Google parent Alphabet is trying to reset the narrative after a punishing stretch on Wall Street. The company ended its longest monthly losing streak in more than a decade heading into September, and this week it rolled out product updates, pricing moves, and legal relief that together paint a more constructive picture for investors.

Two new models, one aggressive cadence

On Wednesday, Google introduced Gemini 3.8 Flash — its third Flash release in six weeks — alongside a dedicated cybersecurity model built for trusted government and enterprise customers. The company describes 3.8 Flash as Gemini's strongest reasoning and coding variant to date, with meaningful gains over 3.7 Flash on software engineering and multi-step workflows.

The model is tuned for two areas where AI vendors are racing to convert research into revenue: coding and agentic tasks. Tulsee Doshi, senior director of product management at Google DeepMind, told CNBC the recent Flash line has "really surprised us in positive ways in their performance" and that the results "give us opportunities to lean into them."

Smaller Flash systems cost less to operate and iterate than Google's largest frontier models, and they are closing the gap on select benchmarks. Even so, Gil Luria, an analyst at D.A. Davidson, said in an interview that the advances are unlikely to pull Alphabet near Anthropic or OpenAI in the enterprise market.

"From a product perspective this model seems to keep Google in the race, but probably won't change the fact they are a distant third in the enterprise market," Luria said. He recommends holding the stock.

Pricing as a competitive weapon

Google is pairing model improvements with a deliberate cost message. Gemini 3.8 Flash is priced at 75 cents per million input tokens and $3.75 per million output tokens — unchanged from the prior Flash model despite the upgrade in coding, agentic, and reasoning capabilities.

  • Gemini Enterprise is adding pay-as-you-go billing, token discounts of up to 20%, monthly caps on agent spending, and a zero-dollar base subscription tier.
  • Materials shared with CNBC show Google directly comparing its approach to Microsoft and Anthropic, arguing that recurring seat fees and separate product licenses make rival offerings more expensive and less flexible.

Scale remains a tailwind: nearly three-quarters of Google Cloud customers already use its AI products, and Google Cloud CEO Thomas Kurian told CNBC those clients are spending about 50% more than their original commitments.

DeepMind's Demis Hassabis outlined a similar playbook at the G20 Innovation meeting on Wednesday, suggesting Gemini could serve as a general-purpose layer that coordinates cheaper, specialized models and agents — a future where breadth rivals raw model supremacy. It was Hassabis's first public appearance since DeepMind's reorganization last month, when he stepped down as CEO to become chairman of the unit.

Cybersecurity at a lower price point

Google says Gemini 3.8 Flash Cyber can detect and patch software vulnerabilities at frontier-level performance while running substantially faster and more affordably than larger systems.

"We're really excited about being able to provide an offering to defenders that is a fraction of the cost, much faster, while still showcasing that frontier-level performance," Doshi said.

Because those capabilities could be misused, access is initially restricted to a small group of trusted government and enterprise defenders through the new Fairwind Program.

High-profile backing — and a heavy bill to pay

Alphabet is pouring enormous sums into AI infrastructure, so management is counting on growth and share gains to justify the investment over time. Berkshire Hathaway CEO Greg Abel offered a public vote of confidence on Wednesday, telling CNBC's Becky Quick the conglomerate sees Alphabet as a winner in AI — a view shaped in part by how Berkshire portfolio companies use Google's technology.

"We have a lot of visibility from within our companies as to how we're using AI, what type of benefits it's delivering, so that brought incremental interest, and then we saw Google as a significant player," Abel said.

Ad business strength meets antitrust relief

While some market experts view Google's AI position as fragile, its advertising engine continues to deliver. Ad revenue grew 14% in the latest quarter, and the unit received another lift on Wednesday when a federal judge overseeing the Justice Department's ad-tech antitrust case ruled that Google will not have to sell its AdX exchange. The court favored behavioral remedies over the structural breakup regulators had sought — a decision that follows a separate ruling last year rejecting calls to force a Chrome divestiture.

Antitrust attorney Wyatt Fore, a partner at Schinder Cantor Lerner, told CNBC it's a "big deal" that courts in both cases opted against breaking up the company, adding that Google is heading into the AI race with "no hands tied behind its back."

Shares still searching for traction

The stock has not yet reflected much of this week's positive news flow. Alphabet shares rose 0.6% on Wednesday after falling more than 1% on Tuesday to open the month, leaving the name slightly lower as September gets underway. The broader setup nonetheless looks healthier than it did after a difficult summer, when Google lost ground in the AI model race, saw high-profile talent departures, and underwent a major restructuring inside DeepMind.

WATCH: Google's Pixel 11 puts Gemini Intelligence at the center of the smartphone