AI Thematic Mutual Funds: The Risks Indian Investors Are Underpricing

Every bull market produces a fund category that looks like it cannot lose. In 2020 it was pharma. In 2021 it was IT. In 2022 it was PSU and defence. In 2026 it is AI. Indian AMCs have read the room: a wave of "AI", "next-gen tech", "innovation", and "Mag 7 access" funds have been launched or repositioned in the last 18 months, often as NFOs at ₹10 NAV, marketed with charts showing what NVIDIA did from 2022 to 2024.

If you are tempted, this article exists to ensure you go in clear-eyed. We are not anti-thematic — we are anti-naive. The risks below are not theoretical; they are the standard failure mode of every prior thematic cycle.

Risk 1 — Sector concentration is the point, and that is the problem

A diversified equity mutual fund holds 40–80 stocks across 8–12 sectors. A thematic AI fund deliberately concentrates in 20–40 stocks across 2–3 sectors (semiconductors, hyperscale cloud, AI software). That concentration is the entire selling proposition — and the entire risk.

The math is unforgiving. When the theme is in favour, concentrated funds outperform by 5–10% per year. When the theme falls out of favour, they underperform by 15–25% per year. Over a full cycle, the asymmetry rarely pays.

For context: between 2000 and 2002 the Nasdaq fell ~78% peak-to-trough. Investors who held US tech ETFs through that period needed 15 years to break even in nominal terms. That is not a minor drawdown — that is a decade and a half of opportunity cost.

Risk 2 — You are buying at the top of the narrative, not the top of the cycle

NFOs are launched when AMCs can sell them, which is the same moment retail euphoria peaks. The fund opens at ₹10 NAV, you put in ₹50,000, and the underlying stocks — already priced for AI dominance — proceed to either underdeliver or simply rerate to fair value. Either way you lose.

This is not a hypothetical pattern. Look at any sectoral NFO from 2021 (especially pharma and IT). Investors who bought at NFO are still underwater four years later, while the same investors who bought a plain Nifty 50 index fund are up 60%+.

The rule: Never buy a thematic NFO. If a theme is genuinely durable, the existing diversified fund or sector index ETF will give you exposure with less narrative tax. NFO unit costs are a feature for the AMC, not for you.

Risk 3 — The seven-stock concentration trap

If your "AI exposure" is a US-tech fund or a Nasdaq 100 feeder, you already own NVIDIA, Microsoft, Apple, Alphabet, Meta, Amazon, and (depending on classification) Tesla. The Mag 7 dominate the S&P 500 and completely dominate the Nasdaq 100.

If you then add a dedicated "AI thematic" fund, you are buying the same seven stocks again — at higher fees, with worse diversification. The marginal addition to your AI exposure is small; the marginal fee + tracking error is large.

Before buying an AI thematic fund, open the portfolios of every equity fund you already own and tally the Mag 7 weight. If it is already above 15% of your equity portfolio, an AI thematic adds risk without adding exposure.

Risk 4 — Equity-looking, non-equity-taxed (again)

Most Indian AI thematic funds are structured as fund-of-funds into overseas AI/tech ETFs. That makes them, for Indian tax purposes, non-equity schemes — taxed at your slab rate with no LTCG concession and no indexation.

If you are in the 30% slab and your AI fund returns 15% per year, your post-tax CAGR is roughly 10.5%. The same 15% pre-tax return from a domestic equity fund leaves you with ~13% post-tax (assuming LTCG at 12.5% on gains above ₹1.25 lakh).

That 2.5 percentage points per year of tax drag compounds to a massive difference over 10–20 years. The fund has to outperform an equivalent Indian equity fund by 2.5% per year just to break even on tax. Most thematic funds, historically, do not.

For details, see our taxation primer.

Risk 5 — Recency bias is dressed up as "secular thesis"

Every thematic pitch sounds the same: "This is not a cycle, this is a generational shift. Mobile, cloud, now AI." Sometimes the thesis is correct (mobile and cloud were, broadly, generational). Sometimes it is wrong (Web3, metaverse, electric two-wheelers as a thematic). And even when the underlying thesis is correct, the stocks that win are rarely the ones the thematic fund holds at launch.

The 2001 internet bust was not a referendum on whether the internet would be big. It was a re-pricing of who would capture the value. Cisco was 90% of the internet's plumbing in 2000 and is roughly flat 25 years later. The internet won; the early internet stocks lost.

The same is plausible for AI: the underlying technology is transformative, but the specific NVIDIA + Microsoft + a-handful-of-others basket may not be the right way to bet on it. A diversified large-cap fund that gradually rebalances toward AI winners as they emerge is structurally more robust than a fund that locks in today's winners.

When does a thematic AI fund actually make sense?

There is a narrow case where it is defensible:

1. You have a fully built diversified portfolio (asset allocation, emergency fund, term insurance, health insurance, retirement on track).
2. You have a specific thesis on which sub-theme of AI (semiconductors, applied AI software, AI infrastructure) you believe is underpriced and not adequately captured by your existing US-broad or Nasdaq holdings.
3. You can size the position at no more than 5% of your equity bucket — small enough to lose without affecting goals.
4. You can hold for at least 7 years to amortise both the volatility and the tax drag.
5. You are not buying at NFO. You are buying an existing fund with at least a 2-year track record and visible portfolio.

If any one of those five conditions fails, you should not own this fund. Buy a plain large-cap index fund and move on with your life — you will almost certainly do better.

What the data says about thematic funds historically

A broad observation across global research (not specific to India, but consistent across markets): thematic equity funds, in aggregate, underperform their broad-market index over 7+ year periods by 1.5–3 percentage points per year after fees. The few that outperform are not predictable in advance — survivorship bias makes the post-hoc winners look skillful, but pre-selection success rates are roughly random.

For Indian-domiciled global thematic funds, the picture is worse because of the tax drag described above.

The honest alternative

If you genuinely want AI exposure in a way that is robust:

1. Own a low-cost US broad-market index fund (S&P 500 or total-market FoF). This gives you ~30% AI/tech exposure embedded, with the rest diversified. As AI winners emerge, the index rebalances toward them automatically.
2. Optionally, a small allocation to Nasdaq 100 FoF. Higher tech tilt without single-stock-narrative risk.
3. A diversified Indian flexi-cap or large-cap fund. Indian IT services companies (TCS, Infosys, Wipro) get most of the second-order AI services revenue — you are already getting Indian AI exposure here.

This stack will give you 80% of the upside of a successful AI bet with 30% of the concentration risk and zero NFO timing risk. See our direct vs regular primer to ensure you are paying the lowest possible fee on each of these.

Worked example: the real cost of the tax drag (hypothetical)

All numbers below are hypothetical and for illustration only. No specific fund or NAV is implied.

Assume two investors, Priya and Rahul, each put ₹5 lakh into funds that both earn 15% per year pre-tax over 10 years.

Priya buys a domestic equity flexi-cap fund (equity fund, Direct plan). After 10 years:

• Corpus: ₹5,00,000 × (1.15)^10 = approximately ₹20,23,000
• Gain above ₹1.25 lakh annual exemption: for a lump-sum investor, the full ₹15,23,000 gain is realised at redemption. LTCG tax at 12.5% on ₹15,23,000 = approximately ₹1,90,375.
• Post-tax corpus: approximately ₹18,32,000

Rahul buys an overseas AI FoF (non-equity fund, same 15% pre-tax CAGR). After 10 years:

• Same pre-tax corpus: ₹20,23,000
• Entire gain of ₹15,23,000 taxed at his slab rate. At 30% slab: tax = approximately ₹4,56,900.
• Post-tax corpus: approximately ₹15,66,000

The tax drag alone costs Rahul approximately ₹2,66,000 — more than half his original principal — compared to Priya holding the identical pre-tax return in an equity fund. The AI fund has to outperform the domestic fund by enough to close that gap before Rahul gets ahead. At 10 years, that required outperformance is roughly 3% per year compounded. Historically, most thematic funds do not deliver that.

This is not a tax technicality. This is the single largest structural drag on overseas thematic funds for Indian investors, and almost no AMC marketing material mentions it clearly.

Common mistakes investors make with thematic funds

1. Treating "AI" as a monolith. AI is a capability, not an industry. The companies that profit from AI include semiconductor fabs, hyperscalers, enterprise software companies, and downstream service companies. A fund marketed as "AI" may be concentrated entirely in US semis — a sub-theme with its own cycle, geopolitical risk (Taiwan, export controls), and valuation dynamics. Understand what the portfolio actually holds, not what the fund name says.

2. Anchoring on NVIDIA's 2022–2024 return. Fund marketing shows NVIDIA's chart. But NVIDIA was also down 65% from peak to trough in 2022. Investors who bought at the 2021 peak waited until early 2024 to recover. The same stock. If you could not hold through that, you could not have captured the famous gain.

3. Using thematic funds as a diversification tool. "I already have a large-cap fund, so this adds something different." A concentrated theme is the opposite of diversification — it adds a correlated cluster risk. During a tech rout, your AI fund, your Nasdaq FoF, and your flexi-cap (which holds IT stocks) will all fall together. The portfolio behaves like one concentrated bet.

4. Ignoring the fee multiple. A typical overseas AI FoF charges 1.0–1.5% TER in Direct plan, on top of the underlying ETF's fee (0.2–0.5%). You are paying 1.2–2.0% total for a passive basket. A domestic index fund charges 0.05–0.10% for similar market-cap exposure to similar companies. Over 15 years, the fee gap alone erodes 20–30% of ending corpus.

5. Conflating "AI will be big" with "this fund will outperform." Even if AI transforms every industry — a plausible thesis — the gain may accrue to consumers, not shareholders, if competition drives down margins. Or the gains go to companies not yet public. The thesis being right does not guarantee the stocks in today's AI fund will outperform.

Questions advisors don't answer honestly about AI thematic funds

Q: My advisor says AI thematic funds give "international diversification." Is that true?

No. Real international diversification means exposure to currencies, economies, and sector cycles that are uncorrelated with India. An AI thematic fund concentrated in US tech moves closely with the Nasdaq. During a global risk-off event, Indian equities and US tech fall together. You are not diversifying — you are adding a correlated high-volatility cluster.

Q: The fund's 2-year return is 45%. Shouldn't I ride this momentum?

The 45% return is behind you. What you are buying is forward return, which depends on what those underlying stocks are priced at today, not what they were in 2022. When every investor already believes AI stocks will keep rising, that belief is already in the price. Momentum works until it doesn't, and the reversal in concentrated thematic funds is fast and severe.

Q: If the AI theme underperforms, can't I just switch to another fund?

Yes, but every switch in a non-equity fund (which most AI FoFs are) triggers a taxable event at slab rate. You are not just taking a performance loss — you are also crystallising a tax event. Frequent switching in non-equity funds is one of the most tax-inefficient strategies available.

Q: The AMC says this fund is "suitable for investors with 3+ year horizon." Is that accurate?

No. Any fund with 60–80% concentrated in tech themes should have a minimum 7–10 year horizon. The 3-year framing is a compliance checkbox, not genuine guidance. Tech downturns can take 3–5 years to recover. A 3-year horizon investor who buys at the wrong point in the thematic cycle has no buffer time.

Q: Why don't distributors mention the tax drag upfront?

Because the fund is still sold primarily in Regular plans through distributors, and the higher TER means higher commission. The tax drag is structurally invisible to an investor focused on headline NAV returns. Ask any distributor recommending an AI FoF: "What is the total cost — fund TER plus underlying ETF TER — and how does the after-tax CAGR compare to a domestic equity fund at the same pre-tax return?" If they don't answer that precisely, you have your answer.

Action checklist

1. Open every equity fund you already own. Tally NVIDIA + Microsoft + Apple + Alphabet + Meta + Amazon weight.
2. If that tally is above 15% of your equity portfolio, you do not need an AI thematic fund.
3. If you still want one: avoid NFOs, cap allocation at 5%, hold 7+ years, pre-commit to not selling during the first 30% drawdown.
4. Re-read the five conditions above. If you cannot honestly check all five, walk away.
5. Allocate the same money to a US broad-market index FoF instead.
6. Run the tax drag calculation above for your own slab rate before committing.

Themes are the easiest sale in mutual funds and the easiest way to underperform a boring index. Be the investor who buys the boring index.

Mutual fund investments are subject to market risks. Read all scheme documents carefully. This article is for educational purposes and is not investment advice.