Aakrit Vaish, founder of conversational AI company Haptik, was reported to have closed the debut fund of his venture fund, Activate, at $105 million (approximately JPY 16.3 billion as of September 2026 at $1≈JPY155; reports also cited roughly INR 1,000 crore, or 10 billion rupees) on September 16, 2026. The co-founder is Pratyush Choudhury, formerly of Together Fund. This is a fund formation with a clear picture of who is deploying how much capital and at what stage of AI startup, and the move by a founder turned investor to bet on the earliest stage of AI gives Japanese companies looking to partner with Indian AI firms a clue for reading the flow of capital and talent.
Activate's $105 million is made up of an $85 million earliest-stage fund called "Inception" and a $20 million growth allocation. The fund completed its first close at $75 million in December 2025, reaching full close within a year of launch. Investments target pre-seed, idea-stage AI-native companies spanning AI applications, foundation models, and physical infrastructure. The fund expects to back 25 to 30 companies, with per-deal check sizes ranging from $500,000 to $3 million. The base figure is $105 million, roughly JPY 16.3 billion in yen terms ($1≈JPY155), and reports also cited a rupee figure of roughly INR 1,000 crore (10 billion rupees, with 1 crore equal to 10 million).
Vaish is an operator who served as CEO of conversational AI company Haptik. Having built a business and taken it through to an exit, the founder is now betting on the earliest-stage AI companies as a capital provider. Activate has already invested in seven stealth-stage companies in consumer AI, AI-enabled services, and frontier technology, and through its growth allocation has backed Sarvam AI, ElevenLabs, and Wispr Flow. It is a structure in which an investor with a builder's perspective gets involved before a company has even fully taken shape.
Check sizes of $500,000 to $3 million per deal are designed to capture companies before they take shape, rather than making large investments at Series A and beyond. By spreading capital across 25 to 30 companies, the fund absorbs the high variance of earliest-stage bets through numbers. Its LPs include Vinod Khosla, General Catalyst, and Vijay Shekhar Sharma, along with more than 50 founders and AI researchers, general partners from a dozen or more global VC firms, and more than 50 family offices and corporations. A defining feature is the depth of the builder network sitting behind the capital.
Activate is backed by partnerships with NVIDIA, OpenAI, Anthropic, AWS, Google Cloud, and Microsoft Azure, as well as a GenAI community of more than 15,000 engineers. Combined with an LP base centered on founders and researchers, this is a structure that supports early-stage companies not just with capital but with a network of people and technology. Whether this earliest-stage strategy pushes companies successfully into their next stage will be measured by the fund's track record going forward.
When Japanese companies engage with AI in India, Activate's model raises the question of at what stage, and through whose network, to enter. The earliest-stage AI companies are selected and nurtured not just by capital, but by the builder network around them. Japanese VCs target India's universities before companies are even formed and Meta's investment in CRED, aimed at securing the founder as well show that early-stage relationships are determined less by the amount of capital and more by who you partner with. Before negotiating investment amounts, Japanese companies need to think first about securing a point of connection into these founder networks.
As capital flows to the earliest-stage AI companies, deal flow also spreads to AI outsourcing and SaaS businesses around them. AI outsourcing venture from a former Infosys chief and Indian SaaS company Rocketlane's Series C As these cases show, fund activity acts as a catalyst that broadens the base of companies that could become outsourcing and partnership targets for Japanese firms.
Vaish has brought the builder's perspective he developed at Haptik into investing, setting up a structure to bet $105 million on 25 to 30 of the earliest-stage AI companies. If Japanese companies want to partner with AI in India, the priority, ahead of negotiating the terms of a large investment, is to build a single point of connection into funds and founder communities like this one, positioning themselves to meet companies early, before they have even fully taken shape. A good starting point is to look at the investment areas of earliest-stage funds like Activate and identify a single technology theme that overlaps with one's own business.
RECENT
2026.09.24Bridgestone India’s 30th-anniversary film builds on the founder’s motto2026.09.24Flipkart’s move: putting the two meanings of "BBD" on Kolkata buses2026.09.24Veteran fertilizer maker KICL enters beverages with coconut water, starting in three South Indian states2026.09.24Pinit delivers saris in under 45 minutes, setting up a mirror and lighting at the customer’s home to choose fromCONTACT
Considering entering the Indian market?
Building on the developments on the ground covered in this article, we will suggest an approach that fits your products.
Book a free consultation →SOJAPAN
We support Japanese companies entering India, from market research through local partner development, test sales, and import.