Ready, set, code: the AI race is on
China just hosted a ‘Global Solidarity in the AI Era’ conference in Shanghai. And you know who was there? Yes, back-flipping robots. Yes, boxing robots. Okay yes, beer-serving robots, too. But also, Intrigue’s very own co-founder Helen Zhang, among thousands of other delegates from all around the world. A robot also scooped ice-cream.
So here are the top four quotes you need to know, and why:
“We should strengthen coordination to form a global AI governance framework that has broad consensus as soon as possible” — China’s premier, Li Qiang
That’s been the main headline this week: China’s call for global rules around AI. Supporters argue it reflects China’s determination to manage global AI risks, promote genuine multilateralism, and include smaller and developing countries at the table.
But critics argue it reflects China’s determination to shape global rules in its own favour, undermine Western regulatory leadership, legitimise its own authoritarian-led AI model, and contrast this ‘responsible leader’ branding against a disruptive US.
But what type of global AI rules does China want…?
“We must choose the open source pathway for the sake of humanity and the Global South” — That was Li Qiang again
Li warned, “if we choose closed source AI, then the future of power rests only with the few, the elites.” Supporters argue this open-source preference reflects China’s pledge to empower the Global South, and democratise the benefits of AI.
But of course, critics argue China just wants to burnish its champion of the poor branding, undermine US-dominated proprietary AI systems, and promote China’s own AI models in markets across the world.
And Li got an amen from an intriguing new voice…
“Unlike OpenAI's closed system, we adopt an open strategy to advance science and technology" — The CEO of Z.ai, Zhang Peng
Remember January? Everyone was hyperventilating at DeepSeek’s claims that it matched or exceeded OpenAI models using just a fraction of the US player’s chips, cash, and time.
Well grab your paper bag, because while DeepSeek’s CEO laid low at Shanghai this week, local rival Z.ai claimed its own GLM-4.5 model is now even better than DeepSeek’s! The world is still scrutinising Zhang’s big claims, including that he pulled off this feat using just eight Nvidia H20 chips (ie, the slower chips specifically designed to fit US chip controls).
But what does the US have to say about all this?
"Winning the AI race is non-negotiable" — US Secretary of State, Marco Rubio
Don't worry, you didn't miss a Rubio trip to speak at China’s AI conference. Rather, the White House released its own AI thinking, wrapped in a 90-day action plan focused on the three pillars of 1) innovation, 2) infrastructure, and 3) international dominance.
It aims to remove regulatory hurdles, ensure any open-source AI aligns with US values, and defend US AI solutions as the global go-to (while curbing China’s access).
The one thing these two competing visions have in common? China only appears twice in America’s vision, while the US doesn’t get directly mentioned in Li’s speech at all.
They’re one other’s elephant in the room, wrestling with — if not yet naming — the other.
Members-only analysis
Intrigue’s Take
Get full access to Jeremy, John and Helen’s unvarnished takes on the world and what it means for you.

