Perplexity CEO Dares Google to Choose Between Ads and AI Innovation

Key Takeaway:

• Perplexity CEO Aravind Srinivas urges Google to choose between protecting ad revenue or embracing AI-driven browsing innovation.
• As Perplexity’s Comet browser pushes AI-first features, a new browser war looms, challenging Google’s traditional business model.

In a candid Reddit AMA, Perplexity AI CEO Aravind Srinivas criticized Google's reluctance to fully embrace AI agents in web browsing. He believes Google faces a critical choice: either commit to supporting autonomous AI features that reduce ad clicks or maintain its ad-driven model and suffer short-term losses to stay competitive.

Srinivas argues that Google’s deeply entrenched advertising structure and bureaucratic layers are impeding innovation, especially as Comet, a new browser from Perplexity, pushes AI agents that summarize content, automate workflows, and offer improved privacy. He described Google as a “giant bureaucratic organisation” constrained by its need to protect ad revenue.

Comet, currently in beta, integrates AI tools directly within a Chromium-based browser, allowing real-time browsing, summarization, and task automation via its “sidecar” assistant. Srinivas warned that large tech firms will likely imitate Comet’s features, but cautioned that Google must choose between innovation and preservation of its existing monetization model.

Industry experts are watching closely as a new "AI browser war" unfolds. While Google may eventually incorporate ideas from Comet, such as Project Mariner, Srinivas remains confident that Perplexity's nimble approach and user-first subscription model give it a competitive edge.

OpenAI Expands Infrastructure with Google Cloud to Power ChatGPT.

Key Takeaway

• OpenAI has partnered with Google Cloud to boost computing power for ChatGPT amid rising infrastructure demands.
• The move marks a shift to a multi-cloud strategy, reducing dependence on Microsoft Azure and enhancing global scalability.

OpenAI has entered into a major cloud partnership with Google Cloud to meet the rising computational demands of its AI models, including ChatGPT. This move, finalized in May 2025, reflects OpenAI’s ongoing strategy to diversify its cloud infrastructure and avoid overreliance on a single provider.

Historically, OpenAI has leaned heavily on Microsoft Azure, thanks to Microsoft’s multi-billion-dollar investment and deep integration with OpenAI’s services. However, with the explosive growth of generative AI and increasing demands for high-performance GPUs, OpenAI has been aggressively expanding its cloud partnerships. The addition of Google Cloud now places the company in a “multi-cloud” model, also involving Oracle and CoreWeave, which recently secured a $12 billion agreement with OpenAI.

By tapping into Google’s global data center network—spanning the U.S., Europe, and Asia—OpenAI gains greater flexibility to manage the heavy compute workloads needed for training and running its large language models. Google, for its part, strengthens its cloud business by onboarding one of the world’s leading AI developers as a client, which not only enhances its credibility but also diversifies its cloud clientele beyond traditional enterprise workloads.

This deal marks a significant step in the ongoing arms race among tech giants to dominate cloud-based AI infrastructure. OpenAI’s multi-cloud strategy ensures resilience, scalability, and availability for its services across different regions and use cases. It also allows the company to better respond to surges in demand for ChatGPT and its API-based offerings, which serve millions of users and enterprise clients daily.

The partnership underscores a broader shift in the tech industry, where high-performance computing for AI is becoming a core battleground. For OpenAI, spreading its workload across multiple providers could mitigate risks, lower costs, and boost its capacity to innovate and iterate at speed.

Google’s AI Can Now Make Phone Calls on Your Behalf

Key Takeaway

• Google's Gemini AI can now call local businesses for users directly through Search to gather information or book services.
• The feature uses Duplex technology and is available in the U.S., with opt-out options for businesses and premium access for AI Pro users.

Google has taken a major step forward in AI-powered assistance by rolling out a new feature in the U.S. that allows its Gemini AI to make phone calls to local businesses directly through Google Search. This tool, first tested earlier this year, lets users request information like pricing, hours of operation, and service availability without ever picking up the phone.

When someone searches for services such as pet grooming, auto repair, or dry cleaning, they may now see an option labeled “Ask for Me.” If selected, Gemini will use Google’s Duplex voice technology to place a call to the business. The AI introduces itself as calling on the user’s behalf, asks relevant questions, and then returns the response to the user via text or email.

This move transforms the search experience into a more active and intelligent assistant. Users can now delegate simple but time-consuming tasks like making inquiries or scheduling appointments. It’s part of Google’s broader strategy to make AI more agent-like, capable of taking real-world actions on behalf of users.

Credit: Google

Businesses that don’t want to participate in this feature can opt out using their Google Business Profile settings. For users, the functionality is available across the U.S., but those subscribed to Google’s AI Pro and AI Ultra plans benefit from more usage credits and access to advanced Gemini models like Gemini 2.5 Pro. These premium tiers also include features like Deep Search, which can generate in-depth research reports on complex topics using AI reasoning.

As AI integration deepens in everyday apps, this feature showcases a new phase of interaction, where digital tools not only inform but also act on our behalf. Google’s move reflects the future of AI as not just a search engine assistant, but a personal concierge for real-world tasks.