Blasé Capital Two Paths to AI

For most people, the choice about an AI tool is by accident, a recommendation from a friend, a shared link, or whatever came first. But the question of which to use is crucial. Global spending on AI is projected to cross $300 billion by 2027. Goldman Sachs estimates that AI can automate up to 25 per cent of work tasks. However, what has happened is not a single dominant model, but a split. At the centre of that split sit two of the most visible platforms: Claude and ChatGPT. From a user’s perspective, they reflect two distinct philosophies. OpenAI, which is behind ChatGPT, was founded on the idea to make AI available widely. Get it out there. Let people use it. The results will speak for themselves. ChatGPT became the fastest-growing product in history, reaching 100 million users in two months. Anthropic, which is behind Claude, was founded by former OpenAI researchers. Their principle is “responsible scaling,” or do not release something unless you are reasonably confident it will not cause harm.

The difference in personality is not accidental. ChatGPT has steadily expanded, and spans text, image generation, code assistance, file handling, and integrations across multiple tools. Claude’s emphasis is on longer context windows, document-heavy workflows, and a tone that prioritises clarity and caution. Although Claude and ChatGPT dominate conversations, this is not a two-horse race. The field has widened faster, and the new entrants are not chasing the same prize. Google’s Gemini is a quiet threat. It sits inside Gmail, Google Docs, Google Search, and Android. It is just there, open, when the user reaches for help. Grok, built by Elon Musk, is less filtered, more willing to engage with contested territory, and deliberately positioned against what its creators describe as the excessive caution of mainstream AI. Perplexity made a smarter strategic bet, and built an identity around the opposite. Meta plays a different game. Its Llama models are open-source, and Meta is not trying to win the chatbot race. It is trying to make the chatbot race irrelevant by handing the infrastructure to anyone who wants to build another infrastructure of their own.

In India, the pricing conversation is complicated. ChatGPT operates four distinct tiers. The free plan is available to anyone. ChatGPT Go was launched at Rs 399 a month. Above this sits ChatGPT Plus at Rs 1,999 a month, and ChatGPT Pro at Rs 19,900. Gemini offers its AI Pro plan at Rs 1,950 a month, with a budget AI Plus tier available at under Rs 500. Perplexity Pro lists at Rs 1,999 a month, but is currently free for a year to Airtel subscribers. Grok’s SuperGrok tier comes in at Rs 700 a month. Meta’s Llama, true to its open-source philosophy, costs nothing. Claude has no India-specific pricing. Claude Pro is billed in US dollars at $20 a month, which effectively costs Rs 2,200-2,400. As of writing, Claude remains the only major AI platform without rupee pricing or UPI integration. The pricing gap is not merely a commercial footnote. What is unfolding in India is less a product competition than a land grab. OpenAI chief executive Sam Altman recently suggested during his India’s visit that India could surpass the US in ChatGPT usage.

With more than 900 million internet users, mobile-first habits, and among the cheapest data costs in the world, India is the largest untapped AI market. Yet, data privacy sits at the centre of this. The Digital Personal Data Protection Act places new obligations on how data is collected, stored and processed. OpenAI nor Anthropic are not headquartered here. Both process data on servers outside India. By default, conversations with ChatGPT can be used to train OpenAI’s models unless users manually opt out, a setting deeply buried. Anthropic’s defaults are stricter. Meta’s open-source Llama models found quiet traction among developers and privacy-conscious organisations. They can be run entirely on a private server. The practical implications are already visible in how organisations are deploying these systems. A mid-sized law firm in Bengaluru uses Claude to review and summarise contracts. A digital marketing and content org in Mumbai runs ChatGPT to generate first drafts, headlines, and campaign briefs at volume. Claude is used when the task requires continuity and attention. The interaction is slower, more iterative, and focused on refining output.

In practice, users move between several. Instead of a single best tool, users are building informal stacks, choosing systems based on the nature of the task. What emerges is not a verdict but a map. ChatGPT for speed, versatility, and the widest toolkit. Claude for depth, long documents, and a more cautious disposition. Gemini for anyone inside Google’s ecosystem. Perplexity for those who need to verify rather than receive. Meta’s Llama for those who need control over where their data goes. In India, the choice is being made not by users but for them through telco bundles, promotional windows, and platform defaults. The models are moving faster than any comparison. What one system did better six months ago may be table stakes today. New versions arrive quarterly. Permanent capabilities evaporate within a product cycle. GPT-5’s launch is a useful reminder that more powerful does not mean useful.