$270 Billion in One Week. Here's What a Decade of Pledges Actually Delivered.

Over the past decade, India has attracted some of the biggest investment announcements on the planet. From Foxconn's semiconductor megaplant to Saudi Arabia's $100 billion commitment to Tesla's repeated teasing of a factory — the pledges have been spectacular. The delivery has been... mixed.

We tracked 14 of the largest investment pledges made to India since 2014. Here is the scorecard:

The pattern is unmistakable. And it splits cleanly into two categories.

Four names stand out on the delivered side: Apple, Samsung, Amazon, and Micron. They share one thing — they had a direct commercial reason to be in India. Not a diplomatic photo-op. Not an MoU at a summit. A business case.

Apple is the clearest example. India's share of global iPhone production went from under 1% in 2017 to roughly 20% by 2025, with $22 billion worth of iPhones assembled in FY25 alone — a 60% year-over-year jump. Two new factories opened in April 2025 (Foxconn Bengaluru and Tata Hosur). The target of 25-32% of global production by FY27 looks achievable. Why? Apple needed to de-risk from China. India was the answer. The incentive was existential, not ceremonial.

Samsung doubled its Noida factory to 120 million units per year by 2020, then expanded into laptop manufacturing in 2025. Amazon has deployed roughly $40 billion across AWS data centers, fulfillment infrastructure, and acquisitions. Micron's $2.75 billion Gujarat semiconductor facility began commercial production in February 2026 — on schedule.

And then there's the other side of the scorecard.

Foxconn-Vedanta's $19.5 billion semiconductor plant is the poster child. Announced with fanfare in September 2022, it unravelled within months. Foxconn pulled out in July 2023, citing a project that "was not moving fast enough." Their technology partner, STMicroelectronics, refused to take a stake. The Indian government raised questions about inflated cost estimates. Vedanta took over but couldn't find a replacement tech partner. As of 2026: zero groundbreaking, zero production, zero jobs. Foxconn later partnered with HCL for a $433 million chip packaging plant — that's 2.2% of the original promise.

Tesla spent nine years teasing India. Elon Musk first mentioned entry in 2016. In 2023, India rolled out the red carpet with import tax cuts tied to a $497 million investment commitment. Musk cancelled his April 2024 visit at the last minute. India's Heavy Industries Minister stated publicly that Tesla "does not want to produce EVs in India." In July 2025, Tesla finally opened a 4,000 sq ft showroom in Mumbai selling imported cars. Nine years of promises. One showroom.

Saudi Arabia's Crown Prince MBS pledged $100 billion during his February 2019 India visit. Actual Saudi investments in India total approximately $10 billion, primarily through Aramco and PIF minority stakes in Jio ($1.5B) and Reliance Retail ($1.3B). The massive Aramco-IOC refinery in Maharashtra has been discussed since 2018 without breaking ground. A "High-Level Task Force on Investment" was only constituted in 2024 — five years after the pledge.

No institution better illustrates the pledge-to-delivery gap than Vibrant Gujarat — the biennial investment summit that has been running since 2003. The headline numbers have been staggering:

But Gujarat Assembly data tells a different story. Across nine editions (2003-2019), the state signed MoUs worth Rs 9,45,158 crore. Actual investment that materialized: Rs 21,557 crore. That's a conversion rate of 2.28%.

Economist Hemant Kumar Shah pointed out that India's total industrial investment during 2007-2010 was approximately Rs 24 trillion — yet the Vibrant Gujarat 2009 summit alone claimed Rs 24.3 trillion in MoUs. The pledges routinely exceed the state's own GDP. The 2009 power sector saw 31 MoUs worth Rs 2.1 lakh crore for 58,000 jobs. Zero materialized.

The Gujarat government dropped 25,975 of its own MoUs, citing "dubious and deteriorating financial condition" of the signing companies.

The pledge-to-delivery gap isn't uniquely Indian. But the data reveals a clear hierarchy of reliability:

Globally, academic research by Abigail Hornstein (China Economic Review, 2017) found that firms fulfil an average of 59% of their investment pledges within two years — but this applies to genuine corporate commitments with contracted amounts, not summit theatre.

The Peterson Institute for International Economics found that Trump's claimed "$17 trillion" in investment pledges globally are largely fiction. Japan's own trade negotiator admitted just 1-2% would be actual equity investment — the rest is loans and loan guarantees. Saudi Arabia's 2017 first-term pledge of "$350 billion in deals" produced roughly $5 billion in actual investment by 2024 — a 1.4% hit rate.

Against this backdrop, the India AI Impact Summit 2026 arrived with the biggest numbers in Indian summit history. Here's what was announced:

The summit also produced the New Delhi Declaration on AI Impact, endorsed by 88 countries including the US, UK, and China. It lays out seven guiding principles — dubbed the "Seven Chakras" — from democratizing AI access to responsible development. It is entirely voluntary and non-binding.

Here's the nuance the headline numbers miss. Not all $270 billion is summit theatre. Some of it is backed by genuine commercial demand.

The global AI infrastructure buildout is real. OpenAI's $500B Stargate project has Tata as a construction partner. NVIDIA's GPU supply to India is contracted and specific — 20,736 Blackwell Ultra GPUs at Yotta's Greater Noida campus, operational by August 2026. Blackstone's $1.2 billion in Neysa is equity financing with a named company and a defined GPU count (20,000+). These are not MoU-grade announcements.

Google's $15 billion includes a new subsea cable system — physical infrastructure that must be built to serve India's growing cloud demand. Microsoft's $3 billion initial tranche is "on track to spend by end of calendar year 2026," with a new Hyderabad data center region announced for mid-2026.

The question isn't whether any of the $270 billion will materialize. It's how much of the headline number is genuine corporate investment vs. aspirational domestic conglomerate projections. If history is any guide, the tech company pledges (Google, Microsoft, NVIDIA) have a 60-80% chance of materializing. The $210 billion from Reliance and Adani? That's a different category entirely.

Buried inside India's own data is proof that rewarding the product, not taxing the process, converts pledges into factories.

The Production Linked Incentive (PLI) scheme, launched in 2020, does something radical by Indian policy standards: it pays companies based on what they actually produce, not what they promise to invest. The results?

The star performer: mobile phones. Exports grew 8x in four years — from Rs 228.70 billion (FY21) to Rs 2 trillion (FY25). Smartphones became India's single largest export category. Apple vendors alone exported $50 billion worth of iPhones through December 2025. Pharma flipped from net importer to net exporter of bulk drugs. These aren't MoU numbers. These are customs receipts.

The lesson is simple: when India rewards output instead of taxing input, money follows. PLI doesn't care about your MoU. It cares about your invoice.

The data points to a clear set of policy shifts — backed by what worked for India (PLI) and what worked globally (Vietnam, Singapore, Ireland).

India's 18% GST on software services taxes the act of building. R&D activity itself is a taxable event. Meanwhile, India spends just 0.64% of GDP on R&D compared to Israel's 6.35%, South Korea's 4.96%, and China's 2.6%. The angel tax (abolished only in 2024 after a decade of damage) literally taxed the process of raising startup funding. Zero-rating R&D and product development activities — not just exports — would directly lower the cost of creating products in India.

India's $283 billion IT sector — largely a services empire rather than a product builder, as explored in our India AI consumer analysis — is over 70% services — building other countries' products, not India's own. Indian IT firms spend 1-2% of revenue on R&D versus 10-15% by global tech leaders. There is no PLI equivalent for software products. A "Software Product PLI" rewarding product revenue (not service billings) would redirect a fraction of India's massive tech workforce toward building products, not just manning call centres. If even 5% of services revenue shifted to product R&D, that's $14 billion annually in product development.

Singapore's Economic Development Board isn't a marketing agency — it's a dealmaker with hard KPIs: grounded investment, operating expenditure, quality jobs. India's state summits produce 2-15% conversion rates because nobody is accountable for converting the MoU into a factory. States should shift from biennial publicity events to permanent facilitation institutions that track and publish conversion rates publicly. Maharashtra's selective approach (signing only assured deals) already achieves 55-75% conversion — proving the model works when the incentive changes.

Vietnam's FDI success isn't about cheaper labour. It's about ready-built factories with digital infrastructure in industrial zones, fast approvals, and tax incentives tied to actual production. Samsung now manufactures half its smartphones in Vietnam. Apple invested $15.84 billion since 2019, with 35 vendor partners. India's recent SEZ reforms — reducing land requirements from 50 to 10 hectares for semiconductors — are steps in the right direction. But Vietnam didn't wait for companies to build their own factories. It built the factories first.

Here's the alarm bell nobody is talking about. India's net FDI plunged 96% in FY2024-25 to just $353 million. Gross FDI rose to $81 billion, but repatriation by foreign firms hit $51.5 billion — the highest in a decade. Companies are taking money out of India faster than new money comes in. Ireland's lesson: investors don't need the lowest tax rate — they need predictability. Ending retrospective taxation, simplifying dispute resolution, and committing to multi-decade policy stability would matter more than any summit pledge.