What are the Laffont brothers (Coatue) saying in 2026?

In 2026, Thomas and Philippe Laffont of Coatue Management articulate a vision of an artificial intelligence "supercycle" that is reshaping markets and the macroeconomy . They project the AI ecosystem's revenue will grow from $140 billion to $300 billion in the current year, and then double to **$600 billion in 2027** [30, 27]. This growth is expected to create a new cohort of mega-cap technology companies, with candidates like SpaceX, OpenAI, Anthropic, and Databricks poised to join the public markets within the next 12 to 24 months [24, 29]. Thomas Laffont asserts that the theory of large language models being mere commodities has been "thoroughly disproven," suggesting durable value creation at the model layer . This technological wave is forecast to expand technology's share of global GDP beyond its current 15% and generate alpha for investors who look beyond the MAG7 to find outperforming companies in semiconductors, power, and software [6, 11]. By 2028, Coatue projects that the combined revenue of leading AI companies could even surpass the total revenue of Microsoft .

The competition for AI dominance is most intense at the infrastructure layer, where access to compute is a primary bottleneck . The Laffonts view NVIDIA GPU allocation as a critical leading indicator of future market leadership among cloud providers . In this "arms race," Microsoft has been particularly aggressive in securing its supply chain and is operating at a massive scale, with its AI operations estimated to be producing **100 trillion tokens per month** [1, 8, 16]. In contrast, Thomas Laffont notes that Amazon's share of GPUs is half its share of the AWS cloud market, which could indicate either that AWS is behind in AI or that it is pursuing a divergent hardware strategy . This focus on compute allocation provides a framework for evaluating the long-term strategic positioning of the major hyperscalers in the AI era [12, 16].

Regarding capital markets, the Laffonts highlight a challenging environment for recent public companies while anticipating a new wave of AI-driven IPOs. The 2021 cohort of traditionally listed companies has performed poorly, remaining **down 50% five years later** [5, 19]. This contrasts with the hyper-efficient growth of the MAG7, which have compounded revenue at over 20% in recent years while their operating expenses grew at only 2% [20, 22]. For the current generation of private companies, Coatue advises a strategic approach based on growth and cash burn, with options ranging from IPO preparation for market leaders to radical "reinvention" for less efficient businesses [6, 14]. Despite the productivity gains from AI, Thomas Laffont notes that none of Coatue's portfolio companies are planning to cut their engineering staff in half, suggesting a focus on reinvesting gains into new product development rather than pure cost reduction [9, 18].

From a macroeconomic perspective, Philippe Laffont presents AI-driven productivity as a powerful counterforce to concerns over the U.S. national debt . He posits that if annual productivity growth can reach **2.5 to 3.5%** over the next decade, the U.S. could substantially reduce its debt-to-GDP ratio, effectively growing its way out of a potential debt spiral . This optimistic scenario may already be getting priced in by the bond market . Looking further ahead, Laffont predicts a fusion of technology and government finance, envisioning a future where the U.S. government issues tokenized bonds directly to consumers, creating new instruments like one-year or 30-year stablecoins [4, 13]. The recent passage of stablecoin legislation is seen as a major step in building the regulatory framework for such financial innovations .