace of change is going to be slower than what people imagine. Because there is risk management. They want to make sure the system does not break. There is a trust and confidence issue with AI, whether the data will go or disappear. It is still not mature. These large enterprises have customer data and are bound by regulations. So people are scared and are bound to be cautious.
Large enterprises are not going to just jump into it. Even today, global enterprises have not used much of AI in their work. What they have done is use it for discovery, research reports, and finding information. But they are not using it in their core business. And many of them have not been successful.
Over the last couple of years, most large enterprises have been watching. They have bought Claude subscription, but when they start using AI, who is going to help them use AI? Do they have internal capability?
They need all the service companies to come and help them use AI, to do applications faster, better and cheaper to run their systems for them and do all the changes that they want.
So these service companies are the ones who are going to make most global enterprises AI first.
And it's going to take seven, eight years. It is not going to happen overnight. When large enterprises start using AI for application development, quite possibly the time required, the effort required could come down by 20 to 40 percent. Now, the key thing is, that means a cost saving can be 10 to 20 percent because the cost of AI is very high.
You have to remember that Claude’s data centre cost is too high, and the subscription does not pay for the cloud. All the AI companies that are releasing new software versions are making huge losses. The annual loss on AI could be up to a hundred billion dollars.
The total investment in AI this year and next year could be two trillion dollars in the US. On two trillion dollars, they must get revenues of $500 billion minimum to be viable. Now the big five are putting $700 billion this year.
So whether this two trillion dollars is an overkill is something that is to be understood. And for all the investment the companies are making, the cost of using AI will go up from whatever it is today.
The full cost is not paid for by subscribers. They are subsidising their consumers. Sooner or later, Anthropic and Chat GPT will have to break even. They cannot burn money for long.
So the cost of AI is also going to be a part of the subscription cost. The key issue for enterprises is that when they start using AI for development, the cost will not come down because the company that does it for them is going to charge more because they have to reskill.
There could be 15-20 percent savings for new applications, which are maybe 25-30 percent of the revenues of IT companies. Now, they will use the money to redeploy into more AI development.
IT firms want to move away from legacy to an AI architecture, which takes time, more effort, and more investment. So any savings in the new application will go back for redeployment. That has been the trend over the last 20 years.
So that is going to happen. More work will come. Third, there's a lot of accumulated backlog of software for all business enterprises. There are existing software subscriptions and contracts.
Software needs many features, and these updates happen over the years. They will see savings, some savings in application, not in maintenance over time. It will take time.
IT service companies will see improvements in productivity. In the last three years, they have seen an improvement of 15- 20 percent. And next year, there will be more work on AI done by all these global firms. So to that extent, the fear is overblown. The productivity will be used by these service companies to do more work.
People said that the mainframe would disappear. Has the mainframe disappeared? Has legacy code disappeared? No.
It's going to take time.
IT firms are not going to see any immediate margin pressure, because they have been preparing for it in the last two years. They are becoming more and more efficient because clients are giving them more work.
AI is promising what SaaS also promised. The do-it-yourself model. What they are saying is you don't need any system integrators to do this work. They have been saying it for long. Has it worked? No.
SaaS has also grown a lot. Others have also grown. SaaS has become another tool to use.
See, the business complexity is huge. Business process is not standardised. People are not standardised. Businesses have history. Businesses have domain. All of it will not collapse.
Among the largest sectors, like healthcare and banking, these are all highly regulated sectors, and you can't throw away legacy systems overnight.
See, in areas like client customer acquisition, customer servicing, AI will have a big impact. They will be able to reach more customers. They will be able to lower the cost and grow faster.
That means the same base of employees can do 1x, 2x business. So the cost will come down overall, but I think employment will remain.
When we brought technology to banks, people said a lot of employees would lose jobs. But total employment in the bank kept going up, because the market also grew. This will happen in all industries.
Discover the latest Business News, Sensex, and Nifty updates. Obtain Personal Finance insights, tax queries, and expert opinions on Moneycontrol or download the Moneycontrol App to stay updated!
Find the best of Al News in one place, specially curated for you every weekend.
Stay on top of the latest tech trends and biggest startup news.
