Major US AI firms are cutting costs to compete with Chinese developers, with OpenAI reducing prices for its GPT-5.6 model by 80%. As the industry shifts from rapid growth to profit focus, market leader DeepSeek is hiking prices ahead of a planned IPO. Meanwhile, India is focusing on expanding its data center infrastructure to support the growing demand for affordable AI computation.
The global artificial intelligence industry is witnessing a significant shift as the era of "growth at any cost" gives way to a focus on economic viability. Intense price competition, originally triggered by Chinese AI developers, has forced leading Western firms to adjust their pricing strategies. In a major response, OpenAI has cut the price of its GPT-5.6 Luna model by 80%, while Anthropic has launched its Claude Opus 5 at roughly half the cost of its previous flagship, the Fable 5. This race to the bottom is driven by enterprises that are increasingly prioritizing cost-effective AI deployment for large-scale production workloads over purely experimental features.
The Shift Toward Profitability
While price-cutting is currently the trend among US hyperscalers, the landscape is complex. DeepSeek, a Chinese firm previously known for aggressive price disruption, has pivoted its strategy. The company recently announced price hikes for its V4 models, citing the need to manage infrastructure capacity and prepare for an upcoming initial public offering (IPO). This change signals that even the most aggressive AI players are now under pressure to demonstrate financial sustainability rather than just market share growth.
For investors, this evolution suggests that the market is moving toward a more mature phase. The high price-to-sales ratios of many AI startups are being scrutinized, and there is growing concern about the risk of "massive capital destruction" if the multi-billion dollar investments in AI infrastructure do not yield sufficient financial returns for the tech giants funding them.
India’s Infrastructure Strategy
While the US and China battle for dominance in model pricing, India is carving out a different role in the global AI ecosystem. Rather than competing directly with frontier model developers, New Delhi’s strategy centers on building the essential "plumbing" for the AI revolution. Through the IndiaAI Mission, the government is providing subsidies for over 45,000 GPU units to boost domestic compute capacity. The goal is to provide a reliable and cost-effective platform for running AI computations for both domestic and international companies.
Local conglomerates and global hyperscalers are partnering to expand data center capacity, which is projected to reach 4–5 GW by 2030. This infrastructure-centric approach aims to leverage the country's cost-conscious talent pool and growing data center footprint. However, challenges remain, specifically the heavy dependence on imported hardware and the difficulty of integrating massive power demands into existing electrical grids.
Investor Monitorables
Looking ahead, market participants should watch the financial reports of global hyperscalers to see if massive spending on AI hardware is translating into actual revenue growth. In the Indian context, the key monitorables will be the speed at which new data center capacity is commissioned and the ability of local firms to secure consistent access to high-end hardware, as geopolitical tensions could impact the global supply chain for memory chips and processors.