Pure software companies without a network effect or proprietary data, such as Calendly and DocuSign, are 'basically dead' because AI makes their core functionality trivial to replicate.
The cost of AI intelligence is decreasing by 10x per year, a deflationary force so powerful that founders should not be overly concerned with current AI expenses as they will 'solve themselves with time'.
San Francisco is the undisputed primary hub for ambitious startup founders due to its critical mass of talent and capital; London and Paris are the best secondary options.
AI is fundamentally breaking traditional hierarchical company structures, making roles like middle management obsolete and shifting the primary operational constraint from employee headcount to AI token consumption.
U.S. export restrictions on frontier AI models have inadvertently created a significant market for sovereign AI and inference capabilities in other countries seeking to reduce their technological dependence.
Foundational Period
Co-founds GoCardless, which becomes one of fewer than five 'kernel' companies that formed the initial London startup ecosystem (claim 13).
Circa 18 months ago
Y Combinator companies reaching Demo Day generated a baseline of revenue per employee that would later serve as a benchmark for AI-driven productivity gains (claim 31).
Circa one year ago
The common view within the tech community was that AI's primary utility was as a productivity tool for enhancing existing workflows, such as using Copilots to make engineers 20% more productive (claim 22).
Present Day
Articulates the view that AI is a foundational technology breaking old business models, citing a 5x increase in revenue per employee at YC companies and a 10x annual decrease in the cost of intelligence (claims 25, 31, 10).
Next 12 months
Predicts that AI systems will begin to automatically route user queries to the most cost-effective models without user intervention, further abstracting away complexity and reducing costs (claim 2).
▶AI as a Foundational Disruption to BusinessMay 2026
This theme posits that AI is not merely an incremental tool but a revolutionary force fundamentally altering business. It argues that AI is making pure software companies obsolete, breaking traditional hierarchical organizational structures, and enabling the creation of 'self-improving' companies built around AI loops.
For investors, this suggests that traditional moats for software companies, such as code complexity, are eroding. Future value will likely be found in companies with proprietary data, unique business context, and the ability to leverage AI to create recursive improvement cycles.
▶The Deflationary Economics of IntelligenceJul 2026
This theme focuses on the rapidly decreasing cost of AI. The assertion is that the cost for a given level of AI intelligence is falling by approximately 10x per year, making current high costs a temporary issue that will resolve itself. This economic shift is leading to a new paradigm where companies prioritize spending on AI model tokens over employee headcount.
Analysts should model AI-native companies with different long-term margin expectations, as their primary cost of goods sold (intelligence) is deflationary, unlike traditional labor costs. This could justify high initial token burn for companies that can achieve market leadership.
▶The Y Combinator AI PlaybookMay 2026
This theme details Y Combinator's strategic response to the AI revolution. YC is actively guiding its portfolio companies by providing significant AI token credits, advising them to disregard short-term AI costs, and re-evaluating founder archetypes. Internally, YC is also transforming its own operations by using AI to automate code fixes and create a 'self-improving' user manual from recorded office hours.
Y Combinator is not just a passive investor in AI; it is functioning as a centralized R&D lab for building AI-native companies, creating a powerful feedback loop where its operational learnings directly inform its investment thesis and portfolio support.
▶The Enduring Importance of Startup GeographyJul 2026
Despite the rise of remote work, there is a strong conviction that physical location remains critical for ambitious startups. San Francisco is identified as the undisputed global center, offering a critical mass of talent, capital, and ambition. London and Paris are positioned as the best secondary options, highlighting the network effects of established ecosystems.
The emphasis on physical hubs suggests that while AI can automate internal processes, the serendipitous interactions and dense networks found in places like San Francisco are still considered irreplaceable for building high-growth companies.