PrimeIdea Ventures
  • Home /
  • Blogs /
  • Why US Earnings Growth Can’t Be Compared With India’s

Why US Earnings Growth Can’t Be Compared With India’s

PRIME ProspectiveUpdated October 31, 2026
Why US Earnings Growth Can’t Be Compared With India’s

Growth is growth. But a developing nation of 146 crore people is not a one-theme market.

The comparison doing the rounds is simple. S&P 500 EPS grew ~53% in April–June 2026. The Nifty managed 13–25%. So capital goes to America.

Yes, earnings growth is earnings growth. A rupee of profit and a dollar of profit both count. If we compare only growth rates, the comparison stands. But the moment part of that growth is circular, the two numbers stop being comparable. To see why, look at how the US number was made, and at who India has been compared with before.

India has always been compared with someone

A few years ago it was Korea, Brazil, Indonesia and South Africa. Today it is the US. Look at what those markets have in common: a handful of companies, often one theme, carrying the whole index.

Korea’s market leans on two memory-chip makers, Samsung and SK Hynix, at ~40% of the index. When doubts over the memory cycle hit in mid-2026, the KOSPI fell ~27% in under a month. Indonesia’s index, heavy with a few banks and conglomerates, lost ~35% in six months on doubts about free float and disclosure. Brazil fell ~14% in a single session on 26 October after its election. South Africa leans on one tech holding group for about a quarter of its market. Annexure F

The triggers differ: memory chips, governance, politics. The weakness is the same. When a market rests on a few pillars, one crack shakes the whole building.

How the US number was made

Out of 500 companies, the profit engine is one idea: AI. 21 AI-linked companies hold ~34% of the S&P 500’s value and earn ~38% of its profit. Annexure A That is not 500 companies growing together. It is one theme carrying the index, which is exactly the Korea pattern on a bigger stage.

Inside that theme, four Big Tech firms sit in a web of deals with the AI start-ups. Annexure B Strip away the press releases and there are only three ways it lifts their EPS:

1Sell to the company you fundedInvest in an AI start-up; it commits to buy your cloud. The money goes out as equity and comes back as revenue.
2Mark up the company you ownWhen the start-up raises money at a higher valuation, partly because of these very deals, the investor books the increase as profit.
3Fund the buyer of your productA chip maker pledges billions to customers who then order its chips.

Now do the division. In April–June 2026, ~$160 bn of Big Tech profit was paper gains on AI stakes, against roughly $864 bn of total S&P 500 profit. That is about 1 in every 5 dollars. Annexure C Not all US EPS is propped up. But a meaningful slice is, and it sits inside the companies that carry a third of the index.

The cracks are already showing

We are not saying the US market will fall or rise. But the AI trade is starting to wobble, and the wobble begins exactly where the paper gains were.

This week Alphabet reported its July–September results. Profit fell 25%. The same investment book that produced huge gains in April–June produced an $8.7 bn loss. Capital spending jumped to $51.4 bn, more than the cash the business generated. The share had its worst day on record. Annexure E

The next morning, Amazon and Intel flagged heavy cash burn, Nasdaq futures dropped about 2.5%, and the Nasdaq closed October down more than 5%. The market is now asking a simple question: show us the revenue.

That is the second lever running in reverse. Paper gains lift EPS on the way up. Paper losses cut it on the way down.

Who wins, and who pays

The winners today are the sellers of chips and cloud. Their order books are real and full. The payers are the AI model makers at the end of the loop. Their spending becomes everyone else’s revenue, and they are not yet earning it back.

Big Tech plans $660–690 bn of AI capex in 2026. The largest AI model maker lost more than it earned last year. Annexure D If their revenue catches up, the loop holds. If not, valuations fall and orders slow.

Same growth number, different engines: the US circular AI investment loop beside India's seven broad economic pillars
Same growth number, different engines: the US AI investment loop on the left, India’s broad pillars on the right. Schematic, not to scale.

India is a different economy

India is the world’s most populous nation, and still developing. It has ~146 crore people, consumption at 61.5% of GDP, and income per person roughly one-thirtieth of America’s. Annexure H

That gap is the growth engine. Profits here come from first home loans, first two-wheelers, insurance policies, power connections and highways. Banks fund it, energy and metals feed it, autos and consumer companies sell into it. Services exports and a large, young, skilled workforce add to it. None of it depends on companies billing each other.

And the profit base is widening. The Nifty 50’s share of Nifty 500 profit fell from 87% to 51%, and mid- and small-caps delivered about two-thirds of FY26 profit growth. Annexure I

To be fair to both sides

US earnings are not fake. The other 493 companies also grew strongly, the chip and cloud sales are real, and the paper gains are legal and disclosed.

India is concentrated too, mainly in banks. Annexure G But banks lending to a growing economy is a very different risk from companies funding each other’s customers. And India has had its own bad year in some large stocks. A broad market does not mean a market that never falls.

What this comparison leaves out. EPS and market value alone ignore brain drain, dependence on foreign technology and data, and the job pressure of a young workforce. Judging India on two numbers is unfair.

What to watch

  • Investment gains and losses in Big Tech results. If they keep swinging profit, earnings quality has not improved.
  • Proof of AI revenue. The loop holds only if end-users start earning what they spend.
  • The big AI listings. Public prices will test the private valuations behind last quarter’s gains.
  • India’s FY27 earnings. Broad credit and consumption growth must show up in the numbers.
VerdictGrowth rates can be compared. Circular growth cannot. Compare what earnings are built on, not just how fast they grow.

Annexures

Data referred to in the article. Tap any “Annexure” link above to jump here.

Annexure A · AI-linked companies in the S&P 500

Share of S&P 500 market valueShare of S&P 500 profit (last 12 months)
Company Value % Cum. Profit % Cum. Chart
NVIDIA 7.85% 7.85% 7.34% 7.34%
Microsoft 5.96% 13.81% 5.42% 12.76%
Alphabet 4.51% 18.32% 8.97% 21.73%
Amazon 3.59% 21.91% 5.13% 26.86%
Broadcom 2.47% 24.38% 1.44% 28.30%
Meta 2.08% 26.46% 2.85% 31.15%
Micron 1.59% 28.05% 3.23% 34.38%
AMD 1.39% 29.44% 0.24% 34.62%
Palantir 0.67% 30.11% 0.11% 34.73%
Intel 0.65% 30.76% Loss 34.73%
Oracle 0.60% 31.36% 0.72% 35.45%
Applied Materials 0.50% 31.86% 0.35% 35.80%
Lam Research 0.48% 32.34% 0.32% 36.12%
KLA 0.33% 32.67% 0.20% 36.32%
Sandisk 0.30% 32.97% 0.42% 36.74%
Marvell 0.29% 33.26% 0.10% 36.84%
Arista Networks 0.28% 33.54% 0.15% 36.99%
Seagate 0.24% 33.78% 0.17% 37.16%
Dell 0.21% 33.99% 0.42% 37.58%
Western Digital 0.20% 34.19% 0.38% 37.96%
Vertiv 0.14% 34.33% 0.08% 38.04%
Total (21 companies) 34.3% 38.0%

Market value as of 29 Oct 2026. Intel made a net loss over the last 12 months.

↑ Back to the article

Annexure B · The AI deal loop

Investor / seller Money going out AI company Money coming back
Microsoft ~$19 bn invested since 2019 OpenAI ~$250 bn Azure cloud commitment
Microsoft Up to $5 bn Anthropic $30 bn Azure cloud commitment
NVIDIA Up to $100 bn OpenAI Purchases of NVIDIA chips
NVIDIA Up to $10 bn Anthropic Use of NVIDIA systems
NVIDIA ~7% stake + $6.3 bn deal CoreWeave Purchases of NVIDIA systems
Amazon $8 bn (2024) + $5 bn, up to $20 bn more (2026) Anthropic ~$100 bn AWS compute over a decade
Alphabet $10 bn, up to $30 bn more (2026), ~14% stake Anthropic Up to 5 GW of Google TPUs over five years
Amazon (AWS) — OpenAI $38 bn AWS deal over seven years
Oracle ~$40 bn of NVIDIA chips OpenAI Data-center capacity for OpenAI
OpenAI Chip purchases AMD Right to take up to 10% of AMD

↑ Back to the article

Annexure C · Paper gains inside Big Tech profit, Apr–Jun 2026

Item $ bn
Alphabet: other income (mainly SpaceX, Anthropic stakes) 97.9
Amazon: non-operating income (mainly Anthropic stake) 53.4
Microsoft + NVIDIA: gains on stakes (combined, approx.) 8.7
Four companies: total “other income” from AI stakes 160+
Same four companies, Jan–Mar 2026 69
Magnificent 7: total net profit 315.6
Magnificent 7: of which investment gains 134.6
Magnificent 7: share from investment gains ~42%
S&P 500: total quarterly profit (approx.) 864
AI-stake gains as share of S&P 500 profit (approx.) ~18%

↑ Back to the article

Annexure D · AI spending vs AI earnings

Item Figure
2026 capex plans: Amazon, Alphabet, Meta, Microsoft, Oracle $660–690 bn
OpenAI revenue, 2025 $13.1 bn
OpenAI operating loss, 2025 $20.9 bn
OpenAI revenue, Apr–Jun 2026 $6.7 bn
OpenAI operating loss, Apr–Jun 2026 $12.3 bn
OpenAI planned compute spend, 2026–30 ~$856 bn
Anthropic revenue, Jul–Sep 2026 $16.8 bn
Anthropic net loss, Jan–Jun 2026 (incl. $72.4 bn non-cash charge) $80.7 bn
Anthropic valuation, Apr 2026 → Aug 2026 IPO target $350 bn → ~$965 bn

↑ Back to the article

Annexure E · Cracks in the AI trade, October 2026

Item Figure
Alphabet net income, Jul–Sep 2026 $26.2 bn (−25%)
Alphabet loss on equity investments, Jul–Sep 2026 $8.7 bn
Alphabet capex, Jul–Sep 2026 $51.4 bn (+114%)
Alphabet free cash flow, Jul–Sep 2026 −$9.1 bn
Alphabet share price, 29 Oct 2026 session −12.6%
Nasdaq 100 futures, morning of 30 Oct 2026 ~−2.5%
Nasdaq Composite, October 2026 −5%+
Nasdaq Composite, 16 Oct 2026 −2.1%
CoreWeave, 16 Oct 2026 −12%

↑ Back to the article

Annexure F · Markets India used to be compared with

Market Concentration 2026 fall Period
South Korea Samsung + SK Hynix ~40% of index −27% 19 Jun – 13 Jul 2026
Indonesia Heavyweight banks & conglomerates −34.7% Jan – Jun 2026
Brazil Large oil & mining weight −13.9% One session, 26 Oct 2026
South Africa One tech holding group ~25% — —

↑ Back to the article

Annexure G · Concentration side by side

Measure S&P 500 Nifty 50 Nifty 500
Top 10 stocks’ share of index ~39% ~52% ~29%
Largest block by index weight Tech
~38%
Financials
~36%
Financials
~31%
Largest block by profit AI-linked
~38%
Financials
~37%
Financials
~38%
Latest annual / quarterly profit growth 29%
(Jul–Sep 26)
9.1%
(FY26)
15.4%
(FY26)

↑ Back to the article

Annexure H · India and the US: the economy behind the index

Item India US
Population ~146 crore ~35 crore
Private consumption, % of GDP 61.5% (FY26) ~68%
Income per person (approx.) ~$2,900 ~$90,000
Median age (approx.) ~29 years ~39 years

↑ Back to the article

Annexure I · Nifty 500: where profit comes from

Item Figure
Profit share: banking, financial services & insurance (FY26) 37%
Profit share: oil & gas (FY26) 13%
Profit share: automobiles (FY26) 10%
Profit share: metals (FY26) 8%
Profit share: technology (FY26) 8%
Profit share: all other sectors (FY26) 24%
Nifty 500 total profit, FY20 ₹3.97 lakh crore
Nifty 500 total profit, FY26 ₹18.1 lakh crore
Nifty 500 profit growth, FY26 15.4%
Nifty 50 profit growth, FY26 9.1%
Nifty 50 share of Nifty 500 profit, FY18 → FY26 87% → 51%
Mid + small-cap share of FY26 profit growth ~2/3
Listed (Nifty 500) profit as % of GDP, FY26 5.2%
Nifty earnings growth forecast, FY27 15–16%

↑ Back to the article

Get Started

Grow your wealth with a

personalised financial plan

Get Started