Google launches a "full-stack" counterattack, forcefully reclaiming dominance in AI!

The market has long underestimated Google’s “full-stack” AI counterattack capabilities.

Recently, Huatai Securities stated in its latest report that, unlike OpenAI, which heavily relies on external computing power (Nvidia) and cloud infrastructure (Microsoft), Google is building a completely self-sufficient ecosystem closed loop, from chips (TPU v7p) to models (Gemini 3.0) to applications (Search + Waymo). They are bullish on Google’s self-sufficient “full-stack” AI ecosystem and capabilities, noting that now is the time to regain the leading position.

The report points out that this closed loop is being converted into solid financial returns: TPU deployment has greatly reduced inference costs, the market share of search has stabilized and rebounded to above 90%, and a robust advertising cash flow provides ample ammunition for intensive capital expenditures (Capex).

Huatai Securities believes that Google’s large-scale self-developed cloud TPUs and supporting software ecosystem are driving the growth and market share of its cloud business; the advertising business has monetization elasticity under Gemini’s empowerment, and abundant cash flow feeds back into AI investments and application implementation. In contrast, OpenAI only possesses large model development capabilities. Though current user numbers are high, C-end payment willingness is low, and the B-end commercialization path has yet to be established, so whether the first-mover advantage can continue remains to be seen.

Huatai Securities has raised Google’s target price to $380, more than 18% higher than the current price, corresponding to 30x PE for 2026.

Hard-tech moat: TPU v7p rivals B300, cloud business growth surpasses AWS

Google’s computing foundation is demonstrating astonishing dominance.

The report states that TPU v7p, as Google’s latest iteration of its self-developed chip, has FP8 computing power up to 4.5PF, directly rivaling Nvidia’s B300 chip. Unlike competitors, Google skipped the Scale-up stage in interconnect architecture and directly realizes Scale-out through optical circuit switching (OCS), with a single cluster expandable to 9,216 pod interconnections.

Huatai Securities believes that unlike competitors who rely on external computing resources, Google started TPU deployment as early as 2016 and already possessed training and inference capabilities in 2017. Currently, it is promoting TPU deployment to third-party cloud service providers, such as Fluidstack, and adopting a usage-based revenue model, which is expected to open up new growth prospects.

At the same time, competitor Anthropic has announced it will use 1 million TPUs in 2026, and OpenAI is also starting to use TPUs to reduce inference costs. Google, relying on learning frameworks like TensorFlow, OpenXLA, and the integration of TPU, has built a combined hardware-software AI ecosystem with the potential to compete with Nvidia’s CUDA.

The report points out that this computing advantage not only serves Google’s own needs but also becomes the engine of cloud business growth.

Huatai Securities indicates that in the third quarter Google Cloud revenue reached $15.2 billion, a year-on-year increase of 34%, second only to Azure’s 40% among major cloud service providers, and clearly ahead of Amazon AWS’s 20% and Oracle’s 28%. Market share of the cloud business rose from 18.6% to 19.3% over the same period last year.

Unbreakable cash cow: Search share returns to 90%, strong advertising business feeds back into AI

Previously, the market worried that large models would disrupt the search advertising model, but data shows Google has stabilized its position.

The report notes that Google's latest release, Gemini 3.0, leads in multiple benchmarks, with comprehensive improvements in inference, agent, and interaction capabilities. Currently, Gemini has 650 million monthly active users, still trailing ChatGPT’s 800 million weekly active users, but with deeper integration with the search business, it is likely to capture Google’s entire site user traffic.

The Chrome browser is accelerating the integration of Gemini features, adding multiple agent functions such as task execution and contextual understanding, making the ecosystem more cohesive. The AI Overviews feature has already served over 2 billion users, providing personalized search results and content generation services. Data shows that in September and October, Google’s search market share rebounded to above 90%.

Huatai Securities analysts note that Google’s search and advertising businesses provide stable cash flow and a massive user base for AI development. Under Gemini's empowerment, the advertising business possesses strong monetization elasticity, able to sufficiently fund ongoing capital expenditures.

Diverse AI ecosystem layout

Beyond core search, advertising, and cloud businesses, Google’s AI strategy is increasingly diversified.

Huatai Securities indicated that the Waymo autonomous driving fleet now has over 2,500 vehicles, operating in cities such as Phoenix, San Francisco, and Los Angeles, with weekly orders exceeding 300,000, representing leading global standards in autonomous driving and physical AI.

The report points out that the AlphaFold project has made breakthrough progress in protein structure prediction, reshaping the AI-based drug development field. Currently, five Nobel Laureates are working or have worked for Google, covering cutting-edge fields such as neural networks, protein structure prediction, and quantum computing.

Huatai Securities mentions that, based on full ecosystem construction, they have revised upward Google’s revenue forecasts for 2025-2027. It is expected that revenue in 2025 will reach $405.169 billion (about $405.2 billion), with net profit forecasts revised up to $131.513 billion.

The report states that Huatai Securities maintains a “Buy” rating, with a target price raised to $380, corresponding to a 30x P/E ratio for the 2026 forecasted earnings per share (EPS).

Risk Disclaimer and Limitation ClausesThe market has risks, and investment must be cautious. This article does not constitute personal investment advice, nor does it take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions contained herein fit their specific situations. Investments made accordingly are at the user’s own risk.