DeepSeek, Alibaba, and a Welcome AI Horse Race
China's AI challengers have flipped chip constraints into a 53x cost advantage, offering supply chains new affordable scenario planning and resilience options in the face of our current trade storm. As dual innovator-operators like Alibaba light the path to genAI value in supply chains, a new AI leadership formula emerges: relentless focus mixed with the drive to discover.
On January 20, Chinese AI startup DeepSeek sent tremors through Silicon Valley when its open-source AI model, R1, dethroned OpenAI's ChatGPT from its perch atop the Apple App Store. The news triggered a market rout as investors fled from once-hot AI GPU chip makers like NVIDIA.
Why the rout? R1’s reasoning capabilities are on par with ChatGPT’s most advanced model, but R1 was reportedly developed for under $5.5 million (some argue $30 million+ including pre-training), a mere fraction of OpenAI’s estimated hundreds of millions. Fewer chips, lower cost, and potentially far lower overall chip demand. The market's reaction was swift and merciless.
Creativity, Meet Constraint
Whether or not these cost differentials withstand engineering scrutiny, we're witnessing an AI efficiency revolution that a few of us saw coming – but most did not.
US export restriction on GPU chip sales to China left DeepSeek working with fewer older-generation chips, sparking the development of Reinforcement Learning with Human Feedback (RLHF), which discovers reasoning patterns organically. The result? DeepSeek was able to improve inference efficiency by over 80% and undercut OpenAI's price by a staggering 53x. American firms were left reeling. VC Marc Andreessen dubbed this AI's "Sputnik moment," invoking Russia's 1957 satellite launch that ignited the Space Race.
The analogy fits, except this race is now swarming with satellites.
Barely five days after DeepSeek's bombshell, Chinese tech giant Alibaba unleashed its Qwen 2.5-Max model, surpassing DeepSeek on all key performance metrics. Somehow, a company with its roots in online retail had morphed into an AI powerhouse. The transformation was head-spinning.
A Threat – and a Gift
The DeepSeek and Alibaba announcements merely preview the coming Chinese AI surge. Tech titans Alibaba, Baidu, Pinduoduo, Tencent Holdings, JD.com, and BYD are training their AI breakthroughs on each other – and on reimagining global supply chains.
Consider Alibaba's two-pronged assault: One arm develops frontier AI models. The other embeds Large Language Models (LLMs) deep within shopping experiences for 1.2 billion active users across Taobao consumer and DingTalk enterprise apps. In 2024, Alibaba’s AI cloud services business soared +200% quarter-over-quarter.
Alibaba has engineered an AI-commerce flywheel, with others now following the same playbook.
The added in-market supply chain experience of Alibaba and its peers makes the AI explosion look decidedly welcome. Yes, security concerns exist, but open-source models like DeepSeek can run on private clouds behind firewalls. Governments and bad actors gain no additional access to data.
Little wonder that LVMH chose to strengthen Alibaba ties in 2025 despite mounting tariff pressures. AI expertise and Chinese market access justify substantial tariff costs. Meanwhile, Apple is weighing replacing OpenAI's ChatGPT with agentic models from Baidu, Tencent, or Zhipu AI to secure Chinese government approval to sell AI-enabled iPhones in China – and because they work.
Now Focus
The AI acceleration shows no signs of slowing. Yet the imperative isn't to chase each innovation like a cat let loose in a barn, as my grandmother would say. The point is to leverage what’s happening in the AI ecosystem and steer it toward enterprise value.
Old-world companies like Domino’s Pizza (+4,600% increase in stock price value, 2010-2023) get this, consistently outperforming valuation growth of tech giants Amazon, Apple, and Alphabet by pointing digitization towards their unique market advantages.
A simple action plan to that end:
- Channel the hype. Funnel the AI news into your strategy forums. Designate technical experts to fact-check and filter. Then regularly challenge your team’s PRFAQ vision for AI’s end-to-end algorithmic supply chain potential.
- Continuously calibrate ROI. Periodically tap Zero100’s AI Blueprint data set to calibrate AI proof of value. Leverage peer networks to augment insight on ROI and scale the potential of emerging AI capabilities. Close the loop on AI experiments to ensure value capture.
- Widen the learning funnel. Leverage internal supply chain-data science fusion teams to test and scale new AI capabilities fast. Tap external technology partnerships to learn from the mistakes of others and gain access to new markets. Keep your radar up for other free untapped learning opportunities in your supply chain ecosystem.
But while agile innovators grab headlines, lack of focus kills 90% of transformations. Amazon founder Jeff Bezos cuts to the heart of it: "I very frequently get the question: 'What's going to change in the next ten years?'... I almost never get the question: 'What's not going to change...?' And I submit to you that that second question is actually the more important of the two."
Whether you're optimizing cold chain deliveries, streamlining automotive parts flow, or managing retail inventory turns, those fundamentals still drive your competitive edge, not the latest algorithm.
Zen and the Art of Chaos Surfing
The best AI advances, like the best supply chain solutions, emerge from the creative tension between ambition and constraint. As my colleague Lauren Acoba notes, AI engineers and supply chain professionals share this DNA – both excel at turning constraints into opportunities, building value despite the daily chaos. That's not just adaptation; it's alchemy.