Two and a half years after declaring artificial intelligence a new epoch for technology, Ben Thompson reflects on how the Big Five tech companies—Apple, Google, Meta, Microsoft, and Amazon—are adjusting their strategies in light of rapid developments and shifting competitive landscapes. Meta´s scramble to upgrade its artificial intelligence talent comes after disappointing Llama 4 results and exposes both vulnerabilities and leadership uncertainties. This new pressure is recasting what was once seen as a sustaining technology for Meta into a source of possible disruption if the gap in foundational model capabilities grows, forcing CEO Mark Zuckerberg to undertake bold—and expensive—recruitment efforts to avoid being left behind.
Apple finds itself in a paradoxical position: while the company lags at the frontier of large language models and related products, its hardware ecosystem is not immediately threatened by artificial intelligence´s ascent. The focus for Apple, Thompson argues, should be to double down on seamless partnerships—most notably with OpenAI—and to expand its hardware leadership into devices optimized for artificial intelligence integration, possibly including significant investments in robotics and next-generation consumer tech. However, Apple´s dedication to a vertically integrated approach could limit its ability to quickly adapt or to acquire breakthrough model talent, especially if regulatory dynamics constrain transatlantic deals.
Google sits on the best technological infrastructure and reaps unique data advantages from its dominance in search, YouTube, and web indexing. The company has rapidly improved its proprietary Gemini models and media generation tools, but faces existential risk from the possibility that artificial intelligence-powered interfaces could erode its core search business model. Google’s response involves fortifying its cloud computing business, leveraging its pricing and model integration edge, and harnessing Android as a launchpad for advanced artificial intelligence experiences. The question remains whether Google can deliver compelling consumer products, not just technical advances.
Microsoft, buoyed by its deep infrastructure and distribution via Azure and Windows, rides its close—but fraying—partnership with OpenAI. While its enterprise and developer offerings benefit from exclusive access to OpenAI’s APIs, tension over business terms and the risk of overreliance on Sam Altman’s ambitions create uncertainty. Microsoft’s best path, Thompson suggests, is to hedge by investing in diverse foundational model providers, including xAI and Mistral, to secure long-term leverage and stave off potential dependency crises.
Amazon, meanwhile, has grown more relevant, shoring up its position with Anthropic and investing early in optionality through projects like Bedrock and Trainium chips. Its infrastructure business, AWS, benefits from increased artificial intelligence-driven usage and enjoys stability absent the existential threats looming over its peers. The e-commerce side also stands to benefit as artificial intelligence becomes more integral to consumer shopping and recommendation systems, so long as competitive threats from alternative marketplaces remain in check.
Thompson also surveys the state of key model makers. OpenAI’s continued consumer traction via ChatGPT puts it fundamentally at odds with platform companies, yet its mass appeal means both Apple and Microsoft have substantial incentives to support integration. Anthropic has captured developer market share and stable enterprise relationships, especially with Amazon, while xAI faces the pitfalls of overcapitalization and limited partners, despite market demand for alternatives to the OpenAI/Anthropic duopoly. The wider industry remains watchful of whether foundation models ultimately prove to be world-destroying disruptors or vital infrastructure with clear business boundaries—the answer will reshape the strategies of every Big Five incumbent.
The analysis closes by highlighting China’s potential role in commoditizing both chips and artificial intelligence, posing another variable that could undercut traditional U.S. tech advantages or reinforce big tech’s dominance depending on global policy and supply dynamics. The strategic landscape of artificial intelligence in 2025 is, then, defined less by certainty than by a mix of calculated bets, shifting alliances, and both new and old business risks shaping the next decade of competition.
