Huawei's Bid to Build Egypt's AI Data Centres Triggers US Counter-Offer

Egypt has become the unlikely front line in the US-China contest over global AI infrastructure, after Huawei submitted a detailed bid to build the Egyptian government's AI data centres, a proposal serious enough to trigger direct US government involvement in assembling a competing offer.

According to Bloomberg, which reviewed documents and spoke to people familiar with the matter, Huawei's tender proposes exporting 1,408 of its top-end Ascend 950-series chips for an AI training cloud, plus 600 additional chips, either more 950-series units or the earlier 910B model, for two inference clusters, the infrastructure needed to actually run AI models once they have been trained. The company laid out a 12-month construction timeline in a proposal document running to 102 pages.

The scale of the offer matters less than what it represents strategically. In raw computing terms, 2,008 Ascend chips amount to roughly the capability of a few hundred top-tier Nvidia processors, orders of magnitude below what major Western cloud operators deploy routinely. But this would reportedly mark the first known export of Huawei's Ascend accelerators after more than a year of attempts to sell them internationally, following an earlier approach to buyers in the Gulf and Southeast Asia that did not result in a completed deal. Winning Egypt's contract would give Huawei a foothold in a region where the United States has already invested heavily in far larger AI data centre infrastructure in the UAE, and would demonstrate genuine foreign demand for Chinese AI chips at a moment when Washington has worked hard to constrain exactly that demand through export controls.

The proposal's partnership structure is drawing particular scrutiny. Huawei's bid pairs its hardware with iFlytek, a Chinese speech recognition and surveillance technology firm, and both companies have sat on the US Commerce Department's Entity List since 2019, Huawei since May and iFlytek since October, the latter specifically over its role enabling surveillance in Xinjiang. The joint proposal's scope extends well beyond a conventional AI training cloud. According to Bloomberg's reporting, it includes vehicle and individual recognition capability run against a national population database, along with what the documents describe as situation awareness, packaged as a single panoramic city-monitoring view. One slide in the presentation reportedly displayed the insignia of China's Ministry of Public Security as an illustrative example of basic video applications, a detail that has done little to ease concerns in Washington about what kind of infrastructure, and what kind of surveillance capability, Egypt would actually be adopting alongside the compute power.

Washington's response has been immediate and concrete rather than merely rhetorical. When the Trump administration learned of Huawei's pitch, the State Department contacted Nvidia, AMD and Microsoft about assembling a US consortium to compete directly for the same contract. A State Department spokesperson described Egypt as one of dozens of countries the US is engaging on AI technology leadership, a notable characterisation given that American AI infrastructure companies have not historically treated Egypt as a priority market, nor has the country featured prominently in the administration's broader AI diplomacy efforts to date. Bloomberg's reporting suggests this may be the first instance of the US and China competing directly for the same government AI infrastructure tender anywhere in the world, a distinction that elevates Egypt's decision well beyond its own borders as a signal of how this broader contest will actually play out at the level of individual national contracts.

Neither Huawei nor Egypt's foreign ministry responded to requests for comment, and Egypt has not indicated which bid it prefers or when a decision might be reached. That ambiguity is itself informative. Egypt sits in a genuinely difficult position: a nation of more than 100 million people with a decades-long Huawei commercial presence and deepening cooperation with China across trade and military ties, while simultaneously needing to weigh the diplomatic and economic consequences of favouring Chinese AI infrastructure over an American alternative, particularly given Washington's demonstrated willingness to treat AI chip export decisions as a matter of active diplomatic engagement rather than passive market competition.

For Gulf enterprises and governments already deep into their own sovereign AI infrastructure build-outs, this contest is worth watching closely as a preview of dynamics that could reach the region's own future procurement decisions. The UAE and Saudi Arabia have so far built their AI infrastructure predominantly around US chip suppliers, a choice reinforced by substantial American investment and, at times, active US government involvement in securing chip export approvals, including the UAE's own move into domestic chip production as it works to reduce dependence on any single foreign supplier over time. Egypt's tender demonstrates that Chinese AI hardware, even offered at a fraction of the compute scale Western firms deploy, is now a genuine option serious enough to draw direct US government intervention, not a hypothetical alternative that can be dismissed on capability grounds alone. The security and surveillance capabilities bundled into Huawei's proposal also underline a distinction regional decision-makers evaluating AI infrastructure partners increasingly need to make explicit: the choice is rarely just about chip performance or price, it is about what governance, surveillance and data-sovereignty architecture comes packaged alongside the hardware, and which government, ultimately, that architecture answers to.