# AI Winners and Losers After Gemini 3

Source: https://www.youtube.com/watch?v=wlBMcCl2WVk
Recap page: https://rapidrecap.app/video/wlBMcCl2WVk
Generated: 2025-11-20T02:32:52.361+00:00

---
## Quick Overview

Gemini 3 solidifies its position as a top-tier AI model, outperforming GPT-5.1 on benchmarks, while the massive strategic investments by Microsoft and Nvidia into Anthropic ($30B compute capacity, $10B and $5B respectively) create a competitive landscape where Google's internal efforts and partnerships, though strong, face challenges from a newly empowered Anthropic.

**Key Points:**
- Gemini 3 beats GPT-5.1 on benchmarks, but Anthropic's Sonnet 4.5 is still ahead in the few categories where Gemini 3 was mixed (like coding).
- Microsoft committed $5 billion and Nvidia committed $10 billion to Anthropic, with Anthropic also purchasing $30 billion in Azure compute capacity.
- Google is seen as a winner due to its internal strengths and the fact that its models are trained entirely on TPUs, avoiding reliance on Nvidia GPUs.
- Meta's Zuch and Wang are noted for getting their groove back, while Elon Musk (Grok) is seen as just doing his own thing.
- Vibe Coders and Google are listed as winners for providing better designs and being the 'Capstone of a three-year clawback,' respectively.
- The overall sentiment suggests that while Google leads in certain areas, the massive investment in Anthropic signals a highly competitive future where companies like Nvidia and Microsoft are deeply embedded.
- The speaker initially had positive experiences with Gemini 3 but notes that users are starting to experiment with Grok 4.5 and other models, suggesting continued volatility in preference.

![Screenshot at 00:07: The main focus is on assessing the 'WINNERS/LOSERS after Gemini 3 Day,' using a list format to quickly categorize the impact of the new model release and related industry moves on key AI players.](https://ss.rapidrecap.app/screens/wlBMcCl2WVk/00-00-07.png)

**Context:** The video analyzes the immediate winners and losers across the AI landscape following the release of Google's Gemini 3 model and major strategic investment announcements involving AI startup Anthropic. The analysis is structured around a subjective 'Winners/Losers' framework applied to major players like OpenAI, Nvidia, Meta, and Google, using early benchmark results and recent partnership news as context.

## Detailed Analysis

The speaker reviews the competitive landscape following the Gemini 3 launch and massive investments into Anthropic. Gemini 3 performs well, beating GPT-5.1 on benchmarks, but Anthropic's Sonnet 4.5 still leads in some areas like coding, creating a 'surprising mixed' result for Gemini 3. Nvidia is categorized as a loser because their business relies on others using GPUs, while Google is a winner for training Gemini entirely on TPUs, indicating an internal advantage. Meta is seen as regaining groove. The massive investment into Anthropic by Microsoft ($5B) and Nvidia ($10B), coupled with Anthropic's $30B commitment to Azure, positions Anthropic as a major contender that others will need to adapt to. The speaker notes that while Google has strong internal capabilities, the market is not strictly polarized; users are already experimenting with Grok 4.5, and the competition remains fierce. The overall implication is that the AI landscape is becoming highly competitive, demanding constant innovation from all major players.

### Gemini 3 Benchmark Performance

- Gemini 3 beats GPT-5.1 on benchmarks
- Sonnet 4.5 still leads in a few mixed categories like coding
- Anthropic's mixed performance is noted.

### Winners/Losers Categorization

- Nvidia is a loser (sleeping on G3, trained on TPUs)
- Grok is doing its own thing
- Meta is getting its groove back
- Google is a winner due to TPU integration.

### Anthropic's Strategic Wins

- Anthropic secured $30B in Azure compute capacity
- Nvidia invested $10B into Anthropic
- Microsoft invested $5B into Anthropic.

### Vibe Coders & Google as Winners

- Vibe Coders are winners for better designs
- Google is the 'Capstone of a three-year clawback' for its perceived behind-the-scenes work.

### Competitive Landscape

- The recent moves suggest a fierce battle, forcing companies to constantly innovate and reorganize internal processes, as seen by Google's DeepMind realignment.

![Screenshot at 00:07: The main focus is on assessing the 'WINNERS/LOSERS after Gemini 3 Day,' using a list format to quickly categorize the impact of the new model release and related industry moves on key AI players.](https://ss.rapidrecap.app/screens/wlBMcCl2WVk/00-00-07.png)
![Screenshot at 00:30: Nathaniel Whittemore's LinkedIn post summarizing immediate winners and losers after the Gemini 3 announcement.](https://ss.rapidrecap.app/screens/wlBMcCl2WVk/00-00-30.png)
![Screenshot at 01:11: Mike Krieger's tweet announcing the significant partnership between Microsoft, Nvidia, and Anthropic.](https://ss.rapidrecap.app/screens/wlBMcCl2WVk/00-01-11.png)
![Screenshot at 01:30: The Kobelssi Letter tweet detailing the multi-billion dollar partnership investments between Nvidia, Microsoft, and Anthropic.](https://ss.rapidrecap.app/screens/wlBMcCl2WVk/00-01-30.png)
![Screenshot at 03:31: Oriol Vinyals' tweet highlighting the pre-training vs. post-training gains in Gemini 3 performance, noting a drastic jump between 2.5 and 3.0.](https://ss.rapidrecap.app/screens/wlBMcCl2WVk/00-03-31.png)
![Screenshot at 04:15: Bindu Reddy's tweet comparing Gemini 3 benchmarks against GPT-5 and Sonnet 4.5, noting Sonnet 4.5 still leads in some areas.](https://ss.rapidrecap.app/screens/wlBMcCl2WVk/00-04-15.png)
![Screenshot at 05:35: A tweet showing a comparison chart where GPT-5.1 still beats Gemini 3 in certain metrics, contradicting the overall hype.](https://ss.rapidrecap.app/screens/wlBMcCl2WVk/00-05-35.png)
![Screenshot at 07:06: A LinkedIn post discussing the implications of the competitive landscape, specifically mentioning the advantage of Google's internal TPU training.](https://ss.rapidrecap.app/screens/wlBMcCl2WVk/00-07-06.png)
![Screenshot at 08:32: A tweet from Deedy suggesting Google is coming out on top due to vertical integration over the full AI value chain.](https://ss.rapidrecap.app/screens/wlBMcCl2WVk/00-08-32.png)
![Screenshot at 11:28: A tweet showing an animation of Replit integrating Gemini 3 into its design experience, demonstrating practical application.](https://ss.rapidrecap.app/screens/wlBMcCl2WVk/00-11-28.png)