Google has recently introduced a new feature called “AI mode” as part of its ongoing advancements in artificial intelligence (AI).

This feature is expected to roll out soon for users in the United States, while a global launch date has yet to be announced. Google describes it as similar to having a chat with a knowledgeable expert in various fields.

This release marks just one of many initiatives that are part of Google’s comprehensive strategy to incorporate AI into its services.

Google’s ambitious strategy is not limited to enhancing individual applications. The company aims to control the entire AI development process—a strategy known as “vertical integration”—which includes everything from designing AI chips to developing user interfaces like Google Maps and Gmail.

Other companies are also pursuing similar integration strategies. For instance, OpenAI recently acquired a startup founded by Apple’s Jony Ive to centralize its hardware development. Amazon is doing the same, owning cloud services, custom chips, and gadget designs, while enhancing AI functionalities.

This trend may signal a shift towards vertical integration among major tech firms and could have major implications for both users and businesses.

The AI ‘tech stack’

The “AI tech stack” consists of various layers, including hardware, software, databases, and servers that work together.

Google’s evolving vertical tech stack has four main components:

1. Hardware layer. Google creates its own AI chips, known as tensor processing units (TPUs), which are said to outperform general-purpose processors in efficiency.

2. Infrastructure layer. The company relies on its cloud infrastructure for its computing, networking, and storage needs, which supports its AI capabilities.

3. Model development layer. Google conducts in-house research to innovate its products and services, focusing on robotics, language models, and computer vision.

4. Data layer. Google continually gathers data from users through its various platforms, including search, maps, and email, making data collection a prerequisite for using its services.

Some believe that vertical integration is an effective strategy across multiple industries, not just tech, but practical challenges exist.

Power Dynamics and Imbalances

Google and OpenAI are among a few companies that dominate the tech market.

This dominance allows these firms to set higher prices and potentially engage in questionable practices in online advertising.

Vertical integration can intensify this imbalance by consolidating the various layers of the AI tech stack within individual companies. Distributing responsibilities across sectors could create fairer competition.

This concentration of power presents challenges for smaller firms, hindering their ability to compete as larger corporations manage everything internally.

Add to this the reduced motivation for innovation aimed at improving the customer experience, as the lack of competition typically spurs creativity.

Data is often referred to as the new gold, especially in the realm of AI, which relies heavily on it. Google’s robust access to user data gives it an edge in the industry.

Concerns with Vertical Integration

The effectiveness of a vertically integrated company hinges on retaining top expertise within its walls. Keeping knowledge centralized can lead to hoarding behaviors.

Studies show that knowledge hoarding diminishes social learning and increases disparities in the market, creating vulnerabilities as potential gains are lost for the sake of exclusivity.

This exclusivity can also reduce the organization’s resilience, as critical points of failure become consolidated.

Risk is better managed when there is diversity in oversight, accountability, and transparency. Collaboration among different industry players relies on these principles.

By centralizing the AI tech stack, a company may remove external checks, reducing interactions with outside product or service providers, which can lead to riskier corporate behavior.

Regulatory bodies can act as a form of external oversight.

However, the current move towards deregulating AI is widening the gap between technology development and regulation.

It also results in greater opacity within large tech firms. A lack of transparency raises questions about their practices, particularly concerning AI and data management—as data practices remain a significant concern.

The growth of vertical integration in AI could further exacerbate issues of transparency and accountability.