AMD CEO Lisa Su in Shanghai predicts five billion daily AI users within five years

Five billion people using AI daily by 2030. That’s the forecast from AMD’s CEO Lisa Su, delivered with corporate certitude at an event in Shanghai. It’s a breathtaking number, a vision of total digital saturation. It’s also, in my view, a masterpiece of strategic misdirection, designed to make us look at the shiny, futuristic horizon while ignoring the cracked foundation beneath our feet.

Let’s be clear: this isn’t a prediction about technology. It’s a sales pitch for a future that companies like AMD need to sell. The “critical turning point” she mentions is less about the AI itself and more about the industry’s desperate need for a new, planet-sized growth narrative. The hype has become the product. By 2030, they don’t just want you using AI; they need you to be utterly dependent on it, running on their silicon, feeding their cloud infrastructure, and justifying the obscene capital expenditure currently reshaping global finance. This isn’t a vision of empowerment; it’s a blueprint for a captive market.

And what does “using AI” even mean? When my smartphone suggests an emoji, is that AI? When my spam filter catches a phishing email? If we define “AI” broadly enough—as the industry has every incentive to do—then we might already be there. The prophecy becomes a tautology: we will use more AI because everything will be rebranded as AI. The five billion figure isn’t about transformative human experience; it’s a metric of pervasive, often invisible, utility. It’s the number of potential data points, the scale of the subscription model, the breadth of the advertising platform.

The most glaring omission in this upbeat narrative is the colossal physical and ethical bill that comes due. Training a single advanced large language model can consume the energy of a small town. Running these models at a global scale requires a planetary network of data centers, guzzling electricity and water in regions often already stressed. To talk of five billion daily users without a concrete, radical plan for sustainable infrastructure is to build a cathedral on quicksand. It’s the tech industry’s classic move: innovate wildly, scale recklessly, and let society deal with the externalities later—if ever.

Then there’s the matter of who isn’t on this train. Five billion implies near-universal access. Yet billions still lack reliable internet, let alone the sophisticated devices and stable power needed for meaningful AI interaction. The “AI for all” story papers over deepening digital divides. The future Su describes will be stratified, not unified. It will be a world where a few build and control the models, a middle layer integrates them for profit, and the majority interact with them through walled gardens, their data extracted as the price of participation. This isn’t democratization; it’s the solidification of a new feudalism with algorithms as the lords.

Shanghai as the venue is no accident. It positions AMD at the nexus of the two dominant forces in this race: American silicon and the Chinese market’s voracious appetite for scale. But it also highlights the geopolitical paradox. The AI future is being promised as a global, borderless phenomenon, yet it is being built on a foundation of bifurcated supply chains, export controls, and competing national interests. The five billion will be interacting with systems that are fundamentally fractured along political lines, creating parallel digital realities. The “global AI adoption” is a convenient fiction.

Lisa Su is a brilliant executive navigating a brutally competitive landscape. She’s playing her part perfectly, offering the market what it craves: a story of limitless growth. But we’ve heard this song before. We were told the smartphone would connect and empower everyone. It did, but it also created an epidemic of distraction, a mental health crisis, and new vectors for surveillance. The blockchain was going to decentralize the internet; instead, it spawned rampant speculation and environmental waste. The pattern is to over-promise the utopian outcome while the dystopian trade-offs are quietly implemented in the backend.

So when I hear “five billion daily users,” I don’t hear a forecast. I hear a warning. I hear the sound of a future being locked in, where convenience is traded for autonomy, where the very fabric of our cognition and creativity is mediated by systems we don’t own and can’t audit. The hype isn’t just about the technology; it’s about engineering our consent to this arrangement. The real question isn’t whether five billion people will use AI by 2030. It’s whether we’ll have any meaningful choice in the matter, and what we’ll have lost by the time we realize we didn’t.