AI Money Flows, But Not Via India

The global race in AI is no longer about who has the best tech but where the money is flowing. And right now, the US is attracting a huge share of global AI investment, with companies spending heavily on building new models, data centres, and AI-led products. 

A report published by Stanford University mentions that $285.9 Bn was accounted for in 2025 for private AI investments in the US, which is more than 23X the $12.4 Bn invested in China. Even with growing concerns about an ‘AI bubble’, investors are putting more money into the space.

India, however, is not part of this bull run. Unfortunately, foreign investors have been pulling out funds and moving towards markets and companies offering stronger exposure to AI-driven growth. 

This is not because Indian companies are not using AI — they are, but mostly to improve efficiency, cut costs, and protect margins. What’s missing is the evidence of AI being used to build new products, platforms, or revenue streams at scale. 

At a time when the next wave of tech growth is taking shape, does India risk being left behind as the world progresses with fresh investments? Let’s talk all about it in this edition of The AI Shift.

Where’s The Capital Flowing

Foreign portfolio investors (FPIs) have been pulling money out of India at a record pace. In March 2026 alone, FPIs sold $12.58 Bn worth of Indian equities, the highest monthly outflow on record, with net selling across 21 of 23 sectors, according to the National Securities Depository Limited (NSDL).

This is not a one-off event. FPI equity assets under custody (AUC) dropped from nearly $930 Bn in September 2024 to around $660 Bn by March 2026, a near 29% decline. Part of this is explained by macro triggers such as geopolitical tensions, inflation concerns, and rupee weakness.

On the opposite spectrum and amid ongoing global volatilities, the US ETF market saw $112.9 Bn in net inflows in March 2026, with equity ETFs alone attracting $56.7 Bn.

At the same time, AI continues to absorb a disproportionate share of global capital. US-based AI companies captured nearly 79% of global AI funding in 2025, as per some estimates.

Rishabh Khandelwal, the founder and CEO of investment advisory firm Sabiduria Capital, said that this is not just about AI exposure.

“The AI trade is a primary driver, but valuations also matter. Indian companies are not cheap compared to other emerging markets. Taiwan, South Korea, and China offer both lower valuations and stronger AI-linked opportunities, making them more attractive within the same emerging market allocation bucket,” he added.

Slow And Steady Won’t Win The AI Race

Indian IT companies are not yet signalling participation in the AI-led value cycle in ways that global investors recognise. The underlying issue is also how AI is being adopted. Across Indian IT firms, AI is largely being deployed for productivity improvements, cost optimisation, and margin protection.

Productivity gains and automation are increasingly being passed on to clients, creating what analysts describe as AI-led deflationary pressure on revenues.

In contrast, US tech companies are using AI to develop new product categories, platform businesses, AI-linked revenue streams and visible capex commitments.

This difference matters because investors are not tracking whether companies are using AI, but whether they are using it to create new value pools.

Without clear signals of AI-led revenue growth, proprietary products or IP, or large-scale AI-native offerings, Indian IT companies risk being considered efficiency stories, not growth stories. That distinction directly impacts how capital is allocated.

Uddeshya Goel, investment advisor at Empirica, said that the gap is also structural. “Indian IT firms are still largely service-led and late to building meaningful AI IPs.”

He pointed out that capital is increasingly flowing toward AI infrastructure and hardware ecosystems, including power systems and data centre supply chains, rather than traditional services. At the same time, large IT firms have prioritised buybacks and incremental capability acquisitions over aggressive AI investments, signalling a wait-and-watch approach.

“The risk is that by the time they pivot meaningfully, the value pools may already be captured,” Goel added.

This is also translating into a more cautious positioning by global investors. Khandelwal notes that foreign investors are actively reassessing their exposure to Indian IT companies, comparing it with opportunities in developed markets such as the US, as well as other Asian markets.

“They are waiting for more clarity before increasing or rebalancing positions, indicating a pause rather than a complete exit,” the investment advisor said.

India’s AI Play Needs Scale

Domestic institutional investors continue to support Indian tech stocks to some extent, but foreign capital, which typically drives higher investment value cycles, is increasingly flowing toward markets where AI-led value creation is visible and scalable. However, the absence from the current AI capital cycle does not necessarily mean India has missed the opportunity entirely.

Experts are of the view that emerging AI-adjacent plays, including data centres, semiconductor-linked manufacturing, and precision engineering, are early proxies to the AI cycle within India.

“It’s not that India has missed the AI trade. There are proxy plays already emerging, and over time, India will catch up,” Khandelwal said.

In terms of building opportunities, Abhishek Srivastava, managing partner at Kae Capital, said that India’s advantage is not in building frontier models but in ‘domain-specific, context-rich applications built for real-world complexity’.

He highlighted areas such as healthcare diagnostics and multilingual AI, where India’s data diversity across languages and use cases becomes a structural advantage. “Playing to your strengths is always a great strategy,” he added.

But some believe that the centre of gravity in AI is shifting from where models are built to where they are applied. Early signs of this are emerging in the application layer, particularly in areas such as consumer AI platforms, AI-led cost disruption to unlock new markets, vernacular AI targeting non-English users, and domain-specific data infrastructure.

However, for India to be seen as part of the global AI investment cycle, the shift requires visible movement from the largest players in the ecosystem. This means transitioning from AI-enabled services to AI-led businesses, from delivery models to product and IP ownership and from efficiency gains to revenue creation.

Without these signals, global investors have little reason to reposition India as an AI destination. Therefore, the immediate challenge is not competing with the US but a meaningful shift towards AI-led growth to attract investor capital.

Top Stories From India & Around The World

• OpenAI CTO Srinivas Narayanan Quits: Srinivas Narayanan has stepped down from his role at OpenAI after a three-year stint, citing personal reasons. His departure comes amid a series of leadership changes as the company expands aggressively in India and the Asia-Pacific region.
• Centre Sets Up AI Governance Body: The Indian government has formed the AI Governance and Economic Group (AIGEG), chaired by Ashwini Vaishnaw, to coordinate AI policymaking across ministries. The body will oversee regulation, assess risks, and guide long-term AI deployment strategies across sectors.
• Nvidia Launches Ising Models: The fabless chipmaker has unveiled the world’s first open AI model family for quantum computing, Ising, designed to improve processor calibration and error correction. The models deliver up to 2.5X faster performance and 3X higher accuracy than existing methods.
• Claude Opus 4.7 Goes Live: Anthropic has launched Claude Opus 4.7, its latest model with improved coding, multimodal capabilities, and stronger instruction-following. The model also introduces higher-resolution image processing, better long-context reasoning, and new safeguards for cybersecurity misuse, while maintaining pricing parity with its predecessor.
• India’s Patent Gap: Filings jumped 30% YoY to 1.43 Lakh in FY26, but grants fell 36%, exposing a widening gap. Legal hurdles, low-quality filings, and administrative delays continue to slow approvals, with startups lagging behind academia in patent contributions. 

The Weekly Buzz: Anthropic Pushes AI Deeper Into Creative Workflows

Anthropic has rolled out Claude Design, a new capability under Claude that lets users generate prototypes, slides, and one-pagers simply by describing what they want.

The launch is powered by Claude Opus 4.7, the company’s most advanced vision model yet, positioning Claude as more than a chatbot. It now acts as a design collaborator that can iterate through conversation, accept inline edits, and adapt via custom sliders. Users can also export outputs to Canva, download as PDF or PPTX, or hand off workflows to Claude Code.

The bigger play is system-level. Claude can reportedly read a team’s codebase and design files to auto-generate and enforce a consistent design system, applying brand rules across projects without manual intervention.

Early reactions reflect a familiar split.

On one side, builders and solo developers are calling it a ‘game-changer’ for rapid prototyping, compressing what used to take days or weeks into a few prompts and iterations. On the other hand, professional designers remain unconvinced. Many argue that outputs still feel template-driven and ‘cookie-cutter’, far from replacing tools like Figma or the depth of human-led design thinking. 

Design incumbents saw pressure, with Figma reportedly sliding around 7% intraday, while Adobe also saw a dip, reflecting investor anxiety over AI-native design tools.

Startup In The Spotlight: Tattvam AI

As AI model innovation accelerates, semiconductor design workflows are emerging as a key bottleneck. Converting chip architectures into manufacturable layouts remains highly manual, time-intensive, and reliant on large engineering teams, slowing down production cycles despite rising global demand for compute.

Founded in 2025 by Bragadeesh Suresh Babu and Lannan J, Tattvam AI is building autonomous AI systems focused on the physical design stage of chip development. The startup operates between chip design firms and fabrication units, aiming to enhance existing electronic design automation toolchains rather than replace them.

At its core, Tattvam uses AI-led decision systems to automate complex parts of physical design, reducing execution timelines from nearly a year to a few weeks. Its solution targets global semiconductor companies, fabless startups, and design services firms looking to accelerate chip development cycles.

On the business side, the startup is exploring pricing models similar to traditional EDA tools, reflecting the capital-intensive and risk-sensitive nature of the industry. Enterprise adoption is expected to hinge on proof-of-performance deployments and early validation.

Tattvam is positioning itself to benefit from the growing demand for faster chip design, operating in an EDA market projected to reach $32.75 Bn by 2032.

Prompt Of The Week

What prompts and hacks are CTOs, CEOs and cofounders using these days to streamline their work?

Here’s the prompt used by Angad Ahluwalia, COO, Arinox AI, to compress the entire business development workflow — from research to strategy to proposal creation — into a single flow. It helps him understand the client and how to win the deal.

“You are a strategic enterprise deal analyst and solution architect for Arinox AI.

Given a company, decision-maker, or opportunity (add input), break it down into:

1. Decision Maker Intelligence: Profile (role, KPIs, incentives), priorities, fears, and success metrics + Buying triggers vs. resistance points

2. Organisation Readiness for AI: Current systems, data maturity, and integration complexity + Where AI will realistically land (not just aspirationally)

3. High-Probability Entry Points: 2–3 use cases most likely to get approved  + Where ROI is visible within 30–60 days

4. Deal Strategy: Best positioning narrative (cost, efficiency, control, revenue) + What NOT to say + Deal first wedge (pilot, COE, or full platform)

5. Hidden Risks & Blind Spots: Political, operational, or technical blockers + Where deals typically stall in similar orgs

6. Action Plan: Next 3 moves to progress the deal

7. Generate Output

• Create a structured, client-ready Arinox AI proposal/dossier
• Include: industry context, problem framing, agentic use cases, solution, architecture [specify], expected outcomes, and phased rollout
• Keep tone: bold, outcome-driven, and executive-ready

Respond in sharp, decision-ready bullets and structured sections. No generic advice.”

Editor’s Note: Some prompts may need to be adjusted by users for best results or may not work as intended for certain users.

[Edited by: Shishir Parasher]
[Creatives by: Varshita Srivastava]

The post AI Money Flows, But Not Via India appeared first on Inc42 Media.