Today in AI — 2 April 2026

Today's top AI news — curated links and commentary on the stories that matter for product builders.

·3 min read

The money flowing into AI has stopped looking like venture capital and started looking like infrastructure spending. Q1 2026 saw $297B in global startup funding with 81% going to AI, seed rounds are pricing like Series As, and the companies receiving that capital are burning it on power plants, chip fabs, and ad platforms. The question is no longer whether AI will be big — it's whether the economics can hold.

The capital supercycle

Four numbers tell the story: $297B in quarterly VC, 81% AI concentration, $40–45M seed valuations, and $7B in strategic bets from Nvidia alone in a single day. Investors are pricing rounds years ahead of revenue, and the biggest chip company on earth is buying equity stakes in its own supply chain rather than waiting for the market to sort itself out. This is what it looks like when an industry decides the future is already here.

AI finds its revenue model

ChatGPT ads hit $100M annualised revenue in six weeks with 600+ advertisers, and self-serve access drops the $200K minimum this month. Meanwhile, Gap is the first major retailer to let customers buy directly inside Google Gemini, and ElevenLabs is packaging voice into enterprise agent platforms. The interface layer is monetising fast — ads, commerce, and enterprise SaaS all at once.

The trust gap widens

A browser extension with 8 million users was silently harvesting conversations from every major AI chatbot, and Perplexity is facing a class-action suit for sharing user data with Meta and Google even in 'Incognito' mode. People are pouring their most sensitive thinking into AI interfaces, and the plumbing behind those interfaces is leaking in every direction.

Physical infrastructure and hardware bets

Meta's Hyperion campus will need 10 gas power plants and 7.5 gigawatts — more than South Dakota uses. On the model side, H Company's Holo3 topped desktop computer-use benchmarks with just 10B active parameters, proving efficiency gains are real. And Nothing is betting that AI belongs on your face and in your ears, not just on your screen.

The industry is simultaneously building revenue engines, burning through capital at record pace, and losing control of user data. If you're building on these platforms, the commercial opportunity is obvious — but so is the operational risk of trusting infrastructure that hasn't yet earned it.


Stay up to date

Get notified when I publish something new, and unsubscribe at any time.

More news