Next phase of consumer AI and the future of batteries

Insights from The Economist, 23rd May 2026 edition

Photo by Roberto Sorin on Unsplash

Next phase of consumer AI

We are moving from the era of Generative AI to the era of Agentic AI.

Generative AI helped us create content, summarize information, answer questions, and write code. Agentic AI extends those capabilities by taking action on our behalf. Instead of simply providing recommendations, AI agents can research options, make decisions, coordinate tasks, and increasingly execute workflows across digital services.

In the west, Google is at the forefront. Google recently unveiled a suite of Gemini-powered agents that can work across its ecosystem of products and services. These agents can help plan trips, monitor information such as stock prices or sports scores, conduct research, and can continue working on tasks in the background, moving AI from a conversational tool toward an always-on virtual assistant.

Chinese technology giants are also integrating AI agents directly into the digital ecosystems that consumers already use every day. Companies such as Alibaba, Tencent, and ByteDance operate massive “super apps” that combine messaging, payments, shopping, entertainment, and logistics into a single platform.

As a result, Chinese consumers increasingly rely on agents for purchasing goods and services. The AI can help research options, recommend products, initiate transactions, and coordinate delivery.

In China, agents can buy stuff for you. Entire lifecycle of a purchase from the research, decision, payment and follow-ups could be done by the agents. They do this by working on top of integrated platforms called “super apps” that are run by China’s biggest AI companies like Tencent, Byte dance and others like them.

As agents become more powerful, a critical question emerges: will companies allow competing agents to access their platforms and services? For example, how willing would Google be to allow ChatGPT or Claude deep access to Gmail, Maps, Android, or other Google services? Similarly, would Apple allow third-party agents to operate seamlessly across its ecosystem?

China provides an early glimpse of this challenge. ByteDance partnered with ZTE to launch an AI-powered smartphone designed around an AI assistant. However, the device failed, as Alibaba and Tencent blocked access to their payment platforms.

Regulators will need to address these issues to ensure fair competition, much as they did during earlier platform battles such as Microsoft’s bundling of Internet Explorer with Windows in the late 1990s.

The second challenge is economic. Most consumer AI services remain either free or heavily subsidized. Yet running large-scale AI systems requires enormous computing infrastructure. Google processes more than 3 quintillion tokens each month. Each quintillion is a trillion trillion. Every token consumed requires computing power, energy, and specialized hardware.

Long term sustainability of Agentic AI will depend on lowering the compute costs through more efficient models and hardware, cap on free services to force more users to a subscription model and increasing the volume of advertising within the model responses.

It will be interesting to see if the winner takes all, leaving other companies in the dust again.

You can read the full articles here and here.

Batteries are dead. Long live batteries

Lithium-ion batteries power everything from phones to electric vehicles. However, scientists believe they are approaching their practical limits. They can catch fire because they use flammable liquid electrolytes, lose performance in cold weather, and may not be able to store much more energy than they already do.

One of the promising alternatives is the solid-state battery. Instead of using a liquid electrolyte, it uses a solid material to move ions between the battery’s two electrodes. This makes the battery safer and allows more energy to be packed into the same space. Some designs store up to 65% more energy than a lithium-ion battery of the same weight.

Before we start welcoming solid state batteries as the next generation, there are a few challenges to be solved.

Cracking and brittleness — As a battery charges and discharges, its materials expand and contract slightly. In solid-state batteries this creates tiny cracks that reduce performance. Researchers in China have developed a layered material combining ceramic and polymer sheets that reduces this risk.

Dendrite formation — Dendrites are tiny needle-like crystals that can grow inside a battery and eventually cause short circuits. It was long believed to be a result of lithium build-up on the electrod’s surface. Researchers at MIT now believe the deeper cause may be chemical changes that weaken the electrode over time. This has shifted attention toward developing more chemically stable electrode materials.

Slow charging — In some solid-state designs, ions move more slowly through the solid electrolyte than they do through liquids. Researchers at Oak Ridge National Laboratory found that adding molecules called zwitterions, which contain both positive and negative charge regions, dramatically increased ion movement in laboratory tests.

Manufacturing challenges — Today’s solid-state battery manufacturing relies on solvents and energy-intensive drying processes. These methods can create microscopic pores that increase the chance of defects. Companies such as Tesla and LG Energy Solution are investing in dry-electrode manufacturing, which could lower costs, reduce energy use, and improve battery performance.

Many of the world’s largest battery makers, including CATL, Samsung, and Toyota, are targeting commercial solid-state batteries around 2027. If these efforts succeed, entire industries from mobile phones & cars to home energy storage systems could be transformed.

You can read the full article here.