DW Newsletter # 198 – DeepSeek, a new contender in the AI arena

 Page, Text

Dear readers,

When DeepSeek first emerged in late 2023, it seemed like another ambitious startup in the crowded AI landscape. Founded by Chinese hedge fund manager Liang Wenfeng, DeepSeek began with modest attention, releasing AI models that showed promise but were overshadowed by industry giants like OpenAI, Meta, and Google. However, in December 2024, the company catapulted into the global spotlight with the launch of DeepSeek R1, an AI model that shook the foundations of the tech world. This development was not just a technological marvel but a geopolitical wake-up call.

DeepSeek R1 stunned the industry with its ability to rival the capabilities of market leaders, achieving near-parity with models like GPT-4 and Gemini. What set it apart, however, was its shockingly low cost. While US companies like Meta and OpenAI pour billions into AI research and infrastructure, DeepSeek reportedly developed its model for just $5.6 million, leveraging less advanced chips that bypassed US export restrictions. The revelation underscored a significant shift in the AI arms race: resourcefulness could rival raw financial power. The app quickly became the most downloaded on Apple’s App Store, disrupting markets and causing US tech stocks to shed nearly $1 trillion in value within days.

 Animal, Fish, Sea Life, Shark, Bird

The conversation shifted to national security as the US government scrambled to respond. DeepSeek’s emergence directly challenged US dominance in AI and called into question the effectiveness of export restrictions imposed by the Biden administration. Returning for his second term, President Donald Trump convened a high-stakes meeting with Nvidia CEO Jensen Huang to discuss tightening chip export policies. Nvidia, whose shares dropped 17% after DeepSeek’s success, found itself at the centre of the geopolitical battle, with its cutting-edge AI chips both coveted and restricted.

The Trump administration is now contemplating additional measures to curb China’s access to advanced technology, including halting Nvidia’s H20 chip exports, initially designed to comply with export controls. Lawmakers from both parties are urging a thorough review of US export policies in light of DeepSeek’s achievements. Reports have even suggested that DeepSeek may have illegally obtained US-restricted chips, prompting investigations by the Commerce Department.

For its part, DeepSeek has embraced the controversy. Liang Wenfeng has framed the company’s success as a triumph of innovation over-dependence on Western technology. However, questions about the company’s transparency linger. Critics argue that DeepSeek’s low-cost claims omit the broader expenses of research and development, while concerns about its alignment with the Chinese government’s strategic goals have fueled scepticism in the West. The fact that DeepSeek’s earlier models censored politically sensitive content only deepens fears about its role in state-led narratives.

DeepSeek’s rise has challenged assumptions about the cost and accessibility of cutting-edge technology, prompting both excitement and unease. While the US tightens its policies and grapples with its response, DeepSeek’s trajectory is proof that the race for AI dominance is no longer confined to Silicon Valley; it’s now a high-stakes game of global power and influence.

Related news:

deepseek nvidia

US President Donald Trump and Nvidia CEO Jensen Huang discussed the challenges posed by China’s advancements in AI, particularly focusing on the impact of Chinese AI company DeepSeek.

Stay updated on DeepSeek developments!

In other news..

DeepSeek sparks AI investment surge in China

Chinese investors are flocking to AI-related stocks, betting that the success of home-grown startup DeepSeek will propel China to the forefront of the AI race amid the escalating Sino-US technology conflict.

Security concerns lead to Australian ban on DeepSeek

Concern over potential security threats has prompted Australia to ban Chinese AI startup DeepSeek from all government devices, joining a growing list of nations restricting its use.

Visit dig.watch now for the latest updates and other topics!

Marko and the Digital Watch team


Highlights from the week of 31-07 February 2025

eu flags in front of european commission

An LLM project to enhance multilingual AI accessibility.

cryptocurrency in the us

Lawmakers aim to clarify cryptocurrency regulations as the industry continues to expand.

deepseek

Domestic technological innovation amid global competition.

flag of usa and china on cracked concrete wall background

The trade war is raging on through retaliatory measures targeting major businesses.

ai brain intelligent ai technology digital graphic design electronics ai machine learning robot human brain science artificial intelligence technology innovation futuristic

Global collaboration on sustainable and inclusive AI development amid geopolitical tensions.

salesforce jobs layoffs ai

Affected employees can apply for other roles, though the company has not specified which divisions will see the biggest cuts.

trump tiktok ban sale bytedance

The app’s future in the US remains uncertain.

scroll

Hidden text unveiled with AI and X-ray tech.

eu sets rules for e commerce

These changes could have significant financial and operational implications for the e-commerce companies involved.

whatsapp 892926 1280

A recent investigation has revealed that dozens of WhatsApp users across the globe were targeted by advanced spyware designed to infiltrate their devices without any action on their part.

deepseek

The app sparks national pride and curiosity in the country.


Reading corner

BLOG featured image 2025 20
www.diplomacy.edu

Imagine 100,000 people standing in utmost silence for 15 minutes, with their phone flashlights pointed at the sky. They pay tribute to the 15 innocent people who died under the rubble of the collapsed, newly reconstructed eave of the railway station in the city of Novi Sad, Serbia. 100,000 students and supporting citizens, standing still and united in palpable silence, in Belgrade, the capital of Serbia. Those present witnessed a moment that sent a chill up their spine.

download 10 1
dig.watch

In this digitally porous media world, information spreads faster than ever, empowering creativity but also unleashing avalanches of misinformation and deepfakes that blur reality. High-profile clashes, from Novak Djokovic boycotting…

BLOG featured image 2025 18
www.diplomacy.edu

Part 2 examines the UN’s funding paradox: 128 nations contributing 2.2% of the budget can outvote major donors. Can China replace US leadership? Is Security Council reform urgent? Explore power shifts, moral authority, and the future of multilateralism.

BLOG DeepSeek some trade related aspects of the breakthrough
www.diplomacy.edu

DeepSeek’s open-source approach promotes accountability and allows developing countries to leverage AI solutions, potentially levelling the playing field. The shift towards open-source models may also encourage smaller players to create tailored AI systems, fostering a collaborative environment rather than a competitive race. Consequently, trade and technological diffusion can significantly enhance global economic growth and innovation.

BLOG featured image 2025 19
www.diplomacy.edu

What if we could overcome overconsumption? Could Wikipedia’s approach of respect and collaboration offer a solution? Aldo Matteucci writes.

Google drops diversity-focused recruitment targets

Google has scrapped its diversity-based hiring goals and is reviewing its inclusion initiatives, aligning with other US businesses scaling back such efforts. Chief People Officer Fiona Cicconi confirmed the company would no longer set aspirational diversity targets, marking a shift from commitments made in 2020 to improve representation across leadership roles.

A goal set by CEO Sundar Pichai in 2020 aimed for 30% of leadership hires to come from underrepresented groups by 2025. Reports indicate Google had reached 60% of its five-year objectives by early 2024. However, recent regulatory filings show that the company removed previous statements about making diversity a core focus, reflecting its broader reassessment of DEI programmes.

Alphabet Workers Union criticised the move, calling it part of a larger trend against workplace diversity in the tech sector. As a federal contractor, Google also cited the need to comply with changing regulations and executive orders affecting DEI initiatives. Internal employee groups such as “Black Googler Network” and ‘Trans at Google’ will remain in place, continuing to inform company policies.

Other major technology firms, including Meta and Amazon, have also reduced their DEI commitments. Conservative groups have challenged corporate diversity policies following a 2023 US Supreme Court ruling against affirmative action in university admissions, increasing pressure on businesses to revise their approaches.

Google targets commercial quantum computing within five years

Google plans to introduce real-world quantum computing applications within five years, according to its head of quantum research, Hartmut Neven. The company’s prediction challenges Nvidia CEO Jensen Huang‘s estimate that practical quantum computing remains at least 15 to 30 years away. Google has been developing quantum technology since 2012 and recently announced a breakthrough that could accelerate the field.

Potential applications for quantum computing include advancements in materials science, drug discovery, and energy solutions. Google has already demonstrated quantum processors capable of solving complex problems far beyond the reach of traditional computers. A recent study published in Nature highlights new progress in quantum simulation, bringing the company closer to commercial viability.

The timeline for practical quantum computing remains uncertain, with predictions varying widely among experts. Nvidia’s scepticism caused a sharp drop in quantum-related stocks, but Google remains optimistic about achieving its goal. Governments and businesses continue to monitor quantum developments closely due to their potential impact on cybersecurity, finance, and healthcare.

Google expands Gemini AI models with cost-effective options

Google has introduced new versions of its Gemini AI models, including the budget-friendly “Flash-Lite,” to compete with lower-cost rivals such as China‘s DeepSeek. The updates include the public release of Gemini 2.0 Flash and the testing of a new ‘Pro’ model. Flash-Lite was developed following positive feedback on the previous Flash 1.5 version, with the goal of making AI more affordable.

Investor scrutiny has increased over the rising costs of AI model development. DeepSeek recently claimed to have spent under $6 million on training a model, significantly less than what US AI firms are believed to invest. The emergence of cheaper alternatives has influenced discussions at Alphabet, Microsoft, and Meta, with all three companies reaffirming their commitment to high AI investment.

Alphabet’s stock declined on Tuesday amid concerns over a planned increase in capital expenditure, which exceeded Wall Street expectations by 29%. Google’s pricing strategy for Gemini Flash-Lite sets its cost at $0.019 per million tokens, placing it between OpenAI’s cost-efficient model at $0.075 and DeepSeek’s current rate of $0.014, which is set to increase soon.

Bloomberg: Google drops pledge to avoid harmful AI uses, including weapons

Google has removed a key passage from its AI principles that previously committed to steering clear of potentially harmful applications, including weapons. The now-missing section, titled ‘AI applications we will not pursue,’ explicitly stated that the company would not develop technologies likely to cause harm, as seen in archived versions of the page reviewed by Bloomberg.

The change has sparked concern among AI ethics experts. Margaret Mitchell, former co-lead of Google’s ethical AI team and now chief ethics scientist at Hugging Face, criticised the move. ‘Having that removed is erasing the work that so many people in the ethical AI space and the activist space as well had done at Google, and more problematically, it means Google will probably now work on deploying technology directly that can kill people,’ she said.

With ethics guardrails shifting, questions remain about how Google will navigate the evolving AI landscape—and whether its revised stance signals a broader industry trend toward prioritising market dominance over ethical considerations.

China targets Apple and Google with antitrust scrutiny, according to Bloomberg

China’s antitrust regulator is reportedly preparing to investigate Apple’s App Store policies and fees, including its 30% commission on in-app purchases and restrictions on external payment services. The move follows recent measures targeting US businesses, including Google and fashion brand Calvin Klein, just as new US tariffs on Chinese goods emerged. Apple’s shares fell 2.6% in premarket trading following the news.

The investigation, led by the State Administration for Market Regulation, comes after ongoing discussions between Chinese regulators, Apple executives, and app developers over the past year. While neither Apple nor the Chinese antitrust regulator has commented on the matter, the move is seen as part of broader scrutiny of US companies operating in China.

In a separate development, Google was also accused of violating China’s anti-monopoly laws, with experts speculating the probe could be linked to Google’s Android operating system and its influence over Chinese mobile manufacturers. Additionally, China’s Commerce Ministry added PVH Corp, the owner of brands like Calvin Klein, to its “unreliable entity” list.

Google search to function more like an AI assistant

Google is set to transform its Search engine into a more advanced AI-driven assistant, CEO Sundar Pichai revealed during an earnings call. The company’s ongoing AI evolution began with controversial “AI overviews” and is now expanding to include new capabilities developed by its research division, DeepMind. Google’s goal is to allow Search to browse the web, analyse information, and deliver direct answers, reducing reliance on traditional search results.

Among the upcoming innovations is Project Astra, a multimodal AI system capable of interpreting live video and responding to real-time questions. Another key development is Gemini Deep Research, an AI agent designed to generate in-depth reports, effectively automating research tasks that users previously conducted themselves. Additionally, Project Mariner could enable AI to interact with websites on behalf of users, potentially reshaping how people navigate the internet.

The shift towards AI-powered Search has sparked debate, particularly among businesses that depend on Google’s traffic and advertising. Google’s first attempt at AI integration resulted in embarrassing errors, such as incorrect and bizarre search responses. Despite initial setbacks, the company is pushing ahead, believing AI-enhanced Search will redefine how people find and interact with information online.

Snap’s new AI model brings text-to-image to mobile devices

Snap has introduced an AI-powered text-to-image model designed to run efficiently on mobile devices, generating high-resolution images in just 1.4 seconds on an iPhone 16 Pro Max. Unlike cloud-based systems, this model operates entirely on the device, reducing costs while maintaining impressive visual quality. The company plans to integrate it into Snapchat’s AI Snaps and Bitmoji features in the coming months.

By developing its own AI model, Snap aims to provide users with more advanced creative tools while lowering operational expenses. The move aligns with a broader trend among tech companies investing heavily in AI to enhance their platforms. Previously, Snap relied on external providers like OpenAI and Google, but its in-house model gives it more control over future innovations.

Snapchat’s AI investment highlights the growing competition in mobile AI technology, with major players racing to deliver faster and more efficient features. As the company prepares to roll out these new capabilities, it remains to be seen how they will shape user experiences and engagement on the platform.

China targets US companies with new tariffs

China has introduced a series of measures targeting US businesses, including Google, farm equipment makers, and the owner of Calvin Klein, following the implementation of new US tariffs on Chinese goods. Among these measures, China launched an investigation into Google for potential violations of anti-monopoly laws, although no further details were provided. Despite its minimal presence in China, Google continues to collaborate with local advertisers.

China’s Ministry of Commerce also added US firms PVH Corp and Illumina to its ‘unreliable entity’ list, accusing them of actions that harmed Chinese companies. The companies could face significant sanctions, including trade freezes and restrictions on foreign staff. PVH has already been under scrutiny for its ties to the Xinjiang region.

As part of a broader response, China imposed 10% tariffs on US farm equipment, which could impact companies like Caterpillar, Deere & Co., and AGCO. Tesla’s Cybertruck, a model that has yet to receive regulatory approval, could also be affected. These new tariffs, set to take effect on February 10, signal an escalation in trade tensions between China and the US, extending beyond the tech sector.

These moves mark a significant increase in trade restrictions, building on previous actions taken under former US President Biden’s administration. Analysts suggest that these measures may be used as leverage, with the potential for de-escalation if either side chooses to back down.

Google parent Alphabet under pressure over AI spending and slowing cloud growth

Alphabet is set to face investor scrutiny over its heavy spending on AI as it prepares to report earnings. Slower revenue growth in advertising and cloud services has raised concerns, especially as competition in AI intensifies. Chinese startup DeepSeek’s launch of low-cost AI models has fuelled worries about an industry price war. Alphabet’s capital expenditure, estimated at $50 billion for last year, is expected to rise further in 2025 to support AI-driven search features and cloud expansion.

Google Cloud’s growth is forecast to slow in the fourth quarter despite high expectations. Analysts suggest that while heavy investment continues, efficiency gains have helped maintain profits. The company’s search and advertising business remains strong, with an expected 11.2% increase in revenue, though this marks a slight slowdown from the previous quarter. Competition from Amazon and TikTok continues to challenge Alphabet’s dominance in search advertising.

Political advertising linked to the US presidential election may have boosted Google’s revenue, following a similar trend at Meta. However, Meta’s cautious outlook for the first quarter has raised concerns about broader ad market trends amid economic uncertainty. Alphabet’s shares have climbed 7% this year after a strong rally in 2023, largely driven by confidence in its AI strategy.

Investors will closely watch whether Alphabet faces the same cloud business challenges as Microsoft, whose Azure growth slowed due to a shift in AI priorities. Google Cloud revenue is expected to rise by 32% in the fourth quarter, slightly down from the 35% growth seen previously but still outpacing Microsoft and Amazon. Maintaining momentum in AI while balancing cloud growth remains a key challenge for Alphabet.