How Upsparks Capital carved out a niche in India’s pre-seed investments

Bengaluru-based proprietary venture capital firm Upsparks Capital has quietly emerged as a significant force in India’s early-stage startup ecosystem. Since 2020, the proprietary fund has deployed Rs 57 crore across 65 investments, maintaining a sharp filter with an acceptance rate below 3% of evaluated startups, said a top executive in an interaction with YourStory.

Backed entirely by the personal capital of its founding partners Shivam Prasad, Mohamad Faraz, and Vinay Jain, Upsparks currently manages assets worth over Rs 198 crore. According to co-founder Faraz, the firm has achieved a multiple on invested capital (MOIC) of 3.5x to date, a performance that places it in the top tier of early-stage domestic funds.

“Early-stage investing today is both sharper and more nuanced than it was five years ago,” Faraz said. “Founder quality has always been central, but what's changed is that great founders now understand fundraising, storytelling, and metrics far earlier in the journey.”

The firm's portfolio includes breakout companies such as marketing tech firm Pixis, Flam, a mixed reality platform, AI startup Smallest.ai, and space-tech venture GalaxEye. Of its investments, 27 portfolio companies have secured follow-on funding totalling Rs 4,500 crore, with Upsparks often writing the first institutional cheque in 80% of its deals. To date, the firm has seen three exits: logistics startup Zypp, sustainable packaging venture Bambrew, and a partial exit in Pixis. Four portfolio companies are classified as “soonicorns,” or startups nearing a $1 billion valuation.

Founded in 2020, the founders launched the firm after exiting their first venture, Solutions Infini. The startup was founded in 2009 and operated in the communications technology space. It later merged with a European firm and was renamed as Kaleyra. It is now a NYSE-listed company.

Upsparks typically invests between Rs 1 crore and Rs 2 crore at the pre-seed and seed stages. The team reviews more than 150 pitch decks each month and moves quickly, often making investment decisions within 10 days of the first meeting. “We’re drawn to the zeal to build—a founder who wakes up every day obsessed with solving the problem,” Faraz said. “We look for clarity of thought on fundamentals, not just buzzwords.”

Riding the AI wave, selectively

Upsparks has witnessed a surge in AI-related startup activity since the launch of ChatGPT in late 2022. While the firm acknowledges the explosion of ideas, many of which are superficial, it remains focused on backing companies solving real operational bottlenecks. “We’ve seen a surge in ChatGPT wrappers with little defensibility,” Faraz observed. “At the same time, we’re also seeing some of the sharpest technical founders we’ve ever met, people building genuinely novel infrastructure, agents, and vertical applications with deep insight.”

The fund's thesis in AI emphasises  defensibility through domain depth, proprietary data, and workflow integration rather than through the model itself. Industry-specific AI tools with operational leverage in sectors such as banking, manufacturing, and compliance-heavy financial services remain a key focus. “The defensibility rarely lies in the model; it lies in distribution, data loops, and domain depth,” Faraz explained.

India, according to Faraz, has a competitive edge at the application layer of AI rather than foundational model development. “The race to build foundational models is capital-intensive, talent-concentrated, and infra-heavy,” he noted. “India’s edge lies in building AI-native applications that solve uniquely Indian problems in healthcare, vernacular commerce, financial inclusion, or agri-tech.”

Infrastructure and execution: Where the real bets lie

While Upsparks is bullish on vertical AI applications, it is equally interested in what it describes as agentic infrastructure and orchestration layers. These are tools that help companies deploy, fine-tune, and monitor large language models (LLMs) at scale. Faraz sees these investments as “multiplier layers” that enable faster and more scalable AI adoption across industries.

The firm applies different frameworks for infrastructure versus application-layer startups. Infrastructure is judged on its ability to catalyse  ecosystems, while applications are evaluated based on their ability to build user habits, achieve workflow lock-in, and harness proprietary data. “We’re excited by both,” Faraz said. “But infrastructure that unlocks velocity in AI adoption is something we’re actively looking to back.”

Upsparks’ investment in Smallest.ai exemplifies its approach. The startup stood out for its speed, clarity, and deep understanding of user pain points. “Early wins depend on tempo and founder decisiveness in execution,” Faraz said. Across high-performing portfolio companies, the firm sees recurring traits: clarity of thought, bias for action, and strong early believer networks.

As a proprietary capital fund, Upsparks operates free from the institutional constraints and LP expectations that often shape traditional VC behaviour. “Without LP pressure, we can move faster, be more founder-first, and back non-consensus bets early,” Faraz said.

The absence of external investors allows the firm to take a thematic yet flexible approach to portfolio construction. While Upsparks maintains areas of focus such as AI infrastructure, fintech rails, and global software, it also leaves room for conviction-led outliers. The firm typically targets 7 to 10 core investments per year and plans to add 25 new startups over the next 24 to 36 months.

Some of its most valuable lessons, Faraz noted, have come from bets that defied traditional theses. “Product obsession and user love can beat TAM assumptions. We’ve passed on some companies where, in hindsight, the founder’s clarity and hunger outweighed early traction,” he admitted.

Looking ahead

Faraz says the fund is constantly scanning for big, non-obvious opportunities and cites one in particular that remains on his wishlist. “We’d love to back someone building a Palantir for India. An intelligence layer for India’s public systems, enterprises, and infrastructure-heavy sectors,” he said. Such a system would enable mission-critical decisions by processing fragmented data across government and industry.

As hype cycles from Web3 to quick commerce to AI continue to reshape the investment landscape, Faraz remains grounded in core questions: Is this a real shift in user behaviour? Is there a non-obvious wedge into the market? Would this still be a good business if the hype disappeared?

“In a world of flashy rounds and big tweets,” Faraz concludes, “Our mission is to back founders at Day 0, with capital, belief, and real operational support, and help them navigate the early stages of company building. We support startups not just with funding, but with GTM guidance, hiring support, downstream investor access, and access to curated tools and benefits. We’re building an ecosystem where founders feel backed, not just financed.”