TL;DR
OpenAI faces its greatest competitive pressure since ChatGPT’s launch as Google’s Gemini 3 outperforms GPT-5 on key benchmarks. With Alphabet approaching $4tn market capitalisation and Anthropic’s enterprise business growing rapidly, the $500bn start-up is grappling with soaring costs and talent retention challenges.
A New Competitive Landscape
Three years after ChatGPT transformed the technology industry, OpenAI’s commanding lead in artificial intelligence is under unprecedented strain. Google’s release of Gemini 3 last week marked a turning point—the model is considered to have leapfrogged GPT-5, achieving gains that have eluded OpenAI in recent months.
“It’s quite a strong difference with the world we had two years ago where OpenAI was leading ahead of everyone else,” said Thomas Wolf, co-founder of Hugging Face. “It’s a new world.”
Even before Gemini 3’s launch, CEO Sam Altman warned staff in a memo to “expect the vibes out there to be rough for a bit” amid “short-term competitive pressure.”
Google’s Remarkable Turnaround
A year ago, many had written off Google’s efforts to narrow OpenAI’s lead. Fears that ChatGPT would cannibalise its search business left Alphabet’s shares lagging behind Big Tech rivals. But Google’s turnaround began with a confident showing at its IO developer conference and the viral success of its Nano Banana AI photo-editing tool.
The Gemini mobile app now has 650mn monthly users, up from 400mn in May. Alphabet’s market capitalisation is approaching $4tn for the first time, driven by confidence that Google can combine its positions in search, cloud infrastructure, and smartphones to serve AI capabilities to billions.
Critically, Google trained Gemini 3 using its own custom tensor processing chips, eliminating reliance on costly Nvidia hardware. Koray Kavukcuoglu, DeepMind’s CTO, said: “I think we have a unique approach there.”
Looking Forward
OpenAI faces a multi-front challenge. It has pledged $1.4tn over eight years on computing power—orders of magnitude more than current sales—requiring partners to use debt financing. Critics suggest the company has overextended itself, pushing out products from coding tools to video app Sora at breakneck pace.
Meanwhile, Anthropic’s Claude has built a strong enterprise following, with its coding tools widely considered best-in-class. Salesforce CEO Marc Benioff declared after testing Gemini 3: “I’ve used ChatGPT every day for 3 years… I’m not going back.”
With 800mn weekly users, ChatGPT remains dominant in overall usage. But time spent chatting with Gemini now exceeds ChatGPT, according to Similarweb data. The AI race has never been more competitive.
Source: Financial Times