Fractal Analytics targets 25 startups to scale enterprise AI
The analytics firm is scouting for acquisitions in healthcare and agentic AI to prepare for a major shift in how businesses deploy large-scale models.
The short version
- Fractal is evaluating 20-25 startups for acquisitions with a preferred size of $50 million to $100 million.
- The company is targeting specific expertise in healthcare, agentic AI, and industrial verticals.
- CEO Srikanth Velamakanni predicts the cost of intelligence from foundation models could fall 93% over the next year.
- Fractal recently reported record Q2 FY27 revenue of ₹956.5 Cr, with healthcare growth reaching 79% year-over-year.
- The firm is seeking 'multiplier' deals that improve its 'right to win' rather than just adding new business units.
Why it matters
This signals a shift from AI experimentation to large-scale industrial deployment. For businesses, it means providers are aggressively building the infrastructure needed to make complex AI tasks cheaper and more reliable.
Aggressive M&A for AI Scale
Fractal Analytics is currently scouting the startup landscape to identify 20-25 potential acquisition targets. The company aims to strengthen its technical capabilities as it prepares for a significant increase in how large enterprises adopt artificial intelligence. This strategy focuses on securing specialized engineering talent and deepening ties with major cloud providers and frontier AI labs.
The company is not looking for a specific target but is open to deals valued up to ₹1,000 Cr. While there is no minimum price for a deal, Fractal prefers acquisitions in the $50 million to $100 million range. This follows a massive screening process where the firm evaluated approximately 800 potential targets over the last 12 months.
CEO Srikanth Velamakanni emphasized that these acquisitions are not intended to be simple growth levers. Instead, the goal is to create a multiplier effect that improves the company's 'right to win' in the market. The board is currently exploring a healthy pipeline of opportunities with sufficient cash on hand to move forward with transactions soon.
The Economics of Intelligence
A primary driver for this acquisition push is the rapidly declining cost of what Velamakanni calls 'fixed intelligence.' He noted that the cost of this intelligence has already dropped 48% quarter-on-quarter. Based on this trend, he predicts that the cost of running intelligence from a foundation model could fall by 93% over the next year.
As these costs decrease, large-scale AI deployment becomes much more attractive for major corporations. Fractal expects these economic shifts to make AI adoption meaningful for businesses as early as FY28. The company believes that as models become cheaper and more capable of handling long-horizon tasks, the impact will be visible in key metrics like revenue per employee.
To prepare for this shift, Fractal has reorganized its business into three distinct pillars. These include AI-led Transformation, AI Foundations, and AI Work & Workforce. This new structure is designed to help clients redesign their internal workflows and manage the governance of agentic AI, which refers to autonomous AI systems capable of taking actions to achieve goals.
Targeting Specialized Verticals
Fractal is specifically looking for startups that offer deep expertise in industrial verticals and have a strong local presence across different geographies. They are prioritizing areas like healthcare and agentic AI engineering. These specific niches are where the company wants to build deeper capabilities to stay ahead of competitors.
The company's recent financial performance highlights the importance of these sectors. In Q2 FY27, the healthcare and life sciences vertical saw a massive 79% year-over-year revenue growth. Meanwhile, the banking, financial services, and insurance sector grew by 30%, and the retail and consumer packaged goods sector rose by 19%.
The acquisition strategy is also a way to bolster the company's technology layer, known as Cogentiq. This platform serves as the common technology foundation across Fractal's three new business pillars. By acquiring specialized firms, Fractal can integrate unique engineering talent into Cogentiq to provide more robust solutions for complex enterprise problems.
Recent Growth and Expansion
Fractal's recent moves show a commitment to both organic growth and strategic 'acqui-hires.' In November, the company acquired PricingOne, a Geneva-based analytics firm, for 1.32 million Swiss francs. This transaction was described as an acqui-hire, a move intended to provide the company's subsidiary, Asper.ai, with a strategic foothold in the European market.
The company's financial health supports this expansion, as it reported record revenue of ₹956.5 Cr in Q2 FY27. This represented a nearly 20% increase from the same quarter the previous year. Net profit also saw a significant jump, more than doubling year-over-year to ₹79.2 Cr, despite some sequential fluctuations in profit and margins.
While investments in AI engineering talent have had a short-term impact on gross margins, the company attributes its overall success to operating leverage. Fractal continues to serve major global clients including Google, Citibank, Nestlé, and Wells Fargo. The company's goal remains to provide the infrastructure and expertise necessary for these giants to navigate the upcoming AI scale-up.
What to watch
- Which of the 20-25 startups Fractal is evaluating will eventually be announced as an acquisition.
- Whether the predicted 93% drop in foundation model costs holds true over the next 12 months.
- How the new 'AI Work & Workforce' pillar performs in terms of revenue growth compared to the healthcare vertical.
Try it
Look up the 'inference cost' of large language models to see how much it costs to run a single query versus training a model.
Sources and links
- ReportingInc42: the article we sourced
First reported by Inc42. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.