Investment research platform Bastion Research has announced that it will host its first in-person investor meet in Surat on 4 April 2026, focusing on helping investors develop a structured approach to analysing stocks.
The event will introduce participants to what the firm calls the “Accept–Reject Framework”, a method designed to identify red flags in company financials, governance practices and long-term business prospects before making an investment decision. The session will also include a detailed analysis of two listed companies, evaluated using Bastion’s internal research approach.
The Surat gathering will be held in an in-person format and is intended as a focused learning session for investors looking to sharpen their stock-selection process.
The announcement comes as the firm also hinted at an upcoming initiative centred on “asymmetric investment opportunities”—situations where markets may temporarily misprice businesses, creating potential opportunities for investors.
Alongside the event announcement, Bastion Research highlighted what it described as the largest peacetime capital deployment in technology history: the rapid expansion of infrastructure supporting artificial intelligence.
At the centre of this transformation are data centres, the physical facilities where AI models such as ChatGPT, Gemini and Claude are trained and operated. Despite often being described as “the cloud”, these systems actually run inside large industrial buildings filled with high-performance computers, storage equipment and networking infrastructure.
AI has dramatically increased the demands placed on such facilities. Compared with traditional computing centres, AI data centres require five to ten times higher power density per server rack, generate significantly more heat and rely on specialised chips produced by a small number of global manufacturers.
Industry analysts describe the AI data centre ecosystem as consisting of four essential components: chips, power, cooling systems and physical infrastructure. Each element represents a distinct investment opportunity, with power supply and cooling technology becoming increasingly critical as computing loads grow.
The development of a hyperscale AI data centre—a facility capable of running thousands of graphics processing units (GPUs) simultaneously—can take several years and involve billions of rupees in investment.
The process typically begins with site selection, where access to reliable electricity has become the primary consideration. Large facilities may require 100 to 500 megawatts of continuous power, prompting developers to prioritise locations with strong grid capacity or renewable energy availability.
Subsequent stages include securing regulatory approvals, constructing a “powered shell” building equipped with electricity, cooling and fibre connectivity, installing specialised hardware, and finally launching full-scale operations once tenants begin deploying computing equipment.
