After Wall Street witnessed the end of its most prolonged monthly decline in over ten years, Alphabet (GOOGL) is entering September with several promising signals for investors. The company had previously closed four consecutive months in the red, marking its longest losing streak since 2015. Now, a flurry of developments is reshaping the narrative for the tech giant.

On Wednesday, Google officially launched Gemini 3.8 Flash, the third model in its Flash series within just six weeks, which serves as tangible proof that its AI services continue to iterate and improve. The company also introduced a new specialized cybersecurity model aimed at trusted government and enterprise clients. Adding to this wave of positive news, Berkshire Hathaway (BRK.B) CEO Greg Abel publicly voiced confidence in Alphabet's standing in the AI sector. Furthermore, a federal judge rejected the U.S. Department of Justice's request on Wednesday to force Google to divest its ad exchange platform AdX, marking another significant victory for the company in its ongoing antitrust battles.

These converging events have handed Alphabet a turning point after a difficult summer. During that period, Google's momentum in the large-model race appeared to wane, several key talents departed, and DeepMind underwent a major internal organizational restructuring. However, the stock has yet to show a substantial rebound. The early days of September have been tepid, with the stock dipping over 1% on Tuesday and rising a modest 0.6% on Wednesday, leaving the month slightly in the red overall. Still, Google has high hopes for Gemini 3.8 Flash, which is specifically designed for coding and agentic tasks — two areas where the AI industry is actively seeking to convert technological capability into corporate revenue.

Google touts this model as its most powerful version of Gemini yet for reasoning and coding, boasting substantial improvements in software engineering and multi-step task handling compared to 3.7 Flash. Tulsee Doshi, Senior Director of Product at Google DeepMind, stated in an interview that the recent spate of Flash model releases "have performed beyond our expectations, opening up a lot of new opportunities." These lightweight Flash models run at a lower cost and iterate faster than the company's top-tier flagship models, and their performance on certain tasks is already approaching that of their larger counterparts. D.A. Davidson analyst Gil Luria acknowledged this progress but noted in an interview that it may still not be enough for Google to close the gap with Anthropic and OpenAI. "On the product front, this model keeps Google in the game, but it's hard to see it changing the status quo of being third in the enterprise market with a significant gap to the leaders," he said, maintaining a Hold rating on the stock.

Pricing has emerged as Google's core competitive weapon. Gemini 3.8 Flash is priced at $0.75 per million input tokens and $3.75 per million output tokens, continuing the introductory promotional pricing from the previous generation despite the upgraded coding, agentic, and reasoning capabilities. The enterprise version of Gemini has also added a pay-as-you-go model, offering up to 20% discounts on tokens, setting monthly spending caps for agents, and introducing a no-monthly-fee subscription option. Google is directly targeting Microsoft (MSFT) and Anthropic, arguing that competitors' fixed per-seat annual fees and separate product licenses make their offerings more expensive and less flexible. The company also leverages its scale: nearly three-quarters of Google Cloud customers are already using Google's AI products, and Google Cloud CEO Thomas Kurian noted that these customers are spending roughly 50% more than their original contracted commitments.

Demis Hassabis of DeepMind has also articulated a future vision aligned with this strategy. Speaking at the G20 Innovation Conference on Wednesday, he suggested that future iterations of Gemini will increasingly act as a general orchestration layer, coordinating multiple low-cost specialized models and AI agents. In this paradigm, ecosystem breadth may prove just as important as single-model performance. This was Hassabis's first public appearance since DeepMind's restructuring last month, during which he transitioned from CEO to a new role as the division's chairman. Google is also emphasizing cost advantages in the cybersecurity domain, claiming that Gemini 3.8 Flash Cyber delivers top-tier vulnerability detection and patching results but runs faster and at a fraction of the cost of larger flagship models. Doshi remarked, "We're excited to offer this solution to security defenders at a tiny fraction of the cost of alternatives, with higher speed, while maintaining state-of-the-art performance." Due to the potential for misuse, Google is initially restricting access to a select group of trusted government and enterprise security defense organizations through its newly established Fairwind program.

For investors, there's a clear trade-off: Alphabet is pouring massive capital into AI infrastructure and must rely on business growth and market share gains to generate returns over the long term. Berkshire's Abel told Becky Quick on Wednesday that he sees Alphabet as a potential winner in the AI race, basing his view partly on how Berkshire's own portfolio companies use Google's technology. "A lot of our companies are using AI, and we can see the value it's delivering firsthand. That's why we've increased our attention on the AI sector, and we consider Google a major player," he said. Despite some market skepticism about Google's position in AI, its advertising business remains robust, with ad revenue growing 14% year-over-year in the latest quarter. This cash engine received another boost this week when the federal judge presiding over the DOJ's ad tech antitrust case ruled that Google does not need to divest its AdX ad exchange, opting for behavioral remedies instead of the structural breakup regulators had sought. This follows a similar ruling last year in another antitrust case where the judge also rejected a motion to force Google to divest its Chrome browser. Antitrust attorney Wyatt Fore, a partner at Shinder Cantor Lerner, emphasized the significance of these decisions: "In both cases, the courts have declined to force a breakup of Google, and that is hugely important. Google can now fully immerse itself in the AI race without the overhang of a potential divestiture."