iPhones Defined the Last Era, Proactive AI Agents Will Define the Next

The past week delivered a flood of AI news that matters to every portfolio tilted toward innovation. Just two years ago, I remarked to Bloomberg that AI is Google’s market to lose. Last week’s Google I/O event was a powerful demonstration of the innovation and execution I envisioned then. Competitors make significant strides. However, Google’s deep integration of AI across its vast ecosystem — from search to Android to cloud — showcases a formidable strategy. Additionally, we saw major updates come from the Microsoft Build event. 

 Here’s a rundown of the key developments last week:

• Alphabet’s I/O unveiled a strategic vision for a proactive, personalized AI. It  showcased Gemini 2.5, the real-time AI assistant Project Astra, the advanced Veo 3 video model, and a $249/mo Google AI Ultra plan to accelerate monetization.
• Microsoft Build championed a vision for an open “agentic web,” pushing the Model Context Protocol (MCP) and new AI Foundry tools for developers.
• Anthropic released Claude 4.0, which now leads benchmark tests for complex reasoning and coding.
• OpenAI made waves by acquiring Jony Ive’s hardware startup. That signals serious intent to design the AI-native devices that will define the post-iPhone era.

The most significant thread connecting these announcements is the shift from AI you command to AI that acts. We are entering the era of the AI agent. Such an agent can understand complex goals and execute multi-step tasks across different applications on your behalf. Think of an assistant that can do more than just find flight options. It can also check your calendar for conflicts, book the best flight based on your known preferences, add it to your itinerary, and notify your contacts, all from a single request.

So, what can these emerging agents do?

• Today: They can draft, translate, query, summarize, and now execute specific instructions inside supported apps. With Google’s Project Astra integrated into Gemini Live on mobile, you can point your camera at something and ask complex questions about it in real-time.
• Coming soon: Google’s Project Mariner aims to let agents juggle up to 10 tasks at once and automatically learn routines you show it. This is where the true innovation lies: in proactive, autonomous assistance.

To make this possible, a new technical standard is emerging: the Model Context Protocol (MCP). Nestled within the developer-focused announcements from Microsoft Build, MCP is the crucial key to unlocking agent capabilities.

In plain English, MCP is like the new HTTP for AI. It lets any data source (an app, a database, a website) tell an AI agent exactly what tools and context to use. It then allows the data source to receive structured actions back. As a common language, MCP will allow AI assistants to securely and reliably interact with both personal apps and enterprise systems. 

The road to a fully agentic future is still being paved. Today’s agents can’t reason over proprietary data. They can’t see or act with perfect reliability in unstructured digital environments. However, the trajectory is clear and the investment implications are profound, extending far beyond the hyperscale cloud providers.

The massive capital spending from Alphabet, Microsoft, and Amazon is the tip of the iceberg of the AI ecosystem. Their combined AI capex targets are north of $255 billion for FY 2025. This capital flows through the entire AI ecosystem. It creates opportunities across multiple subsectors that are essential for bringing AI agents to life:

• Cloud infrastructure & innovative datacenters: This capex directly funds the build-out of next-generation datacenters. The intense heat generated by AI chips is driving a revolution in innovative datacenter technologies. Those innovations include direct-to-chip and immersion liquid cooling, creating a new class of leaders in thermal management.
• Edge compute & connectivity: For an AI assistant like Project Astra to respond in real-time, computation must happen instantly. This requires moving processing away from centralized clouds and closer to the user. The requirement fuels demand for edge computing hardware and the advanced connectivity solutions (like 5G and next-gen Wi-Fi) that provide the necessary low-latency bandwidth.
• Cybersecurity: Autonomous AI agents will access and act upon sensitive personal and corporate data. In doing so, they will create new and complex attack surfaces. This elevates the need for robust, AI-powered cybersecurity frameworks to protect data, manage permissions, and defend against new threat vectors. This also includes big data/analytics tech to make sure everything is functioning properly.

This interconnected system is the engine of AI’s continued growth. For investors, navigating this complex landscape requires a strategy that captures value at every layer of the technology stack. The ROBO Global Artificial Intelligence Index (THNQ) is designed to do precisely that. The index tracks the leaders among the AI model creators and cloud giants. It also tracks leaders of the mission-critical subsectors providing the enabling hardware, advanced networking, and robust security that form the very backbone of the coming agentic AI era.

Additionally, our next webinar, “The dynamic trends shaping robotics, AI, and healthcare,” will be Thursday, June 5, 2025, at 11 a.m. ET. Register here.

THNQ is the underlying index for the ROBO Global Artificial Intelligence ETF (THNQ C+) and the L&G Artificial Intelligence UCITS ETF (AIAI.LN).

For more news, information, and analysis, visit the Artificial Intelligence Channel.

VettaFi is the index provider for THNQ ETF and AIAI.LN, for which it receives an index licensing fee. However, THNQ ETF and AIAI.LN are not issued, sponsored, endorsed, or sold by VettaFi. VettaFi and its affiliates have no obligation or liability in connection with the issuance, administration, marketing, or trading of THNQ ETF and AIAI.LN.