top of page

AI Publications

Public·8 members

AI cold war

  1. DeepSeek has gained momentum in Emerging Market

  2. Despite some limitations, a free-source model attracts many companies

  3. China also offers much cheaper energy cost than US

  4. This results from long-term energy production investments

  5. Despite enormous amount of US investments, my opinion is that Data Centers in some sense lacks gain of scale (each question to answer is different)



FULL STORY OF DEEP SEEK:

Microsoft warns that China is winning AI race outside the West

https://www.ft.com/content/f7a5b184-1fef-4f02-b957-4c2b07adf91f


Cristina Criddle in San Francisco

Published Jan 13 2026


Microsoft has warned that US AI groups are being outpaced by Chinese rivals in the battle for users outside the west, as China combines low-cost “open” models with hefty state subsidies to gain an edge.


Brad Smith, Microsoft’s president, told the FT that the rapid adoption of Chinese AI start-up DeepSeek’s technology in emerging markets such as Africa underscores the competition American groups face around the world.


“We have to recognise that right now, unlike a year ago, China has an open-source model, and increasingly more than one, that is competitive,” he said. “They benefit from subsidisation by the Chinese government. They benefit from subsidies that enable [them] to basically undercut American companies based on price.”


Smith’s comments come as new research from Microsoft found that the release of DeepSeek’s R1 large language model a year ago helped accelerate the uptake of AI worldwide, particularly in the global south, due to its “accessibility and low cost”. 


That has also led to China overtaking the US in the global market for so-called open AI models, which are often free to use, modify and integrate by developers.


In contrast, US tech groups such as OpenAI, Google and Anthropic have instead focused on maintaining full control of their most advanced technology, profiting from it through customer subscriptions or enterprise deals.


Microsoft’s research, which is based on usage data from the tech group’s products, estimated the Chinese group has an 18 per cent share of the AI market in Ethiopia and 17 per cent in Zimbabwe.



Smith said African countries would need broader investment from “international development banks” or “lending facilities” in order to build data centres and subsidise electricity costs.


“If we rely on private capital flows alone, I don’t think that will be sufficient to compete with a competitor that is subsidised to the degree that Chinese companies often are, especially in those parts of the world.”


However, Bright Simons, vice-president at the Imani think-tank in Ghana and an expert on AI, said there was no “scientifically rigorous way” to determine whether DeepSeek was forging ahead in Africa, but that open-source Chinese AI systems offered cheap alternatives.


“Africans can’t afford very expensive solutions apart from open source, so you have to go to [Meta’s] Llama or Chinese options,” he said. They were also using homegrown small language models, such as Masakhane, a pan-African model, and the South African InkubaLM, Simons said.


Microsoft’s research also found that in countries where US technology products are limited or restricted, DeepSeek had gained a considerable lead, with a 56 per cent market share in Belarus, 49 per cent in Cuba and 43 per cent in Russia.



DeepSeek shocked Silicon Valley when it introduced its powerful AI reasoning model R1 last year, which it claimed was trained at a lower cost with less computing power.


DeepSeek is expected to release its long-awaited new AI model before the lunar new year holiday.


Microsoft’s research showed that AI adoption is concentrated in developed countries, with nearly a quarter of the global north using AI in the fourth quarter of 2025, compared with 14 per cent of the global south, and 16 per cent globally. 


Smith said the growing gap was a “cause for concern” and cautioned that “if we don’t address a growing AI divide, it’s likely to perpetuate and broaden the great economic divide between north and south”.


DeepSeek spreads across China with Beijing’s backing


He noted that this was a crucial battlefront in the US’s competitive race with China, and said more investment was required from private companies to build data centres and provide skills training, as well as investment from governments and financial institutions.


“What we do have is, as American companies, a stronger reputation for trust. We have access to better chips than the Chinese companies do . . . [but] you always have to compete on price,” Smith added.


He warned that a lack of attention to the uptake of AI in regions such as Africa, with young, fast-growing populations, could lead to the rise of systems that are not aligned with democratic values.


“If American tech companies or western governments were to close their eyes to the future in Africa, they would be closing their eyes to the future of the world more broadly, and I think that would be a grave mistake.”

16 Views
JA Soler
JA Soler
Jan 18

Helcio thank you for sharing. Your post is a sharp reminder that the AI race is no longer just about model quality — it’s about distribution, price, and geopolitics.


What Microsoft is highlighting here is uncomfortable but real: open(-ish) models + state subsidies + emerging-market focus is a powerful combo. DeepSeek didn’t “win” on raw capability alone; it won on accessibility and economics, especially where budgets, infrastructure, and energy costs matter most.


Meanwhile, US players (OpenAI, Google, Anthropic) have optimized for control, margins, and enterprise value — a rational strategy, but one that leaves space elsewhere. If you don’t show up with affordable, deployable options, someone else will.


The deeper issue isn’t “China vs the US”, it’s whether the global south becomes a first-class participant in the AI economy or a downstream consumer of subsidised tech. If infrastructure, skills, and power costs aren’t addressed, the market will naturally gravitate to whoever can undercut on price — values come later.


This is less a warning about DeepSeek (DeepSeek) and more a warning about strategy blind spots. In AI, trust matters — but only if people can afford the product.

bottom of page