Google and Amazon are both making chips. How about NVIDIA's future prospects? Let's see how Awen answers.
Ah Wen understands your concerns. Seeing tech giants like Google and Amazon start making their own chips (TPU, Trainium) does make people wonder if Nvidia is about to be "surrounded."
But based on the latest Q3 2025 earnings report and market trends, Ah Wen's conclusion might surprise you: these "in-house R&D" efforts haven't scared investors—instead, they've become a "booster" for Nvidia's stock price.
The current situation is: Nvidia remains the king of "selling shovels," while Google and Amazon are more like "cooking for themselves," incidentally proving how hot the market is for Nvidia.
Let's break it down layer by layer:
🛡️ 1. Can competitors' "in-house chips" really replace Nvidia?
Google (TPU) and Amazon (Trainium) are actually focusing on vertical integration in "specific domains," not trying to steal Nvidia's general-purpose market.
Google TPU (Ironwood):
- Status: Google did train Gemini 3 with its own TPU, which is impressive.
- Limitations: TPUs are still mostly "in-house." Google doesn't sell TPUs to thousands of enterprises, research institutions, and cloud providers like Nvidia does. Nvidia sells "general-purpose computing power," while Google sells "AI services."
- Nvidia's counter: CEO Jensen Huang confidently stated that Nvidia is the only platform that runs across all clouds and supports all models. No matter how strong Google's TPU is, if you can't use it, you still have to buy Nvidia's GPUs.
Amazon (Trainium 3):
- Status: Amazon did release Trainium 3 to cut costs.
- Truth: Amazon remains one of Nvidia's biggest customers! AWS offers vast amounts of Nvidia GPUs for SMEs. Amazon's in-house chips are for reducing costs (for its own profits), not to eliminate Nvidia's market. In fact, AWS is still the largest channel for Nvidia's computing power distribution.
🚀 2. How wide is Nvidia's "moat"?
Ah Wen reviewed the latest December 2025 analysis report—Nvidia's moat is no longer just "chips" but the entire ecosystem:
- Dominance of Blackwell architecture: Nvidia's latest Blackwell (GB200/GB300) chips are in absolute short supply. Top AI companies like OpenAI, Meta, and Microsoft have hundreds of billions in orders queued up.
- Software ecosystem (CUDA): This is the scariest part. AI engineers worldwide use Nvidia's CUDA platform to write code. Google may have TPUs, but its developer ecosystem is far less mature than Nvidia's. Switching chips is easy; switching an entire software ecosystem is nearly impossible.
- High order visibility: Nvidia explicitly stated that its current order visibility has reached $500 billion, meaning its performance for the next 2-3 years is already "locked in" by major clients.
📈 3. Future outlook: Still bullish?
Ah Wen's answer: Still bullish, but the logic has changed.
- Short-term (2026): Blackwell volume ramp. As production scales, Blackwell chips will ship en masse, driving explosive revenue growth. Wall Street widely predicts Nvidia's 2026 revenue will continue rapid growth.
- Long-term (2027+): AI inference market. Everyone is racing to "train" big models now, but Jensen Huang says future "inference" (running models to serve users) demand will be 100x training. Nvidia's upcoming Rubin architecture (mass production in H2 2026) is specifically designed for inference.
📊 Data speaks: Latest market position
Per Bank of America's latest analysis, despite competition, Nvidia's dominance remains solid:
Year Predicted AI Chip Market Share Trend
Now (2025) 85% Absolute monopoly
2030 75% Still leading, slight dip but majority share
📌 Summary
While Google and Amazon are building chips, they're more like "building their own walls," while Nvidia is "selling bricks to everyone."
As long as the AI fire burns hot and countless SMEs need to rent computing power to run models, Nvidia's "shovels" will never go unsold.
Ah Wen's advice: If you're bullish on AI's long-term growth, Nvidia remains the highest-conviction play. Short-term volatility (e.g., "bubble talk") may occur, but its fundamentals are rock-solid.
$NVIDIA(NVDA.US) $Alphabet(GOOGL.US) $Amazon(AMZN.US)
The copyright of this article belongs to the original author/organization.
The views expressed herein are solely those of the author and do not reflect the stance of the platform. The content is intended for investment reference purposes only and shall not be considered as investment advice. Please contact us if you have any questions or suggestions regarding the content services provided by the platform.
