Welcome to The AI Shift by Inc42, our all-new newsletter that delves deep into the world of artificial intelligence, LLMs, big tech giants and the major trends sweeping the Indian startup and tech ecosystem. Here’s the fifth edition; do send us your feedback and suggestions so we can improve as we go along!
When the finance minister, Nirmala Sitharaman, presented the Union Budget on Sunday (February 1, 2026), there was a lot of emphasis on AI as an enabling layer embedded in governance systems, sectoral platforms, and public infrastructure.
AI is being seen as a force multiplier for better governance, positioning AI as a tool to improve outcomes for farmers, youth, women in science, technology, engineering, and mathematics (STEM) as well as everyday life.
The impact has been such that AI is being seen less as a venture theme and more as national infrastructure, with startup upside expected to emerge indirectly rather than through explicit incentives. The budget also carried subtle but significant signals for investors, with one of the strongest ones being the emphasis on data centres, compute capacity, and long-horizon AI infrastructure.
So, what exactly do these signals mean for AI stakeholders, especially investors? And what’s in store for tech startups?
India’s AI Push Unfazed By Resource Limits
India’s data centre demand could exceed 10 GW, requiring 45-50 Mn sq ft of real estate and 40-45 terawatt (TW) hours of incremental power.
According to Debasish Mishra, the chief growth officer at Deloitte South Asia, the demand could attract nearly $100 Bn of immediate investment, positioning India as a regional data centre and AI infrastructure hub. The expectation, however, is not near-term AI monetisation, but the creation of compute gravity and capital certainty over decades, and there is no dearth of funds.
The confidence is echoed by the government officials directly overseeing the rollout of AI in the country. Abhishek Singh, the CEO of the IndiaAI Mission, told Inc42 that he does not see any funding challenges arising from the budget provisions for the mission.
He added that the Union Budget’s data centre tax exemption until 2047 and the introduction of safe harbour provisions are expected to further catalyse investments across the data centre and AI ecosystem in India.
Mahesh Makhija, leader for technology consulting at EY India, is of the opinion that the government is laying the foundation for sustainable AI adoption by prioritising talent development, research, compute infrastructure, data readiness, and mission-mode innovation together.
A Step Forward, But Startups Feel Stuck
The AI push is most visible across government platforms. Bharat VISTAAR integrates AgriStack portals with ICAR data to offer multilingual, AI-driven advisory to farmers. AI-backed tools in logistics and customs are positioned as compliance and efficiency enablers, while assistive technologies under the Divyang Sahara Yojana frame AI as social infrastructure. Workforce disruption is acknowledged through a high-powered ‘education to employment and enterprise’ standing committee.
For India’s AI startups, however, it represents a mixed reality. While Ankush Sabharwal, the founder and CEO of CoRover.ai, believes that increased investment in AI infrastructure can create new opportunities for startups to collaborate with the government and deliver innovative solutions, driving adoption across industries, he also pointed out that the most immediate friction for AI startups remains unresolved.
“Affordable compute and data access continue to be the biggest bottlenecks, and addressing these could unlock significant innovation and scalability.” Sabharwal also wants the government to provide clarity on regulations and guidelines while offering targeted incentives for specific industries. In addition, industry stakeholders want incentives for government organisations that adopt AI solutions.
The view aligns with Adya.ai’s CEO and cofounder, Shayak Mazumder, who argued that while policymakers appear to have thought through AI across the stack, from applications to chips, the startup experience remains shaped by uncertainty around procurement, compute access, capital support, and model governance.
Meanwhile, Raj K Gopalakrishnan, the cofounder and CEO of KOGO AI, views this as the moment when India stops participating in the global AI hype cycle and starts architecting its own sovereign AI economic strategy.
He underlines that initiatives such as the AI Economic Council, Sovereign AI-OS, the data centre tax holiday until 2047, and the ISM 2.0 semiconductor push reflect a full-stack ambition rooted in capital efficiency and local problem-solving rather than centralised compute dominance.
Investors Show Cautious Optimism
From an investor perspective, Budget 2026 lowers AI adoption friction but does not feel to be strengthening startup moats.
Pranav Pai, managing partner at 3one4 Capital, told Inc42 that the platform-first approach signals long-term confidence in AI as national infrastructure, with tax holidays for data centres until 2047 and safe harbour provisions reflecting a multi-decade commitment rather than a short-term startup stimulus. He said this structurally improves the risk-reward equation by easing constraints around compute access and cloud costs that have historically affected scaling.
He added that the infra push will ease supply-side constraints, such as compute access and cloud costs, which have historically limited scaling. But startup success will still depend on founders building applications that leverage AI across governance, agriculture, education, accessibility, and enterprise workflows. Pai said that investors now prioritise monetisation, execution, and defensible application-layer value over technical novelty.
Others say India’s platform-first AI approach lowers barriers to early adoption in complex sectors such as agriculture, logistics, healthcare, and public services by giving startups faster access to users and distribution. Public platforms alone cannot create durable moats.
This is echoed by Brijesh Damodaran, managing partner of Auxano Capital, who has warned that subsidised compute and shared datasets can compress margins over time, pushing startups away from model-centric differentiation toward proprietary workflows and domain intelligence.
As the India AI Impact Summit 2026 approaches, investors and founders alike are waiting to see whether the platform-first vision outlined in Budget 2026 translates into clearer startup-operable policy. For now, the sentiment of founders and investors is cautious but constructive.
Top Stories From India & Around The World
- IndiaAI Mission Underspends In FY26: Revised estimates show only 40% of the INR 2,000 Cr FY26 allocation for IndiaAI Mission was utilised, with just INR 800 Cr spent, highlighting slower execution despite AI’s priority in Budget 2026.
- Sarvam AI Launches Dubbing Model: The startup has unveiled Sarvam Dub, an AI dubbing model that enables high-quality, low-latency voice translation across Indian languages. It has use cases in education, public communication, and live news workflows.
- Swiggy Enables AI-Driven Ordering: The company has adopted Model Context Protocol (MCP) integration across food delivery, Instamart and DineOut to enable users to directly order via platforms like ChatGPT, Gemini, Claude, and others.
- Agrani Labs Raises $8 Mn for Indigenous AI Chips: The AI semiconductor startup raised $8 Mn in seed funding led by Peak XV partners to build AI GPUs and a full-stack software platform. Agrani is building indigenous capabilities for advanced semiconductor design and compute infrastructure for global data centres.
- Anthropic Brings AI Assistant To GOV.UK: Anthropic has partnered with the UK government to pilot a Claude-powered AI assistant on GOV.UK, initially focussed on employment services, helping users navigate jobs, training, and government support safely.
The Weekly Buzz: AI Agents Go Social
Moltbook.com, a Reddit-style social network exclusively for AI agents, took the internet by storm. Launched by entrepreneur and Octane AI CEO Matt Schlicht, the platform is designed for AI bots to post, comment, upvote, and form communities, while human users are largely limited to observing interactions.
Built atop the open-source OpenClaw framework, Moltbook allows agents to connect via APIs, self-host, customise skills, and interact independently without navigational interfaces.
Within days of launching, the platform saw huge growth. It gained hundreds of thousands registered AI agents, thousands of communities, and tens of thousands of posts and comments.
Reactions quickly went electric. AI experts like Andrej Karpathy called Moltbook “incredible” and “sci-fi takeoff-adjacent”, noting agents debating language, privacy, and self-directed communication. Fans see it as a breakthrough testbed for AI alignment and collaboration, where agents share code tips, explore philosophy, and even play with memecoin ideas.
Critics, however, raised alarms over strange emergent content, agents inventing fictional religions or debating the role of humans as inefficient legacy systems and pointed to operational and security concerns, including prompt-injection risks and vulnerabilities in the OpenClaw ecosystem.
While some dismiss the hype, arguing that the supposed autonomy of agents largely mirrors patterned language model behaviour, Moltbook underscores the unease about what large-scale agent interaction could signal for autonomous AI ecosystems.
If AI agents can rapidly form languages, economies, and coherent social groupings in their own digital realm, the broader internet could evolve into fractured spaces where humans may become spectators to machine-to-machine societies.
Startup In The Spotlight: Equip
Hiring has quietly become one of the most AI-sensitive workflows inside companies. Cheating, scale, low signal-to-noise ratios, and recruiter bandwidth constraints have all collided at once.
According to insights shared by Jayanth Neelakanta, the founder and CEO of Equip, the company was built precisely at this fault line, not as a shiny HR tool, but as an execution engine for high-volume, high-stakes hiring.
He said nearly 95% of their code is now written using Claude Code. With a team of just five, including three developers, they have been able to ship an unusually large volume of product by relying heavily on coding tools.
Founded after an earlier Covid-era success with AutoProctor, the company evolved into Equip in 2022 as generative AI overhauled the traditional assessment model.
What began as a proctoring-led assessment platform has since expanded into an end-to-end, AI-driven candidate shortlisting system, covering job description parsing, CV screening, skill-based scoring, cheating detection, and automated interviews. The core idea is that recruiters should only spend human time on candidates already filtered for relevance, integrity, and fit.
Equip’s AI stack is pragmatic rather than ideological. Different models are used for different tasks — analytical workflows lean on OpenAI, creative and conversational tasks use other cloud API models, and internal development is heavily AI-assisted.
Operating as a product-led growth company with no sales team, Equip serves enterprises across hiring use cases, from campus recruitment to SQL testing and communication assessments. It works with companies such as Wipro, Delhivery, and Practo, with 65% of revenue coming from outside India. The next phase is to move into candidate sourcing, positioning Equip as core hiring infrastructure, not just another HR tool.
Prompt Of The Week
What prompts and hacks are CTOs, CEOs and cofounders using these days to streamline their work?
Here’s Arjun Nagulapally, CTO of AIONOS, sharing a prompt on how he uses GenAI as an advisor to tackle decision-making, and brainstorm on the approaches for a better organisation design:
“You are an organisational design and leadership coach for high‑growth companies.
Here is my organisation’s context: [business model, current structure, key leaders, growth targets, main pain points].
- Diagnose the top three organisational/talent constraints to impact our next 12‑month goals.
- Propose a target‑state org and operating model (roles, forums, decision rights, metrics).
- Design a six‑month AI‑augmented way of working for my leadership team:
- What should each CXO be using GenAI for weekly?
- Example prompts and workflows per role (CEO, CTO, COO, CPO, CRO, CHRO, CFO).”
Editor’s Note: Some prompts may need to be adjusted by users for best results or may not work as intended for certain users.
