Krutrim, India’s first generative AI unicorn, announced Tuesday it is shifting from AI model development to cloud services after a year of limited product updates and significant workforce reductions. The Bengaluru-based startup, valued at $1 billion in January 2024, told TechCrunch the pivot follows a business overhaul in late 2025 that included reallocating capital and talent while pausing chip design efforts.
The move reflects broader challenges facing AI startups attempting to build large-scale models while competing against well-funded giants like OpenAI and Anthropic. Krutrim generated ₹3 billion ($31.52 million) in revenue for fiscal year 2026, a threefold increase from the previous year.
Startup Funding Landscape Shows Mixed Signals
While Krutrim retreats from ambitious AI model development, other startups are securing significant backing for specialized applications. Kanvas Biosciences received new funding from the Bill & Melinda Gates Foundation to develop synthetic bacterial microbiomes delivered in pill form, targeting environmental enteric dysfunction affecting 150 million children worldwide.
The startup, founded in 2020 by CEO Matthew Cheng, uses machine learning and spatial imagery to create what it calls a “Google Maps” for the microbiome. Kanvas can pack 145 different bacterial strains into a single pill, significantly more than competing microbiome treatments that contain fewer than a dozen strains.
Meanwhile, European startups continue attracting investor attention despite not reaching the scale of companies like Mistral AI. TechCrunch identified 21 European startups worth watching, including Prague-based BottleCap AI and defense tech company Alta Ares, which develops AI-powered counter-drone systems.
Security Incidents Highlight Startup Vulnerabilities
AI evaluation startup Braintrust faced a security breach that forced the company to urge all customers to rotate their API keys after unauthorized access to an Amazon Web Services account containing customer secrets. The Monday email to customers confirmed “unauthorized access” but stated the company found no evidence of broader exposure beyond one impacted customer.
Braintrust spokesperson Martin Bergman said the company sent the warning “out of an abundance of caution” and confirmed a security incident without evidence of an actual breach. The startup locked down the compromised account, audited related systems, and rotated internal secrets while investigating the cause.
Funding Opportunities Remain Strong for Early-Stage Companies
TechCrunch’s Startup Battlefield 200 applications close May 27, offering pre-Series A companies access to $100,000 equity-free funding, global visibility, and direct VC feedback. The competition takes place at TechCrunch Disrupt 2026 in San Francisco, where 10,000+ attendees including top-tier investors will evaluate startups.
The program seeks ambitious early-stage companies building innovative products across all industries globally. Most selected companies are pre-Series A, though some Series A startups may qualify. Previous Startup Battlefield participants include Dropbox, Discord, Fitbit, Trello, and Mint.
Applicants need a functional MVP and clear product demo, with the TechCrunch editorial team looking for founders building with vision, execution, and real market impact. Thousands apply annually, but only 200 are selected for the main competition.
Geographic Disparities in AI Investment Continue
Krutrim’s struggles highlight the challenges facing AI startups outside major funding hubs. While the company achieved unicorn status early, it has cut more than 200 roles across multiple rounds over the past year and pulled its Kruti AI assistant app from app stores in April.
The startup’s limited public activity contrasts sharply with rival Sarvam, which participated in multiple sessions at India’s AI Impact Summit in New Delhi, showcasing new open-source models and commercial partnerships. Krutrim notably did not appear at the six-day event where global players like Anthropic, Google, and OpenAI participated.
Founded by Bhavish Aggarwal, who also leads ride-hailing firm Ola and EV maker Ola Electric, Krutrim initially positioned itself as a domestic alternative to Western AI models. The startup’s $50 million Series A at a $1 billion valuation reflected early investor enthusiasm for India’s homegrown AI ambitions, even as AI funding in India remains far smaller than in the United States.
What This Means
Krutrim’s pivot from AI model development to cloud services signals a broader market reality: building competitive large language models requires massive capital and technical resources that favor well-funded incumbents. The shift suggests that many AI startups may find more sustainable business models in specialized applications or infrastructure services rather than competing directly with OpenAI, Anthropic, and Google.
For investors, this trend indicates opportunities in vertical AI applications like Kanvas Biosciences’ healthcare focus, rather than horizontal model development. The continued strength of funding programs like Startup Battlefield 200 and targeted investments from foundations suggests capital remains available for startups with clear use cases and defensible market positions.
The security incident at Braintrust also highlights the operational challenges facing AI startups as they handle sensitive customer data and API keys, requiring robust security infrastructure that adds to operational complexity and costs.
FAQ
What caused Krutrim to pivot from AI models to cloud services?
Krutrim faced the economic reality of competing against well-funded AI giants like OpenAI and Anthropic. The company underwent a business overhaul in late 2025, cut over 200 jobs, and paused chip design efforts before announcing the cloud services pivot.
How much funding is available through Startup Battlefield 200?
Selected startups receive $100,000 in equity-free funding plus global visibility at TechCrunch Disrupt 2026. Applications close May 27, with 200 companies selected from thousands of global applicants across all industries.
What security measures should AI startups implement after the Braintrust breach?
The Braintrust incident highlights the need for regular API key rotation, restricted access controls, and comprehensive auditing of cloud accounts containing customer secrets. Startups should also have incident response plans and clear customer communication protocols ready.
