JonasBarrelov
๐˜ฟ๐™š๐™š๐™ฅ๐™Ž๐™š๐™š๐™ โ€™๐™จ ๐˜ผ๐™„ ๐˜ฝ๐™ง๐™š๐™–๐™ ๐™ฉ๐™๐™ง๐™ค๐™ช๐™œ๐™: ๐˜ผ ๐™‚๐™–๐™ข๐™š-๐˜พ๐™๐™–๐™ฃ๐™œ๐™š๐™ง ๐™›๐™ค๐™ง ๐˜ฝ๐™ž๐™œ ๐™๐™š๐™˜๐™? Apple, Amazon, Microsoft, Alphabet (Google), Meta, Nvidia, and Tesla have been spending hundreds of billions to build the most powerful AI models. But last week, a Chinese startup, DeepSeek, shook things up with its R1 AI model, which delivers high performance at a fraction of the cost. ๐™’๐™๐™ฎ ๐˜ฟ๐™š๐™š๐™ฅ๐™Ž๐™š๐™š๐™ โ€™๐™จ ๐˜ผ๐™„ ๐™ˆ๐™–๐™ฉ๐™ฉ๐™š๐™ง๐™จ DeepSeekโ€™s R1 model is open-source, meaning anyone can access, study, and improve it. But the big deal is that Itโ€™s cheaper to train and run than models from OpenAI and Google. If U.S. tech firms can figure out how to replicate its efficiency, AI could become much cheaperโ€”great for some companies, but a nightmare for those who just spent billions on AI data centers. ๐˜ฝ๐™ช๐™ฉ ๐™„๐™จ ๐˜ฟ๐™š๐™š๐™ฅ๐™Ž๐™š๐™š๐™  ๐™๐™ค๐™ค ๐™‚๐™ค๐™ค๐™™ ๐™ฉ๐™ค ๐˜ฝ๐™š ๐™๐™ง๐™ช๐™š? Some experts are skeptical. The Chinese government may be funding DeepSeekโ€™s research and deployment, which could mean itโ€™s not a fair comparison to AI models built by U.S. companies that rely on private investment. However, DeepSeek has released research papers explaining its methods, and AI labs worldwide are now racing to see if they can replicate it. If they can, it could mark the beginning of a new, lower-cost era for AI, forcing companies like Microsoft, Google, and Amazon to rethink their massive AI spending. ๐˜ผ๐™ฅ๐™ฅ๐™ก๐™š: $AAPL (Apple) Apple has spent far less on AI than its rivals, which some saw as a mistake. But if AI models become smaller and more efficient, Appleโ€™s strategy might actually pay off. AI could run directly on iPhones instead of needing the cloud, making Appleโ€™s devices even more powerful. ๐˜ผ๐™ข๐™–๐™ฏ๐™ค๐™ฃ: $AMZN (Amazon.com Inc) Amazonโ€™s biggest moneymaker is AWS (Amazon Web Services), its cloud computing service. But if AI can run on smaller, local systems instead of the cloud, AWS could lose a major source of revenue. Amazon will still use AI for better shopping recommendations and ad targeting, but its high-margin cloud business could take a hit. ๐™‰๐™ซ๐™ž๐™™๐™ž๐™–: $NVDA (NVIDIA Corporation) Nvidia has been the biggest winner in the AI boom, but its stock dropped 17% last week as investors worried that cheaper AI could lower demand for its expensive AI chips. On the other hand, some argue that lower costs could increase AI adoption, keeping demand high for Nvidiaโ€™s hardware. ๐™๐™š๐™จ๐™ก๐™–: $TSLA (Tesla Motors, Inc.) Tesla isnโ€™t focused on building AI modelsโ€”it applies AI to self-driving, robotics, and automation. Cheaper AI could cut Teslaโ€™s costs, making it easier to scale up its technology and gain a competitive edge. ๐™ˆ๐™ž๐™˜๐™ง๐™ค๐™จ๐™ค๐™›๐™ฉ, ๐™‚๐™ค๐™ค๐™œ๐™ก๐™š, ๐™–๐™ฃ๐™™ ๐™ˆ๐™š๐™ฉ๐™–: Microsoft: Plans to spend $ 80B on AI this year, but cheaper AI could make it easier to roll out tools like Copilot, which helps businesses automate tasks in Office apps. Google (Alphabet): Its cloud business grew 35% last quarter, but if AI becomes cheaper, its huge investments in AI infrastructure might not pay off. That said, Googleโ€™s core business (search, YouTube, ads) benefits from cheaper AI. Meta: Spending billions on AI for internal use, but cheaper AI could improve ad targeting and user engagement, making its social media platforms even more profitable. ๐˜ฟ๐™ž๐™™ ๐˜ฝ๐™ž๐™œ ๐™๐™š๐™˜๐™ ๐™…๐™ช๐™จ๐™ฉ ๐™’๐™–๐™จ๐™ฉ๐™š ๐˜ฝ๐™ž๐™ก๐™ก๐™ž๐™ค๐™ฃ๐™จ? Amazon, Microsoft, Google, and Meta have spent $ 212 billion on AI infrastructure in the last year alone. But if DeepSeekโ€™s approach is truly a game-changer, these companies may have overspent on expensive AI data centers that are no longer necessary. For now, the AI race is still on, but DeepSeek has completely changed the conversation. If R1โ€™s efficiency is replicable, Big Tech may need to rethink its entire AI strategy.
Like CommentShare
1 reply
null
.