Francesco Russo
๐ŸฅŠ ๐—ก๐˜ƒ๐—ถ๐—ฑ๐—ถ๐—ฎ ๐˜ƒ๐˜€. ๐—š๐—ผ๐—ผ๐—ด๐—น๐—ฒ: ๐—ง๐—ต๐—ฒ $๐Ÿญ ๐—ง๐—ฟ๐—ถ๐—น๐—น๐—ถ๐—ผ๐—ป ๐—ช๐—ฎ๐—ฟ ๐—›๐—ฎ๐˜€ ๐—•๐—ฒ๐—ด๐˜‚๐—ป ๐Ÿคฏ โ€‹Everyone is talking about Nvidia. They are the "king" of the AI revolution, and for good reason! ๐Ÿ‘‘ But while everyone looks at them, a massive shift is happening quietly in the background. โ€‹There is a new challenger gaining serious traction: TPUs. โ€‹And the company behind them? It used to be Nvidiaโ€™s biggest customer, but now it might become their fiercest rival: Google. ๐ŸฅŠ โ€‹Here is the simple breakdown of the "Chip War" and why it matters for your portfolio. ๐Ÿ‘‡ โ€‹โž–โž–โž–โž–โž–โž–โž–โž–โž–โž–โž–โž–โž– โ€‹๐Ÿ› ๏ธ ๐—ง๐—ต๐—ฒ ๐—ฆ๐—ถ๐—บ๐—ฝ๐—น๐—ฒ ๐—”๐—ป๐—ฎ๐—น๐—ผ๐—ด๐˜†: ๐—š๐—ฃ๐—จ ๐˜ƒ๐˜€. ๐—ง๐—ฃ๐—จ โ€‹To understand the difference, imagine you need to transport people: โ€‹๐Ÿข GPUs (Nvidia): Think of these as SUVs. They are amazing at everything. You can drive them off-road, on the highway, or in the city. They are flexible and can handle any task you throw at them (gaming, video editing, and AI training). โ€‹๐Ÿš€ TPUs (Google): Think of these as Formula 1 Cars. They are terrible at going off-road or picking up groceries. But if you put them on a race track (AI math), they are unbeatable. They are built for one specific purpose: to make AI calculations extremely fast and cheap. โ€‹๐Ÿ’ก The Takeaway: Nvidia wins on flexibility. Google wins on efficiency. โ€‹โž–โž–โž–โž–โž–โž–โž–โž–โž–โž–โž–โž–โž– โ€‹โš”๏ธ ๐—™๐—ฟ๐—ผ๐—บ ๐—•๐—ฒ๐˜€๐˜ ๐—™๐—ฟ๐—ถ๐—ฒ๐—ป๐—ฑ ๐˜๐—ผ ๐—ช๐—ผ๐—ฟ๐˜€๐˜ ๐—˜๐—ป๐—ฒ๐—บ๐˜† โ€‹For years, Google bought thousands of Nvidia chips. But Google realized something important: โ€‹"Why should we pay Nvidia a premium when we can build our own chips that work better for our own AI?" โ€‹So, they built the TPU (Tensor Processing Unit). โ€‹This is a huge threat to Nvidia for two reasons: โ€‹Google doesn't need to buy as many Nvidia chips anymore. (Lost revenue for Nvidia). โ€‹Google is now renting these chips to others. Companies like Apple, Meta, OpenAI and Anthropic are starting to use Google's TPUs instead of just Nvidia's GPUs to save money. โ€‹โž–โž–โž–โž–โž–โž–โž–โž–โž–โž–โž–โž–โž– โ€‹๐Ÿ”ฎ ๐—ง๐—ต๐—ฒ ๐—™๐˜‚๐˜๐˜‚๐—ฟ๐—ฒ: ๐—ง๐—ฟ๐—ฎ๐—ถ๐—ป๐—ถ๐—ป๐—ด ๐˜ƒ๐˜€. ๐—œ๐—ป๐—ณ๐—ฒ๐—ฟ๐—ฒ๐—ป๐—ฐ๐—ฒ โ€‹This is the most important part for investors! ๐Ÿ’ฐ โ€‹Training (The Learning Phase): When an AI is "learning" (reading the whole internet), it needs flexibility. Nvidia GPUs are still the King here. โ€‹Inference (The Using Phase): When you actually use ChatGPT or Gemini to answer a question, you want it to be cheap and fast. Google TPUs are winning here because they use less electricity and cost less to run. โ€‹Prediction: As AI becomes more common, the demand for "Inference" (running the AI) will explode. This gives Google a massive long-term advantage. ๐Ÿ“ˆ โ€‹โž–โž–โž–โž–โž–โž–โž–โž–โž–โž–โž–โž–โž– โ€‹๐Ÿค‘ ๐—ช๐—ต๐—ฎ๐˜ ๐——๐—ผ๐—ฒ๐˜€ ๐—ง๐—ต๐—ถ๐˜€ ๐— ๐—ฒ๐—ฎ๐—ป ๐—™๐—ผ๐—ฟ ๐—ฌ๐—ผ๐˜‚๐—ฟ ๐—œ๐—ป๐˜ƒ๐—ฒ๐˜€๐˜๐—บ๐—ฒ๐—ป๐˜๐˜€? โ€‹Should you sell Nvidia? No. But you need to watch the landscape change. โ€‹๐Ÿ‚ The Bull Case for Nvidia (NVDA): They are still the standard. Everyone knows how to use them, and they have the best software. They are selling the "shovels" in the gold rush. โ€‹๐Ÿ‚ The Bull Case for Google : They are becoming "vertically integrated." They own the Data, the AI Model (Gemini), and now The Chip. This means they can offer AI services cheaper than anyone else, which leads to higher profit margins. โ€‹Bottom line: The chip war is just getting started. Don't just look at the hardware seller; look at the companies building their own hardware to win the efficiency race! ๐Ÿ โ€‹๐Ÿ‘‡ Tell me in the comments: Do you think Google can actually dethrone Nvidia, or is the "Green Team" too big to fail? $GOOGL (Alphabet Inc Class A) $NVDA (NVIDIA Corporation) $MSFT (Microsoft) $META (Meta Platforms Inc) $BTC
1 reply
null
.