From hype to hard returns: The startup reset for 2026

As India’s startup ecosystem looks ahead to 2026, the mood is noticeably more grounded. The era of growth-at-all-costs is fading, replaced by tougher questions around execution, AI readiness, and sustainable profitability. Scale alone is no longer enough; what matters now is how efficiently, intelligently, and responsibly startups grow.

These shifts took center stage at a Snowflake panel discussion titled ‘Future-ready Startups: Data, Product Innovation & Capital Strategy for 2026’. Industry leaders including Dushyant Bhatt, Chief Technology & Product Officer, The Hosteller; Anirudh Damani, Managing Partner, Artha Venture Fund; Milind Borgikar, CTO, Ayekart Fintech; and Pravin Fernandes, Head of Commercial Business (Digital Natives & Mid-Market), Snowflake; came together to unpack what the next phase truly demands.

The discussion, moderated by Shivani Muthanna, Senior Director, Content Partnerships, YourStory, delved into varied aspects, from agentic AI-ready infrastructure and proprietary data strategies to regulatory-grade automation and capital efficiency. The takeaway was clear: in 2026, execution will outweigh experimentation, and outcomes will matter more than ambition.

The shifts defining 2026

Dushyant Bhatt, CTO of the Hosteller, opened the discussion by outlining his plans for 2026. He emphasized the importance of streamlining scattered data from multiple Point of Sale (POS) sources across the business’s 80+ properties and leveraging AI for precise decision-making. This would ensure clean customer data and maintain full compliance with government regulations to drive smooth operations. 

Anirudh Damani, Managing Partner at Artha Venture Fund, shared a candid outlook for 2026, calling AI inevitable across businesses. However, he cautioned against chasing large language models. He argued that India continues to lag in foundational AI resources, and instead advocated for applied AI that drives real efficiency in B2B/B2C models. Reflecting this approach, nearly 25% of Artha Venture’s portfolio is now focused on applied AI.

Damani also emphasized the urgency of profitability from day one, pointing to the shutdown of over 54,000 startups in India. “India is a very, very expensive capital market. We don't have unlimited capital,” he said, adding that even traditional asset classes are giving outsized returns. “Capital is fungible and liquid; it flows to where returns are highest.” For founders, the implication is stark: if a startup’s Return on Equity (ROE) fails to perform an investor’s my Return on Investment (ROI, it risks eroding shareholder value. “That’s why you have to show profits,” he said.

Milind Borgikar, co-founding CTO at Ayekart Fintech, highlighted the company’s operational scale, having moved goods worth Rs 8,500 crore in 3.5 years. In December 2025, Ayekart launched Antriva, a merchant-agnostic platform designed to embed finance for into supply chains, enabling timely supplier payments and smoother buyer deliveries. As an NBFC-licensed entity, Borgikar noted that Ayekart will continue to prioritize regulatory compliance in 2026. 

Pravin Fernandes, Head of Commercial Business (Digital Natives & Mid-Market) at Snowflake, detailed the company’s internal shift toward agentic AI on its data platform. Under CEO Sridhar Ramaswamy, Snowflake retired more than 2,000+ Tableau licenses, adopted Snowflake Intelligence UI to deliver real-time customer insight. What once took four hours of preparation can now be accessed in seconds via a mobile app, significantly boosting sales productivity, reducing hiring needs, and accelerating AI adoption for customers.

Data strategies for 2026

As competition intensifies, robust data strategies are becoming non-negotiable for startups. In 2026, data will be the backbone of AI-led efficiency, anomaly detection, and scalable operations. Panelists shared how startups should rethink data management to stay future-ready.  

Bhatt spoke about enforcing a "zero sheets" policy across the company’s 80 properties and 1,000+ staff, eliminating Excel and Google Sheets entirely. Instead, all operational data, from guest check-ins and service requests to meter readings and laundry tracking, is captured within proprietary in-house systems.

This approach has helped the Hosteller generate rich datasets that power low-cost AI tools, automate workflows, and detect anomalies such as excess electricity consumption during low occupancy, triggering automated shutdowns. AI is also used to analyze more than 10 lakh guest reviews, categorize multilingual feedback, run sentiment analysis, identifying property-specific issues and enabling targeted interventions to reduce operational leakages.

Damani emphasized that data depth is critical for credible AI adoption, especially in B2B startups. He shared an instance where a startup was rejected after conducting only 12 transactions while heavily relying on ChatGPT. Limited datasets, he warned, lead to ineffective training and hallucinations - unacceptable in customer-facing tools that demand near-zero error rates. True applied or agentic AI must be built on large, proprietary, domain-specific datasets. Damani urged founders to invest early in data scientists and teams with proven experience in large-scale product builds, cautioning that hype-driven shortcuts rarely sustain.

For Borgikar, the focus is on targeted automation at critical data touchpoints. Ayekart digitized Goods Received Notes for clients such as Zepto and QSR chains like Subway, replacing paper-based verification with automated emails and validation links, eliminating fraud over the last six months. Additional safeguards such as watermarks, OTP verification, and DigiLocker integration have strengthened bill discounting workflows, improving trust and efficiency across supply chains.

Fernandes stated that the data conversation has evolved rapidly. While startups once focused on breaking data silos, the critical question for 2026 is whether their architecture is agentic AI-ready. He outlined three imperatives: unified agentic AI infrastructure, strong governance with auditable AI-agent logs and PII masking, and vectorized architectures capable of handling both structured and unstructured data. Together, these form the foundation for scalable, future-proof AI operations.

AI: Making workflows work

Across the discussion, speakers advocated a practical and cautious approach to AI adoption, focused on improving decision-making while navigating regulatory and operational constraints. 

Borgikar spoke about balancing AI-led automation with regulatory transparency. As a regulated entity, Ayekart prioritizes explainable decisions over opaque machine-learning models. The company begins by clearly defining business decisions and then introduces automation through “human-in-the-loop” systems, asking "why" automation is needed before focusing on "how".

To maintain clarity, AI-generated outputs are labelled with UX badges, while unstructured LLM responses are reformatted into structured JSON to ensure consistency. Ayekart also applies layered intelligence selectively, using AI primarily for classification and context-building rather than end-to-end automation. “Because we deal with regulators, we don’t have much leverage in adding more AI-powered automation,” Borgikar noted. “But AI has been immensely helpful in improving decision-making and productivity.”

From an investor’s perspective, Damani highlighted how even VC firms struggle with fragmented data despite investing in AI. Many still rely on Excel sheets, with startup information scattered across inconsistent financial models, decks, compliance records, and naming conventions, which makes quick, holistic assessment difficult.

To address this, Artha built a custom Asana AI engine to enrich and standardize data across 140 portfolio companies, integrating information from financials, Ministry of Corporate Affairs records, Goods and Services Tax (GST) filings, and founder backgrounds. With tighter deal timelines and increased competition, Damani said AI-driven insights are increasingly replacing traditional “people-to-people” trust. Referencing Karza Technologies (now Perfios), which scaled its valuation from Rs 8 crore to Rs 600 crore by verifying data accuracy in under an hour, he shared plans to hire a CTO and build deeper in-house AI capabilities next year.

The panelists agreed that AI delivers real productivity gains only when tied directly to revenue or decision efficiency. Fernandes warned startups against chasing AI hype without clear ROI. Boards often push teams to deploy AI for tasks like call summaries or email parsing, but these rarely translate into cost reduction or revenue growth—especially in India, where labour remains inexpensive. CFOs will need clearer incentives to fund AI adoption, he said, adding that true AI-led productivity comes from targeted, revenue-linked use cases, not generic demos.