We asked top AI models their best market picks for 2025. And were surprised....

— Sam Altman, CEO, OpenAI

Who do you think is the greatest investor of all time? Warren Buffett? George Soros? Charlie Munger? All great choices, but fall way short. The greatest money manager the world has ever seen is a name not many are familiar with, at least in this part of the world.

Jim Simons, who passed away in May 2024, was a renowned mathematical physicist who made invaluable contributions to string theory. He also set up the super-successful hedge fund Renaissance Technologies or RenTech, which is unlike any firm on Wall Street. RenTech is staffed with PhDs in fields like mathematics, physics and statistics, but not finance. In fact, it avoids hiring anyone with even a whiff of Wall Street credentials. RenTech’s flagship Medallion Fund was set up in 1988. Over the next more than three decades, it generated mind-boggling annualized returns of 62% (before fees) and 37% returns net of fees.

A sum of $10,000 invested in the Medallion Fund at its launch would turn into a whopping $420 million by 2021, compared to just $400,000 in the S&P-500 or $1.5 million in Warren Buffett's Berkshire Hathaway.

Simon and his team of geeks used complex mathematical models trained on vast amounts of data to predict the price movement of securities. While many other quantitative trading firms employ the same techniques, what set RenTech apart was both the breadth of data it mined and the robust (and secretive) mathematical models it ran on powerful supercomputers to gain an edge over every buyer and seller in the market.

Some consider Simons to be the first person to truly monetize machine learning (ML) on Wall Street.

With the world at the cusp of a new era of artificial intelligence (AI), many people are salivating at the prospect of harnessing the power of AI, ML and Big Data, which were hitherto the exclusive preserve of the intellectual elite.

Of course, for those looking to experiment with AI in investing, expecting to become Jim Simons overnight would be a fool’s errand. Also, following the basic tenets of statistics, if everyone has an edge, then no one has an edge, and we are back to square one.

That being said, the rapid progression in AI models’ capabilities and their jaw-dropping output across a wide range of fields makes one wonder how they would approach the dicey world of Dalal Street.

Model portfolio – AI edition

2024 saw real estate, healthcare and consumer indices topping the sectoral charts. But which Indian sectors are expected to emerge as the winners in 2025?

To find out, we ran multiple prompts on three popular generative AI chatbots – OpenAI’s ChatGPT, Grok (developed by Elon Musk-led xAI), and Google’s Gemini 2.0. For ChatGPT, arguably the most widely used AI tool in the world today, both the free version and its top priced variant (costing $200 a month) were used.

All chatbots added the usual caveats to their answers – that predicting stock market movements is inherently a speculative exercise and market conditions can change due to a host of factors, including economic policies, global events, and technological advancements.

Here’s a distillation of their responses to questions regarding which sectors in India to invest in for 2025:

Grok

A) IT: There is anticipation for strong performance due to the ongoing digital transformation and increasing demand for IT services, both domestically and internationally.

B) Financial Services: Including private sector banks and other financial institutions, this sector is expected to do well with India's economic growth and increasing financial inclusion.

C) Consumer Durables: Continued growth in urban and rural consumption, driven by rising disposable incomes and consumer spending, suggests this sector might remain a top performer.

D) Healthcare: With an ageing population and increasing healthcare awareness, alongside growth in pharmaceuticals and medical equipment, healthcare is poised for significant gains.

E) Realty: The sector could see continued growth with the stabilization of interest rates and ongoing urbanization trends in India. 

F) Hotels, Hospitals, Recycling, Waste Management, CDMO (Contract Development and Manufacturing Organization), Power Transmission, and Textiles: These sectors were specifically highlighted in recent discussions as sectors to watch for potential outperformance in 2025.

Gemini 2.0

Apart from IT and financial services, it endorsed two additional sectors:

A) Industrials: The government’s focus on infrastructure development is expected to drive growth in the industrial sector. Initiatives like ‘Make in India’ could further boost manufacturing activity and benefit industrial companies.

B) Chemicals: Growing domestic demand from various end-user industries is expected to support the chemicals sector. The sector also has export potential, which could contribute to its growth, it said.

ChatGPT

The free version of the tool had arguably the weakest set of responses. It usually just collates articles from different news sites (including Mint) and links back to their original source. There was no clear confirmation of its own preferences and thinking process.

However, once we got to its top-of-the-line Pro version, the results got more engaging. Here are its top choices for the new year and its investment rationale in short:

A) Nifty Bank, Nifty Financial Services: India’s expanding middle class and increasing formalization of the economy support robust credit demand. Over the past decade, financial penetration has grown significantly, with banks and non-banking financial companies (NBFCs) tapping into previously unserved or underserved markets. By 2025, continued improvements in financial inclusion, digital payment ecosystems, and lending platforms could drive earnings growth.

After a multi-year period of deleveraging, many Indian corporates have stronger balance sheets, improving the quality of bank credit portfolios and potentially leading to lower non-performing assets (NPAs) over time.

B) Nifty IT: India’s IT sector is closely tied to global technology spending. With enterprises worldwide continuing to invest heavily in cloud computing, artificial intelligence, machine learning, cybersecurity, and data analytics, well-established IT service providers stand to benefit. India’s IT firms are increasingly moving up the value chain, engaging in consultative, high-margin digital transformation projects. By 2025, this shift could reinforce revenue visibility and operating margin resilience.

C) Nifty Infra: The Indian government’s focus on infrastructure—ranging from roads, railways, ports, and airports to smart cities—underlines a multi-year capital expenditure cycle. Infrastructure development creates downstream benefits for related industries such as cement, steel, construction, engineering, and heavy machinery. An upswing in infrastructure investment by 2025 could translate into sustainable earnings momentum for companies within this sectoral basket.

D) Nifty Auto: The automotive sector stands at the cusp of a generational shift toward electric vehicles (EVs) and cleaner mobility solutions. Government incentives, declining battery costs, and rising consumer awareness should support EV adoption.

As global manufacturers diversify supply chains away from concentrated geographies, India’s component suppliers and OEMs could increase their share in global automotive value chains. The country’s established expertise in small car manufacturing and ability to produce at scale offers a comparative advantage. Improved rural incomes and changing mobility patterns (the “personal vehicle" push post-pandemic) may sustain domestic demand growth.

E) Nifty Energy / Renewable-Focused Stocks: With India’s commitment to achieving ambitious renewable energy targets (450 GW+ of renewable capacity by 2030) and global ESG (Environmental, Social, Governance) investment flows, companies aligned with solar, wind, and green hydrogen are well-positioned.

Interestingly, ChatGPT Pro also generated a short section titled ‘Potential Dark Horses’:

F) Nifty Pharma: Continued demand for generics in developed markets, rising healthcare expenditure domestically, and increased R&D capabilities can revitalize pharma performance. While the sector has been cyclical, long-term secular tailwinds remain intact, particularly if companies move up the value chain into complex generics and speciality products.

G) Nifty FMCG: India’s demographic dividend, urbanization, and premiumization trends could keep FMCG demand resilient. Though valuations tend to remain elevated, stable cash flows and brand strength may support steady long-term returns.

Peering into the black box

What to make of these results?

For one, diversity of opinion is as much a feature of humans as it is of AI models.

“We do not expect human beings to give us consistent answers for a query, and the same holds true for AI," Daniel Lemire, computer science professor at the Université du Québec and among the world’s top 2% scientists (according to Stanford rankings of 2024), told Mint.

“As for the source of the differences, there are different training data, and models are complemented with different runtime sources. Indeed, they benefit from a form of RAG (retrieval-augmented generation) compared to conventional search engines. Also, even with the exact same prompt, large language models do not, in general, provide the exact same answer," he added.

How a user forms the question (prompt engineering, in AI lingo) is also a key determinant of the final output.

“For example, we noticed that when we asked ChatGPT about the sectors to invest in 2025, it referenced Mint, Reuters, YouTube videos, etc. While Grok referenced many websites including groww.in, investindia.gov.in, indmoney.com, 5paisa.com and many social media posts from x.com (formerly Twitter). Gemini referenced different sets of websites and articles," Aniruddha Chakrabarti, partner, AI and Cloud, Grant Thornton Bharat, told Mint.

Another point to note is that two sectors featured in the answers of all the three models—financial services and IT. India’s financial services stocks are among the traditional favourites of foreign investors, while the country's IT sector commands the lion’s share of the global outsourcing market.

Does the recommendation of these two sectors indicate that AI models have ingested more data and headlines on these two sectors compared to the others?

“I would say that you definitively cannot trust the models to be unbiased," Lemire said cryptically, but did not elaborate.

Chakrabarti was more forthcoming.

“Apart from financial services and IT, we also noticed that most of the popular gen AI models and chatbots mentioned renewable energy and electric vehicles. These gen AI models depend on publicly available internet data—and since many financial advisers and common users are more aware of these two sectors, there’s more mention about them on various websites. So, these chatbots and AI models are picking up these two sectors more," he said.

Human element

‘AI will make humans redundant’ looms large in the feverish imaginations of technophobes, socialists and other assorted Luddites. But if the results of this experiment are anything to go by, that day will take a long time to come.

“Looks like AI models are recommending pretty much all the sectors, isn’t it?" chuckles Nirav Karkera, head of research at fintech platform Fisdom. “It is largely a play on the macros, you can’t go wrong there. It seems to be an aggregation of all headlines off the web. It looks like a consensus exercise, not an analytical exercise," Karkera told Mint.

From an investor’s point of view, a more granular approach is needed to outperform the peers.

“For example, financial services is too broad a theme. Within financial services, we like asset management companies (AMCs). There are limited listed players, but there are sectoral tailwinds like rising capital market penetration and favourable operational metrics. If an AMC has AUMs of ₹2 trillion, it can double it to ₹4 trillion without having to deploy excess capital. Insurance is also a good bet, though valuations are on the higher side currently," Karkera said.

“Similarly, in the IT space, we like the larger players as revenue has started to trickle in from their large order books, especially in areas like digitalization and AI. The US is pressing the pedal on the growth front, which too would benefit the IT players," he added.

Infra, industrials, etc. are all overplayed themes right now, Karkera further said. Yes, the macro tailwinds are there, but thanks to the stretched valuations, the margin of safety is very less. “We can say that all good news has already been priced in, and any hint of trouble can trigger a tsunami of selling, so investors should be careful."

Many experts say that the next gen tools like AI should be used by common investors not to replace financial advisers but to augment them.

“We suggest a two-phased approach. In the first phase, AI models and gen AI chatbots should be used for the initial research and shortlisting of sectors, stocks and mutual funds. Then, in the second phase, responses and suggestions from AI models should be validated (as AI models can sometimes give wrong responses, which is also known as hallucinations) and then further optimized to suit the requirement of the particular user," Chakrabarti opined.

Perhaps a fitting end to this story would be what ChatGPT reminded us after every prompt.

“Actual outcomes (of the predictions) will depend on macroeconomic stability, geopolitical factors, execution of government initiatives, and global market conditions. As always, a diversified approach and regular portfolio reviews are prudent."