Microsoft Breaks With OpenAI, Xi Targets US AI Dominance

Two pressure points on OpenAI’s position sharpened this week: its most important commercial partner is quietly working against it in enterprise sales, and China’s president used a global stage to frame US AI dominance as a threat to international order.

Key points

• Microsoft is instructing its sales teams to talk down OpenAI and Anthropic to enterprise customers, according to reporting from CXOToday, following CEO Satya Nadella’s public criticism of both firms.
• Nadella has accused OpenAI and Anthropic of being oriented toward “gobbling up the economy,” signaling a genuine philosophical rift rather than a tactical spat.
• At the World Artificial Intelligence Conference in Shanghai, Xi Jinping said AI “should not be a solo performance by a single country” and called for intergovernmental cooperation, with 29 countries signing on to a new AI cooperation body headquartered in Shanghai.
• A June poll found 69% of Americans support public ownership of roughly half of big AI’s equity, while OpenAI’s conversion proposal offered approximately 5% to its nonprofit arm, a gap that continues to fuel political scrutiny of the restructuring.
• Apple’s lawsuit against OpenAI, previously covered on this site, is proceeding with allegations that OpenAI’s hardware engineering division is structurally compromised, with Apple seeking an injunction that could disrupt OpenAI’s physical product ambitions.

Is Microsoft now OpenAI’s competitor?

The relationship between Microsoft and OpenAI has always been structurally awkward. Microsoft invested heavily, embedded OpenAI models across its product stack, and then watched OpenAI build out its own consumer and enterprise channels in direct competition. What’s new is the apparent decision to operationalize that tension at the sales level.

Nadella’s “gobbling up the economy” framing is notable because it mirrors the kind of language regulators and critics use. Whether this is a negotiating posture ahead of contract renewals or a genuine strategic pivot toward Microsoft’s own models remains unclear from the sourcing available. But telling a sales force to actively undercut a partner is a concrete action, not just an executive’s conference remark.

For investors tracking OpenAI’s path to an IPO, enterprise revenue concentration risk just became more visible. If Microsoft steers large accounts toward its own AI products, OpenAI’s standalone commercial trajectory takes a hit at exactly the moment the company needs to demonstrate durable, partner-independent revenue.

What does Xi’s AI pitch mean for OpenAI specifically?

Xi’s remarks at WAIC were framed around governance and sovereignty, but the commercial subtext matters. Chinese open-source models are winning customers abroad partly on cost and customizability. The source notes that companies like Siemens are adopting Chinese models as alternatives to the closed systems of US firms including OpenAI. Xi’s call to “oppose overstretching the national security concept” is a direct counter to US export controls, which restrict Chinese access to top-tier American AI.

Separately, both OpenAI and Anthropic are reported to have temporarily withheld recent model releases due to US government concerns about potential exploitation by hackers. That detail, if accurate, illustrates the growing friction between commercial release timelines and national security review processes. Neither company has publicly confirmed the specifics.

The new intergovernmental AI body, the World Artificial Intelligence Cooperation Organization, counts 29 member countries but notably excludes the US and EU. Its practical influence on OpenAI’s operations is likely minimal in the near term, but it does represent a competing governance framework taking shape outside the Western regulatory conversation.

The public ownership gap, still unresolved

The arithmetic here remains striking. A June poll cited by Medium put support for roughly 50% public ownership of major AI companies at 69% among Americans. OpenAI’s nonprofit conversion proposal, by contrast, offered around 5% to the public-interest side of the ledger. That mismatch has not closed, and it continues to give state attorneys general and congressional critics a politically durable line of attack.

This is not a new story, but it is a live one. OpenAI’s conversion from nonprofit to for-profit structure is still pending regulatory sign-off in multiple jurisdictions. The public sentiment data keeps the political cost of approving the current terms high for any official inclined to scrutinize the deal.

Today’s edition reflects a relatively quiet 48-hour news cycle for OpenAI specifically. The Microsoft sales posture and the Apple injunction threat are the two items with the clearest near-term implications for the company’s business and IPO preparation. The geopolitical backdrop is worth watching but does not present an immediate operational event.