What Five Frontier AI CEOs Said in Delhi
Superintelligence timelines, $200 billion in investment, and the question no one agreed on
The AI Impact Summit 2026 wrapped up in Delhi early in 2026. Nearly 300,000 attendees. Over 100 countries. The first global AI summit was held in the Global South.
But what caught my attention was who showed up and what they said.
Sam Altman (OpenAI), Sundar Pichai (Google), Demis Hassabis (Google DeepMind), Dario Amodei (Anthropic), and Arthur Mensch (Mistral), the people building the most advanced AI systems in the world, all spoke. And they didn't all agree.
Here's what stood out for me.
The timelines are compressing
Altman predicted that by the end of 2028, "more of the world's intellectual capacity could reside inside of data centers than outside of them." He's talking about early forms of superintelligence; AI that could run companies or conduct research at the level of top scientists.
Hassabis put AGI at "maybe the next five to eight years" and compared the coming changes to "10 times the impact of the Industrial Revolution, but happening at 10 times the speed."
Whether these predictions are accurate, I don't know. But these are the timelines the people building the technology are working with. That's worth noting.
India as the test case
Every major frontier lab announced India initiatives. OpenAI is opening offices in Bengaluru and Mumbai. Google committed $15 billion in infrastructure investment. Anthropic opened a Bengaluru office and added support for 10 Indian languages. Mistral praised India's "self-reliance" approach.
Over $200 billion in AI investment was committed to India at the summit.
This matters beyond India. It signals how frontier AI companies are thinking about emerging markets. India is the test case; 1.4 billion people, a large developer population, and existing digital infrastructure. If these investments work, expect similar playbooks in other developing economies.
Amodei made an interesting point: AI could deliver around 10% growth in developed economies, but potentially 25% in India due to "catch-up effects." If true, AI may benefit developing economies more than developed ones.
They don't agree
One moment went viral. It was when Prime Minister Modi gathered the tech leaders for a group photo and lifted Sam Altman's hand on one side, Sundar Pichai's on the other. Everyone joined in, except Altman and Amodei, standing side by side, who raised their fists instead of holding hands.
One investor posted: "When AGI? The day Dario and Sam hold hands."
The context: Amodei was a co-founder at OpenAI before leaving to start Anthropic. Last month, Anthropic ran Super Bowl ads that Altman called "clearly dishonest." At the summit, Amodei talked about "red lines"; autonomous weapons and mass surveillance; that Anthropic won't cross.
The point is that these companies are building similar technology with different philosophies. That's important to note as this will shape further conversations we'll see in the coming months and years.
The sovereignty question
Arthur Mensch from Mistral raised something you don't always hear from US labs: as AI becomes critical infrastructure, who controls it?
His argument: every organization running AI; governments, hospitals, public institutions; needs the "turn-on and turn-off button." Dependency on providers who could withdraw access isn't acceptable.
This is going to matter more as AI moves from an optional tool to essential infrastructure.
The bottom line
Five frontier AI leaders. Different predictions on timelines. Different approaches to safety. All are investing heavily outside the US and Europe.
If you're watching the AI space, this summit was a useful snapshot of where the people building the technology think it's headed, and where they disagree.
