The Prompt: Insights from OpenAI Global Affairs

Chinese Progress at the Front

Read more

a year ago · 16 likes · OpenAI Global Affairs

It’s a striking move—one that elevates Zhipu from an industry competitor to something more strategic. And it raises an obvious question: why now?

This post looks at what Zhipu is actually doing, how it fits into China’s broader AI strategy, and what OpenAI might be signaling—intentionally or not—about its own evolving role in the global AI landscape. Because beneath the details lies a deeper shift: AI is no longer just a race between models. It’s a contest over who defines the infrastructure, the norms, and the future of digital power.

So what is Zhipu, really?

Zhipu AI is one of China’s so-called “AI tigers,” a term the Chinese press uses for the country’s new generation of large model startups. It was spun out of Tsinghua University in 2019 and has since become a flagship in Beijing’s push for AI sovereignty. Backed by over $1.4 billion in state-linked investment, Zhipu sits at the center of a coordinated effort to develop and export a Chinese AI stack, from chips to code to governance.

It’s also currently the most aggressive of China’s foundation model firms in going global.

Zhipu is opening regional offices in the UAE, Singapore, Malaysia, and the UK. It’s launching joint innovation centers in Indonesia and Vietnam. It’s marketing “sovereign LLMs” to governments looking to control their own models and infrastructure. It’s building those systems on Huawei Ascend hardware and has received capital from funds like Saudi Aramco’s Prosperity7. One of its Dubai labs is focused specifically on content moderation and “AI safety.”

In April, Zhipu filed IPO paperwork with China’s securities regulator, making it the first of China’s “Six Little Dragons” to formally start the listing process. It’s burning cash fast - losing about 2 billion yuan last year against 300 million in revenue - but the move is as much about strategic positioning as it is about funding. Zhipu wants to be first to market, and first to claim the title of China’s “first large model stock.”

Zhipu has strong ties to Beijing’s AI establishment. It incubated at the state-backed Beijing Academy of Artificial Intelligence (BAAI), which pioneered the WuDao and WuJie model families. Its leadership regularly meets with senior CCP officials. It’s part of the Digital Silk Road strategy, which is trying to build an end-to-end Chinese tech ecosystem that can be exported globally - especially to Belt and Road partner nations.

Zhipu isn’t just selling AI. It’s selling a framework: A self-contained system of infrastructure, governance, and values that countries can plug into without relying on the West.

Why is OpenAI suddenly talking about Zhipu

OpenAI didn’t have to name Zhipu. It could have focused on DeepSeek, or the state-backed Tongyi Qianwen models from Alibaba. But it didn’t. Instead, it framed Zhipu as a global actor exporting not just models, but standards. A builder of AI systems that are, in OpenAI’s words, “responsible, transparent, and audit-ready” - a clear signal that Zhipu is pitching itself as the OpenAI alternative to the Global South.

Why highlight this now? Because OpenAI is changing too.

The company recently launched OpenAI for Government, signed a $200 million deal with the U.S. Department of Defense, and is leading the $500 billion Stargate UAE project with Oracle, Cisco, Nvidia, and SoftBank. It’s becoming more than a model company. It’s becoming infrastructure. And with that comes a shift in posture—from innovator to geopolitical player.

In that context, Zhipu’s expansion isn’t just a foreign competitor but a foil- a narrative frame that helps OpenAI position itself to U.S. and allied governments as the necessary counterweight. The responsible stack. The democratic stack. The stack you can trust.

Framing Zhipu as a CCP-aligned soft-power tool doesn’t just raise alarm, it also answers a strategic question that gets at when and how the defense establishment will work its way into the AI race: Why should Washington and its partners back OpenAI, and not just let the market decide?

OpenAI is saying: Because the other side already is.

Is Zhipu really that threatening?

In terms of model performance, customer traction, or economic scale? Not yet. In terms of reach? Possibly.

Zhipu is still early. Its IPO is months away. Its revenue is small. Its expansion into ASEAN and Gulf markets is promising but limited. And domestically, it’s under pressure. It’s fighting off competition from Alibaba, Baidu, and ByteDance - all of whom can undercut Zhipu on API pricing and bundle models with cloud services at massive scale.

There’s also the issue of stickiness. Many of the government clients Zhipu is courting - health ministries, education bureaus, local SOEs - may be eager to experiment with AI, but they’re also pragmatic. If U.S. or European providers can offer more powerful tools, or easier integration, or stronger diplomatic ties, loyalties can shift.

But that’s not the only risk.

The deeper risk, and the one OpenAI seems to be calling out, is that Zhipu is embedding systems before anyone else shows up. Infrastructure has a way of entrenching itself. Once you’ve trained your model on a Huawei box, integrated your content policies with a “safety lab,” and co-branded your national chatbot with a Chinese firm - you’re no longer just using their tools. You’re aligned with their stack.

That kind of alignment is hard to reverse.

The real contest: narrative vs. infrastructure

Still, what Zhipu is doing - and this is where it matters - is shifting the frame. It’s offering governments a story: You don’t have to rent your AI future from California. You can build your own - with us.

That’s the promise. And it comes bundled with hardware, model weights, safety protocols, and the political reassurance of co-ownership. Train your model on Huawei gear. Plug it into a national platform. Co-brand it. Regulate it your way. Zhipu isn’t just selling AI - it’s selling sovereignty, or at least a version of it that’s legible to countries caught between the U.S., China, and the demands of rapid digital modernization.

Is OpenAI scared of Zhipu? Doubtful. The positioning of the claim is more like “We see the game they’re playing - and we need to play it, too.”

Because underneath all the rhetoric about safety, hallucinations, and open models, the real competition isn’t about whose chatbot scores higher on a benchmark. It’s about who shows up first with a full stack. Who builds the default. Who embeds standards before anyone else gets a chance to.

That’s the world Zhipu is trying to build. And it may be the world OpenAI is trying to prevent. Either way, the fight is already on. And if we want to understand what’s at stake, let’s stop watching the models and start watching the infrastructure.

I finally got around to playing both the domestic and international versions of Black Myth: Wukong, and I found that they’re essentially the same. The gameplay is almost identical, with only a few tweaks in the international release aimed at appealing to global audiences. More than 90% of the game’s players are based in China, despite its success abroad. This highlights China’s growing ability to not only create globally competitive products but also to meet the demands of its own rapidly expanding digital market. Black Myth: Wukong shows that domestic success doesn’t necessarily require foreign markets—it can thrive without them, even as it finds international recognition.

This shift is also playing out in other digital media. I took my kids to see Ne Zha 2 this weekend, a movie that has dominated the box office in China, with over 90% of its audience coming from within the country. The cute Lotus god has an ungodly number of marketing tie-ins on Chinese social media right now. There’s plenty of media to consume in China, but I’ve never seen Chinese audiences turning to locally produced content so enthusiastically. There is a growing appetite among Chinese consumers for content that reflects their culture. As China continues to produce more of its own blockbusters, it’s likely that fewer foreign films and games will make it to the Chinese market in the future. I’m just happy to not be dragged along to another Avengers movie.

Looking at global digital export numbers for 2023, China is already a significant player. The country exported $207 billion in digital services last year, just behind the U.S. and India. But more important than China’s current standing is the fact that it’s on track to produce more and more of the content it once relied on importing. Rednote, a Chinese app made for the domestic market, unexpectedly gained traction internationally after the TikTok drama in the U.S. Similarly, Temu, a China-based e-commerce platform, has become a household name in the U.S. The rise of these products shows how China is establishing itself not just as a producer of cheap goods, but as a powerhouse of digital innovation that can thrive in foreign markets.

What’s fascinating about this shift is China’s ability to regulate its digital space in a way that other countries can’t. While Western countries often deal with a complex web of regulations across different states or regions, China has the advantage of centralized control. This has allowed the country to steer its creative industries in ways that cater to both domestic tastes and international markets. Black Myth: Wukong is a perfect example: while the Chinese version sticks to local standards, the global version ramps up fantastical elements to cater to international audiences. This dual approach, shaped by China’s unique regulatory environment, is something that gives Chinese creators an edge in the global market.

By controlling what gets produced, how it’s regulated, and how it’s consumed, China is not only producing more of its own digital content, but also setting the stage to become a major player in global exports. It’s no longer about simply exporting Chinese products—it’s about strengthening China’s digital economy and reducing its dependence on imports, particularly in industries like film, gaming, and e-commerce. The digital future for China isn’t just about making more things for export, but about meeting the demands of a rapidly growing domestic market and establishing global relevance at the same time.

As the country’s digital industries continue to mature, we’ll likely see more Chinese products—whether games, films, or apps—take a larger share of the global market. But more importantly, China’s growing digital production capacity will mean it will need to import less from the U.S. or other countries. What’s happening in China is a reflection of a much broader shift, where digital sovereignty and local production are taking the lead.

One of the best takes on this came from Biden’s FTC Chair, Lina Khan in a recent New York Times op-ed. Khan makes a point that’s rarely acknowledged in the U.S.: China’s so-called "crackdowns" on Big Tech may be less about stifling innovation than about fostering competition. In the U.S., the reality is that Big Tech often functions as a brake on innovation. The larger these firms get, the more they prioritize rent extraction over genuine competition, buying up rivals and cementing their dominance rather than taking risks on disruptive new ideas.

This is one of the paradoxes of capitalism: left unchecked, it tends to create monopolies that block the very competition that’s supposed to drive progress. That’s exactly what’s happened in the U.S., where tech giants, despite all their money and legal protections, are now looking increasingly vulnerable to a more nimble and dynamic AI scene emerging from China.

DeepSeek’s big reveal underscores this shift. It’s not necessarily that their models are superior to OpenAI’s or Google’s in sheer power, but the way they’ve shown how stacking AI models is proving to be highly effective. This kind of efficiency and iteration—the ability to build on existing technologies in novel ways—is something that entrenched monopolies struggle with. When you’ve spent years consolidating your power and eliminating competition, the last thing you want to do is introduce something that might disrupt your own dominance.

Then there’s the familiar “China stole our data” refrain, which is looking increasingly flimsy. If data scraping is the crime, then U.S. companies are just as guilty. ChatGPT and other models were trained on vast swaths of internet data, much of it scraped without explicit permission. The real issue isn’t data theft—it’s about who controls access to these models and, ultimately, the narrative around them. The U.S. and China have taken similar approaches to AI development, yet only one of them is constantly accused of playing dirty.

Meanwhile, the fate of NVIDIA is unclear, as is the effect on its stock price. Will U.S. policymakers continue to push for tighter export restrictions, hoping to choke off China’s AI ambitions by limiting access to cutting-edge semiconductors? The logic of containment seems to be that if AI is a nuclear weapon then chips are the enriched uranium. But recent developments suggest China may not need NVIDIA chips to keep advancing in AI. If that’s the case, then what exactly is the goal of these export bans? If China can keep innovating without U.S. hardware, then restricting chip exports is less about technological advantage and more about political posturing.

The DeepSeek moment isn’t just about a single company—it’s a symptom of a much bigger shift in global innovation. For decades, the U.S. has dominated the tech industry, assuming that its lead was unshakable. But now, with the rise of new AI players in China, that dominance is being tested in ways that Silicon Valley—and Washington—weren’t prepared for. The question now isn’t whether China can catch up. It’s whether the U.S. can get out of its own way long enough to stay ahead.

Last week’s TikTok debacle was the kind of chaotic, vaguely absurd moment that defines the modern internet. One minute, 170 million Americans had access to the app that revolutionized short-form video; the next, they were signing up en masse for Xiaohongshu, a platform best described as Instagram meets Pinterest with a heavy dose of Chinese internet culture. And then, days later, some of those same users started getting banned for violating Xiaohongshu’s content guidelines.

A perfect loop. The most online among us fled a government crackdown on one Chinese app by jumping to another, only to get hit with that government’s own flavor of censorship.

TikTok has always been an outsider in the American social media ecosystem. While Twitter, Facebook, Instagram, and YouTube have spent years refining their algorithms into an increasingly predictable paste of ad-friendly engagement, TikTok felt unpredictable, chaotic, alive. And that’s exactly why it took years for legacy media, business leaders, and politicians to figure out how to engage with it. Today, more than half of Americans use the platform. It has reshaped not only entertainment but also commerce, cultural trends, and even political discourse.

So what happens if it goes away?

The Winners

1. Meta, YouTube, and Snap: If TikTok disappears from the U.S., the biggest beneficiaries are the same American tech giants that have spent the past few years scrambling to replicate its success. Instagram Reels, YouTube Shorts, and Snapchat Spotlight were all launched as direct responses to TikTok’s dominance. And while none of them ever quite captured the same magic, the absence of TikTok would almost certainly drive users (and ad dollars) back to these platforms.

2. Washington’s National Security Hawks: Whether TikTok actually poses a national security risk is debatable, but its ban would be a massive win for those in D.C. who want to decouple the U.S. and Chinese tech ecosystems. The argument has always been that TikTok, owned by China-based ByteDance, could be used as a tool of influence or surveillance by the Chinese Communist Party. That concern isn’t entirely unfounded—China’s internet laws do require companies to cooperate with government requests—but it also raises serious questions about what counts as an acceptable level of risk in a global internet economy.

3. Oracle: This part hasn’t been discussed nearly enough. Oracle is TikTok’s biggest U.S. cloud partner, hosting vast amounts of the platform’s data on its servers. But Oracle also has deep ties to the U.S. government, and it’s not about to pick a fight with Washington. If the ban is enforced, ByteDance won’t have much of a choice—TikTok's U.S. data infrastructure depends on Oracle, and Oracle isn’t going to ignore a Supreme Court ruling. Whether ByteDance truly intends to fight this, or if it’s just making noise before ultimately divesting, remains to be seen.

The Losers

1. Small and Mid-Sized Businesses: TikTok isn’t just for dance trends and viral challenges. It’s been a lifeline for thousands of small businesses, many of which wouldn’t exist without it. The platform’s algorithm is eerily good at surfacing niche products to exactly the right audience, making it a powerful tool for independent brands. In 2023 alone, TikTok contributed $5.3 billion in tax revenue to the U.S., driven in large part by small businesses. Those businesses aren’t moving to Instagram Reels overnight.

2. Creators Who Built Their Careers on TikTok: TikTok didn’t just create a new type of content; it created a new type of creator. Many of the biggest influencers on the platform built their followings from scratch, in a way that wasn’t possible on older social networks. While some will successfully migrate to YouTube or Instagram, others will struggle to rebuild their audience in an ecosystem that isn’t designed for discovery in the same way.

3. The Idea of an Open Internet: TikTok has been operating in the U.S. for seven years. If it was truly a national security threat, why is it only now facing an existential crisis? The reality is, this sets a precedent. The U.S. government can now ban an app that’s too competitive for Silicon Valley, too popular among young people, or too politically inconvenient. If Washington can justify banning TikTok, what’s stopping a future administration from banning any platform that doesn’t fit its interests?

And it’s not just TikTok that could be at risk. ChatGPT, for instance, is already being discussed in the context of regulation, particularly around misinformation, AI-generated content, and its broader impact on the workforce. If a precedent is set that apps with security or content moderation concerns can be banned, how long before other platforms—whether AI-driven tools or social networks—face similar scrutiny?

The 小红书 Detour

For a brief moment, the chaos of the TikTok ban led to something kind of beautiful. Thousands of Americans flooded Xiaohongshu (“Red Note” though I’ve always preferred “Little Red Book”), where Chinese creators—many of whom also post on Instagram and YouTube—were waiting for them. It was one of the rare moments of direct online interaction between Chinese and American netizens, a reminder that despite political tensions, the internet still has the potential to connect people.

And then, in true Chinese internet fashion, many of those Americans started getting banned. Xiaohongshu is built for China’s heavily controlled online space, and its content moderation rules are very different from TikTok’s. American users quickly learned that what flies on TikTok—or even Instagram—doesn’t always fly on a platform built under Beijing’s watchful eye.

But Xiaohongshu was never going to be a long-term replacement. Its reach is limited, its advertising infrastructure isn’t designed for international businesses, and onboarding a new platform is a huge ask for marketing teams, most of whom have no experience with the Chinese side, much less with the complexity of working directly with foreign entities. Who’s going to pay all the newly minted American 网红？Like most observers, I predict Xiaohongshu is a fun experiment, but not a real shift in the social media landscape. Too bad.

The Bigger Picture

TikTok’s potential ban isn’t just about one app. It’s about the future of the internet. Who gets to control it? What counts as a security threat? How much power should the government have over the platforms people use to communicate, create, and do business? These are the real questions at stake. And regardless of what happens to TikTok, they aren’t going away anytime soon.

Among the attendees of the GoGlobal Forum 2024, hosted by EqualOcean in Shanghai last month, this challenge was front and center. The forum, which brought together business leaders, investors, and policy analysts, tackled the big question: how can Chinese companies thrive in an era where geopolitical risk is as inevitable as it is unpredictable?

From Globalization to Regionalization

The era of "hyper-globalization," where goods and ideas flowed freely across borders, has given way to something far more constrained. Trade blocs like the CPTPP and RCEP, along with initiatives from BRICS (or whatever the acronym is now), are carving out regional spheres of influence. Meanwhile, the U.S. is reshaping its own trade policies, creating a system designed not for efficiency but for security—prioritizing alliances with "friendly" nations and restricting access to critical technologies.

For Chinese companies, this new world order demands a shift in mindset. The idea of a singular global strategy is obsolete. Instead, success lies in understanding and adapting to the unique rules of each region. A case in point is Latin America, particularly Mexico.

The Opportunity in Mexico

Mexico represents a strategic beachhead for Chinese firms looking to enter North America. It’s close to the U.S., boasts a strong manufacturing base, and operates within the framework of the USMCA. But these advantages come with caveats. Trade policies are in flux, and regional dynamics are often unpredictable. Companies must navigate a complex web of local regulations and workforce challenges.

Speakers stressed the importance of localized globalization - a concept cribbed from a 2019 report from McKinsey - that emphasizes integrating into local markets while retaining global ambitions. It’s a balancing act, and getting it right is essential for companies aiming to establish trust and credibility in foreign markets.

The Resilience Playbook

One of the most compelling discussions at the forum centered on how companies can build resilience in the face of geopolitical risks. The speakers outlined a framework that includes five key pillars:

1. Flexibility in Business Models: Supply chains must be agile enough to withstand disruptions.

2. Reputation Management: In a world of increasing scrutiny, maintaining a positive corporate image is non-negotiable.

3. Regulatory Savvy: Understanding and adhering to local rules is critical for avoiding operational hiccups.

4. Technological Investment: From automation to data analytics, technology is a crucial tool for navigating complexity.

5. Financial Foresight: Companies need to prepare for currency volatility, sanctions, and shifting trade tariffs.

This decentralization - how companies can move from centralized to decentralized and distributed operational models - is key to building the kind of flexibility needed to weather geopolitical storms.

What the U.S.-China Relationship Means for Business

It’s impossible to talk about geopolitics without addressing the elephant in the room: the strained relationship between the U.S. and China. Policies under both Biden and Trump have profoundly impacted Chinese companies operating internationally.

The Biden administration’s Inflation Reduction Act, for example, effectively penalizes electric vehicles that use Chinese-made batteries by excluding them from subsidies. Similarly, the “small yard, high fence” approach focuses on safeguarding critical technologies and resources. These measures, though not explicitly anti-China, create significant challenges for Chinese firms. And with the imminent return of Trump to the White House, the stakes are about to rise further.

The takeaway? Companies need to monitor these developments closely and adapt proactively. This isn’t just about reacting to policies—it’s about anticipating them.

Collaboration Over Conquest

What stood out most at the forum wasn’t just the analysis but the tone of the conversation. One attendee, a global marklet analyst at Oppo, told me it was eye-opening to see geopolitical conditions framed as opportunities for collaboration and cooperation, rather than the "conquest" mindset that typically characterizes Chinese firms' overseas expansion plans. Instead of treating international markets as arenas for domination, the emphasis was on integration—working with local stakeholders, adapting to regional nuances, and finding mutually beneficial paths forward.

From Efficiency to Security

Another recurring theme was the shift in global priorities. The old trade orthodoxy—maximize efficiency, lower costs, and scale production—has been overtaken by a new mantra: prioritize security and resilience. This shift has reshaped supply chains and trade relationships. Nearshoring, friendshoring, and even "ally-shoring" have become the strategies of choice for many nations.

For Chinese companies, this means aligning their strategies with the realities of a fractured global economy. It’s not enough to have a great product or competitive pricing—success now depends on understanding the geopolitical undercurrents shaping the markets they enter.

The Way Forward

Cautious optimism is key. After all, how can things be worse than they already are? The challenges are real, but so are the opportunities. I heard a Chinese proverb from Li Bai invoked more than once: “长风破浪会有时”—“Long winds and breaking waves will eventually lead to success.” Yes, the waves are high, and the winds are unpredictable, but with the right tools, businesses can navigate them.

In the end, the message was clear: the path forward isn’t about avoiding risk—it’s about managing it. Companies that can embrace collaboration, build resilience, and adapt to regional dynamics will not only survive in this new global order but thrive.

The US Dominates AI Funding — But for How Long?

In 2023, US startups captured 71% of global AI equity funding — a significant share, considering how fragmented the AI landscape is across continents. Europe and Asia lag far behind with only 14% and 13%, respectively.

This massive funding influx has allowed US firms to build infrastructure, scale AI models, and drive research that is pushing the field forward. The lion’s share of this investment is funneled into areas like machine learning frameworks, cloud-based AI services, and, of course, the training of massive LLMs like GPT and its successors. The speed at which US companies like OpenAI, Google, and Anthropic have advanced in this field is a direct result of their ability to attract capital.

That being said, funding is just one piece of the puzzle. AI scalability, driven by cloud infrastructure and access to high-performance GPUs (like Nvidia’s A100), is also a critical factor. The US has a stronghold in both, but China's growing cloud infrastructure and its burgeoning AI hardware market may soon give it an edge.

Talent: Still Centered in the US, But China Is Catching Up

In addition to funding, the US has another advantage: AI talent. Over 40% of the world’s AI companies right now are based in the US, which creates a concentrated innovation ecosystem. This is no different from previous tech booms, where graduates from MIT, Stanford, Carnegie Mellon and the like produce world-class researchers in deep learning, computer vision, and NLP.

But China is making substantial strides. In 2023, Chinese universities and research institutions published nearly as many papers on AI as their US counterparts. This surge in academic output is backed by significant state and private investments in AI research. BAT are not just developing cutting-edge models, they’re also scaling up AI talent by offering competitive salaries and recruiting top-tier scientists from around the world.

China’s LLM Push: The Next Big Threat to US Dominance

China’s biggest challenge to the US’s AI lead is in the development of large language models (LLMs). While the US is home to some of the most well-known LLMs, including OpenAI’s GPT-4 and Google’s PaLM, China is catching up quickly with its own offerings.

Alibaba’s Qwen2, for example, has been making headlines as one of the top models on Hugging Face’s leaderboard. This model, along with others like Meta’s LLaMA-3 and Cohere’s Command R+, is pushing the boundaries of performance, particularly in natural language understanding and generation. What makes China’s strategy particularly interesting is its focus on open-source AI — a space where the US has yet to fully capitalize.

Chinese tech giants are actively building and investing in open-source LLMs to give developers and startups access to powerful tools without being locked into proprietary models like GPT-4 or BERT. This could fundamentally change the AI landscape by democratizing access to high-quality AI models, while also creating new competition for US-based companies.

The Infrastructure and Innovation Race: Who Has the Edge?

While the US still has the lead in AI funding and talent, China is focusing heavily on AI infrastructure to catch up. China’s growing dominance in AI-specific hardware, including AI chips designed by companies like Huawei and SMIC (Semiconductor Manufacturing International Corporation), gives it a potential advantage in training large models without relying on US-based providers like Nvidia.

Furthermore, China’s state-backed funding for AI startups and research centers is designed to foster rapid development of AI technologies at scale. The US has its private sector giants, but the Chinese government’s ability to direct large sums of investment into strategic areas — including AI models, hardware, and talent — could create a more resilient ecosystem in the long run.

Can the US Hold Its Lead?

The US is still leading in several key areas, particularly when it comes to capital and talent. However, China is closing the gap, especially in AI infrastructure and LLM development. Over the next 3-5 years, we’ll likely see both countries continue to pour significant resources into AI, creating a more competitive landscape.

This competition is particularly important for industries like healthcare, autonomous vehicles, finance, and national security, where AI will play a pivotal role in innovation and economic growth. The next big question: Can the US hold its lead in LLMs, or will China’s investments in open-source AI and hardware give it an edge?

As the arms race heats up, we’ll continue to see which country can truly dominate AI — and how this competition will shape the future of tech, global power dynamics, and innovation.

Subscribe now