FPI return to India a question of when, not if: Amundi CIO Vincent Mortier

The rise in long-term US interest rates due to massive public borrowing and artificial intelligence (AI)-related expenditure could preclude the Fed from raising short-term rates further as the rise at the longer end of the yield curve is already impacting housing and rent markets, said Vincent Mortier, group CIO of France's Amundi, Europe's largest portfolio management firm. Amundi has €2.6 trillion in assets under management and is an investor in SBI Funds Management Ltd, India's largest asset management company.

Mortier expects India's strong macroeconomic fundamentals to draw back foreign investors in the likely event of a moderation in the AI earnings supercycle and oil prices due to a resolution of the West Asian and Russia-Ukraine conflicts.

Here are some edited excerpts from an interview.

Even as US equity markets trade near record highs, markets like India are taking a hit from elevated interest rates in advanced economies. What is your view?

When you look at the forward estimation of inflation in the US, which is sitting on a mountain of $40 trillion of public debt, it's around 2-2.2%. So the market is not buying the idea that inflation will last forever. The rates have increased quite rapidly at the short end of the curve. So I don't believe the Fed will continue to hike a lot. I think it will pause. I don't believe it (rate hike) will happen because inflation will go down on account of (higher) productivity due to AI and because persistently higher rates at the long end of the curve will put some pressure on housing and rent and do the tightening work for the Fed.

The long end of the curve has seen rates rising because of heavy public borrowing and demand by corporates for AI-related business. So you have borrowers competing and heavy supply (of bonds) expected at the longer end. This will cause the dollar to weaken and benefit emerging markets like India.

But the conflicts in Europe and West Asia are raging on , driving up energy prices…

Hopefully, energy prices will go down in the next month if the Ukraine and Middle East conflicts are resolved. So, we might have pleasant surprises on inflation.

You also said that inflation would moderate because of AI. How would that play out?

AI has had a big impact on US inflation because of the demand for memory chips, which is resulting in price increases and big investments in data centres. But then, in a year's time, maybe memory chips will decrease a lot in price because new capacities are being created.

Then there is innovation from China as well. Consequently, the marginal cost (the cost of producing additional quantities) of producing companies, which increasingly use AI for many tasks, would drop to zero. All this would give rise to deflationary pressure because it would cost much less to produce some services and there would be some impact on the job market. Salaries may not increase a lot and you could end up with fewer well-paying jobs. The Industrial Revolution hit blue-collar jobs. The tech revolution will hit white-collar employees. This will also add to deflationary pressure.

FPIs have been selling Indian equities for the past two years in favour of the AI play. Do you foresee a reversal of these outflows?

It's a question of when, not if, that will happen. Investors' earnings expectations from AI companies are rising a lot. At the end of the day, the massive investments in LLMs (large language models) and hyperscalers are on the balance sheets of these companies and these investments are amortized. So, you have a big disconnect between accounting earnings and cash flows.

That's an issue because in the next few years you will have the amortization of the previous investments, which will impinge on earnings. So, in order to continue to grow earnings, you really need to have super-strong revenue growth to compensate for the amortization of trillions of dollars on the balance sheet. That is possible if the number of tokens (AI computing measure) and their price increases. But the price of tokens will come down because of the additional capacity being created, as we discussed earlier.

Also, companies using AI are asking: "Do we really need so many costly tokens to perform tasks that can be performed by a standard model?" So, the price of tokens will gradually decline, which is good news. But this could be a challenge for the 15-20 names that have driven the US markets on the basis of superlative earnings expectations.

So you're saying that a pause in interest-rate cycle in the US and the challenges to the AI supercycle would eventually result in FPIs returning to India?

The AI dynamic will change and losers will become winners because they will be the ones who are using AI in their own real economy and not the ones who have invested into the chips. When you look at GDP growth of emerging economies, particularly India, it's way better than anywhere else. And when you look at the public finances, I mean, if you take a blind test… Say you take a few countries. You don't show the name. You just put the metrics: debt to GDP, deficit, current account, reserves—all the economic metrics. I mean, you will say, okay, India is developed, while the US is developing.

If you're in equities for the long-term, saving up for retirement, you aren't even in a bear market in India yet. We need to step back and relax as history has shown that even more severe bear markets with 40-50% drawdowns have bounced back from trough to peak in a few years. If you backtest it, the recovery happens in 8 to 10 years at the most.

The US has barred big Indian outsourcing giants from the employment-based green card programme. How will this affect our IT firms?

That’s a good question as the first reaction would be to say that AI will be negative for IT software services. The reality is more nuanced. What we see is that some software companies are able to incorporate AI into their offering and to offer other services customers value.

So it's not only negative or only positive. It will really depend on each company, and this is where I believe some Indian companies will benefit—by embedding AI into a superior value proposition. Some will be challenged, and while they won’t disappear, growth will be more difficult. So I think performance across the sector will be more dispersed.

As Europe's biggest asset manager, what's your own bias? Do you invest in AI companies?

We get a mandate from our clients and our funds are benchmarked to an index like MSCI for global equities. A lot of AI companies make up these benchmarks, so even though we are underweight these companies, we have them as part of our funds. As more and more people are buying indexes or we are measured against the benchmark, we tend to buy all the same things.