E-COMMERCE 2030
E-commerce $10 trillion: 1 in 3 dollars via feed or agent.
Key figures
Global retail e-commerce 2025: $6.9 trillion, 21% of retail.
Consensus $9 to $10 trillion by 2030, about 23% of retail.
Social commerce, buying inside TikTok, instagram, wechat: $1.2 trillion in 2025, $1.5 to $2.3 trillion by 2030.
TikTok Shop is already the size of eBay.
AI agents completing purchases by 2030: 15 to 25% of US online sales, $3 to $5 trillion worldwide, on the consultancies' count.
My call: 1 in 3 online dollars is bought in a feed or by an agent by 2030, up from about 1 in 6 today.
E-commerce is the online channel inside retail: $6.9 trillion in 2025, 1 dollar in 5. Consensus: $9 to $10 trillion by 2030, still barely 1 in 4. Big number, slow share. Wrong story.
The real shift is not online versus offline. It is website versus everywhere. Search becomes chat. Browsing becomes a feed. Checkout moves into TikTok, WhatsApp, marketplaces and AI agents. The growth lands in the feed and the agent, not the cart.
My call: by 2030, 1 in 3 online dollars is bought in a feed or by an agent, up from about 1 in 6 today. Social commerce runs $1.2 trillion now and $2 trillion by 2030, TikTok Shop is already the size of eBay, and the consultancies count $3 to $5 trillion of agent-led buying by 2030. The website stops owning the sale.
The gatekeepers know it. Amazon won a court order blocking Perplexity's shopping agent while building agents of its own. Sellers say omnichannel. The tape says most brands still force you through a slow website, a broken search box and a checkout that demands an account. That is not commerce. That is a leak.
The warning: agents make products easy to compare and brands easy to replace. If your catalog is messy, your pricing opaque and your proof weak, the agent does not fall in love with you. It buys the cheapest acceptable answer. By 2030 the winner is not the prettiest website. It is the product easiest for humans and machines to find, trust, buy and return.