a16z partner Erik Torenberg has laid out a blunt view of where venture capital is headed: the old boutique model is losing ground, and the firms built to win in the next phase will be scaled platforms that combine capital with recruiting, go-to-market help, legal support, finance, public relations, and government access.
The article, written by Torenberg and later organized and translated by Deep Tide, argues that venture capital is no longer defined only by picking the right startups. Software has become a core part of the US economy, private companies are staying private for longer, and founders now expect investors to help them compete rather than simply write a check and wait.
Winning allocation matters as much as picking the right company
Torenberg says top-tier VCs used to live mainly on judgment. In earlier decades, getting into a promising round was easier, so the key skill was deciding which companies were worth backing. That setup has changed. Founders can get term sheets more easily, top rounds attract intense competition, and access itself has become part of the edge.
His point is simple: if a firm cannot win entry into the best deals, superior selection alone does not solve the problem. He links this shift to a larger number of VC firms, heavier competition for talent and market share across startups, and the fact that companies can remain private long enough for later-stage investors to still earn venture-style returns. He also notes that some of the strongest companies are easier for the market to recognize early, especially when repeat founders launch new ventures.
AI and capital intensity are pushing the model toward scale
Torenberg pushes back on the idea that the number of real winners is fixed. He writes that roughly 15 companies a year once reached $100 million in revenue, while the figure is now about 150. In his argument, not only are there more winners, but the winners are far larger than before. The ceiling for startup outcomes, he says, has moved from $1 billion to $10 billion, and now to $1 trillion or beyond.
He ties that expansion to a broader economic change. Software is no longer a side sector populated by outsiders; it now sits at the center of the US economy. The country’s leading corporate names have shifted from General Electric and ExxonMobil to Google, Amazon, and Nvidia. He also writes that private technology companies are equivalent to 22% of the S&P 500.
That backdrop matters because many of the leading companies in the AI cycle are far more capital intensive than earlier software businesses. Torenberg points to OpenAI, Anthropic, xAI, and Waymo as examples of companies that raised large sums at high valuations from the start. In his view, modern frontier technology companies often need hundreds of millions of dollars because infrastructure, compute, and technical development are expensive, and current funding is flowing more toward R&D and capital expenditure than toward the sales-and-marketing expansion seen in 2021.
Why large firms claim an advantage
He frames the shift as “Firm > Fund.” A fund, in his definition, is optimized to generate carry with minimal headcount and time. A firm has an added objective: build a compounding source of competitive advantage. That means using management fees to create a moat, one that improves deal access and founder support over time.
Torenberg points to Y Combinator as an early example of a scaled venture platform that built real structural advantages rather than relying only on brand. He places a16z in the same conversation. For founders, the pitch is not subtle: a large platform can offer support across hiring, GTM, legal, finance, PR, and policy, while also bringing deeper pools of capital and broader networks.
He also cites a broader market signal. Of the world’s 10 largest companies, 8 are headquartered on the US West Coast and were venture-backed. In his telling, that is evidence that scaled firms are not corrupting venture capital; they are adapting the industry to the scale of the companies it now finances.
A barbell market, not a single winner-take-all structure
Torenberg does not argue that only giant firms will survive. His view is that venture capital is moving toward a barbell structure: a small number of very large platforms on one side, and many focused specialist firms on the other. Both ends can work. The firms in the middle face the hardest position, too large to thrive as niche specialists, but too small to match the capital, services, and access offered by scaled platforms.
His conclusion is direct. Software and AI are changing the shape of companies, and venture capital is changing with them. The older story that treats boutique investing as the pure form of VC and scale as a distortion no longer fits the market he describes.