Hajra Naz
Alphabet shares are soaring. Investors are excited about the company’s AI progress. Google has started shipping advanced AI products. With Gemini 3 and Ironwood, CEO Sundar Pichai seems to have assembled the pieces for the company’s AI offerings, experts told CNBC.
Competition remains fierce. Other AI companies are closing in fast. When ChatGPT launched in 2022, Google appeared flat-footed. Now, Gemini 3 and the Ironwood AI chip are generating buzz. Analysts call it Alphabet’s AI comeback.
The AI Comeback Begins
Google kicked off November with Ironwood, its seventh-generation tensor processing unit, or TPU. The company claims it allows customers to run and scale the largest, most data-intensive AI models in existence. Last week, Google unveiled Gemini 3. The model reportedly needs less prompting and delivers smarter, more accurate answers than its predecessors.
The excitement is real. Salesforce CEO Marc Benioff posted on X about Gemini 3. He said that after using ChatGPT daily for three years, he switched to Gemini 3. “The leap is insane,” Benioff wrote. “Everything is sharper and faster. It feels like the world just changed, again.”
While many tech stocks started the week down, Alphabet shares surged. They rose over 5% on Monday, adding to last week’s 8% gain. Warren Buffett’s Berkshire Hathaway disclosed a $4.3 billion stake in Alphabet at the end of Q3.
Alphabet shares are up nearly 70% this year. They outperformed Meta by more than 50 percentage points. Last week, Alphabet’s market cap surpassed Microsoft’s. All this came even as Nvidia reported strong Q3 revenue and guidance.
Melius Research analyst Ben Reitzes explained the paradox. Many AI stocks sold off after Nvidia’s earnings. He said the cause is Alphabet’s AI comeback.
Putting the Pieces Together
Experts say Gemini 3 and Ironwood show Google has finally assembled its AI strategy. Michael Nathanson, co-founder of Moffett Nathanson, highlighted the difference. Google now serves a broad spectrum of customers, from consumers to enterprise clients. This was a challenge after ChatGPT’s launch.
Three years ago, Google seemed lost. Critics doubted Sundar Pichai. Some even called him a failure. Today, experts say Google has a significant advantage.
Early AI launches were imperfect. In 2024, Google pulled its Imagen 2 image generation product after users noticed historical inaccuracies. AI Overviews initially offered faulty advice before adding guardrails corrected it. Gil Luria of DA Davidson noted Google had the tech but needed to assemble it correctly.
Gemini 3 launched quickly after Gemini 2.5 in the spring, which was already impressive. Google’s Nano Banana image generation features also attracted attention. Gemini topped the Apple App Store in September, surpassing ChatGPT. Last week, Google released Nano Banana Pro.
Google leverages YouTube for AI training. Its vast video content offers a competitive edge in image and video generation. Mike Gualtieri of Forrester Research called it “a huge competitive advantage.” He doubts OpenAI or Anthropic can match that data scale.
Google also integrates AI into enterprise products. Its cloud unit drives strong sales. Q3 earnings reported the first $100 billion quarter, fueled by cloud growth. The cloud unit holds a $155 billion backlog and powers AI services.
The AI Chip Advantage
Ironwood is nearly 30 times more power-efficient than Google’s first TPU in 2018. Google’s ASIC chips are a secret weapon in the AI wars. Deals with customers like Anthropic are worth billions.
Meta may also use Google TPUs in its data centers. Nvidia’s stock dropped 3% after the news. The rise of TPUs threatens Nvidia’s dominance in AI chips. Luria explained the advantage of owning the full stack. Optimizing AI models specifically for TPUs boosts performance.
Google’s AI enterprise solutions, combined with TPUs and Gemini 3 integration in consumer products, drive Wall Street optimism. The company’s AI ecosystem strengthens both enterprise and consumer offerings.
Tight Competition
Google’s lead is real but fragile. Competitors remain aggressive. Anthropic launched Opus 4.5, and OpenAI updated GPT-5 for efficiency and conversational improvements. Forrester’s Gualtieri said frontier AI models remain neck and neck.
The AI race is expensive. Alphabet, Meta, Microsoft and Amazon expect capital expenditures to exceed $380 billion this year. Experts predict that frontier AI models may become commoditized. Multiple companies may survive but compete on performance and efficiency.
Google must double its serving capacity every six months to meet AI demand, according to company executives. Amin Vahdat, Google Cloud VP, emphasized AI infrastructure as the most critical and costly part of the race.
Nvidia still holds over 90% of the AI chip market. Google’s TPUs gain attention but remain ASIC-based, designed for specific tasks. Nvidia’s Blackwell chips are more flexible and powerful.
Consumer adoption also lags. Google’s Gemini app has 650 million monthly active users. AI Overviews has 2 billion monthly users. ChatGPT reaches 700 million weekly users. Experts note hallucinations and lower engagement as challenges for Google.
“Yes, Google has got its act together,” Luria said. “But that doesn’t mean they’ve won.”
