Ankur Crawford: ‘When Software Begins to Write Software, Innovation Is Exponential’

Episode Summary

Executive Summary: Ankor Crawford argued that recent market volatility mostly reflected fear rather than deteriorating fundamentals, while also acknowledging a real consumer slowdown. She sees AI as an early-stage, economy-wide productivity revolution akin to the steam engine, with major implications for winners, losers, valuations, labor, and portfolio construction. Alger is positioning around concentrated, non-diversified growth portfolios to capture these shifts.

Main Topics: Personal background and path to investing (Priority: 5/5): Crawford described an unconventional upbringing across multiple countries and schools, plus engineering training at Berkeley and Stanford, which shaped her curiosity and analytical style before she moved from Intel and a Merrill Lynch internship into asset management. How she interprets market volatility (Priority: 5/5): She distinguished fear-driven volatility from fundamentals-driven declines, arguing the early-August selloff was largely a fear trade fueled by macro uncertainty, geopolitical noise, jobs data, and AI-bubble concerns. Consumer slowdown and bottom-up work (Priority: 4/5): Crawford said evidence is emerging that the U.S. consumer is weakening, which matters for company-level forecasts and market-size assumptions in sectors like digital advertising and consumer-facing businesses. AI as a structural transformation (Priority: 5/5): She framed gen AI as a democratization of compute that will drive large productivity gains, lower the cost of deploying advanced software, and create a long runway for change across industries and portfolios. Historical parallels and labor disruption (Priority: 4/5): Crawford compared AI’s likely impact to the steam engine and Industrial Revolution, emphasizing that technology can reshape labor, production, and geopolitical power, while forcing retraining and labor-market adaptation. Winners, losers, and valuation resets (Priority: 5/5): She argued AI will create broad second-order winners in infrastructure, utilities, industrials, and select materials, while putting pressure on software margins and forcing investors to rethink valuation frameworks. Portfolio construction, concentration, and Alger’s strategy (Priority: 4/5): Crawford discussed the challenge of benchmark concentration, SEC diversification constraints, and why Alger launched a concentrated, non-diversified strategy to pursue higher conviction alpha in a market dominated by megacaps.

Key Arguments: The recent selloff was largely fear-based, not purely fundamental, so selling into it would have been the wrong response. The U.S. consumer is clearly slowing, and that should feed into bottom-up revenue and market-size assumptions. ChatGPT marked a major step because it democratized compute, making AI economically viable for many more businesses. AI’s productivity gains are likely to outweigh deployment costs, especially over the next two to five years. The steam engine is a useful analogy because a major technology shift can slowly but permanently rewire industry structure, labor, and geopolitical power. Companies that are hard to “AI” because they involve physical assets or scarce expertise may deserve premium valuations. Software businesses may face margin pressure as AI begins to write software and automate coding tasks. Growth vs. value should not be treated as a single binary; some “value” names are cyclical while others are secular decline/value traps. Higher rates hurt duration assets in 2022, but once rates stabilize or fall, growth can regain leadership—though a rapid cut could also signal recession risk. Market and benchmark concentration have made active management harder; concentrated, non-diversified portfolios can better match the current market structure. AI will likely reshape capital spending, data centers, power demand, and industrial supply chains, creating new winners beyond semiconductors.

Data Points: Schools attended in childhood: 8 - Crawford said she had attended eight different schools by age 12 or 13 due to a global upbringing. Countries visited by early teens: 25 - She said she had been to 25 different countries while growing up. PhD completion year: 2004 - She earned her PhD in materials science and engineering from Stanford in 2004. Years she had been thinking about AI/machine learning: since 2015-2016 - Crawford said Alger had been studying AI and machine learning since roughly 2015 and 2016. Data center share of U.S. market: 2% to about 10% by 2030 - She cited Alger research on rising electricity demand from data centers. Cadence product design cycle reduction: 6 years to 3 years - She said Cadence used AI to cut its hardware design cycle in half. Engineer productivity gain: 20% to 30% - Crawford said Cadence engineers are about 20-30% more efficient using AI. Typical chip design engineer fully loaded cost: $300,000 to $500,000 - She used this range to illustrate the dollar value of AI-driven productivity. Concentrated strategy holdings: 20 to 30 holdings - She said the new concentrated equity strategy is designed to hold roughly 20-30 names. Benchmark weight of top three megacaps: over 30% combined - She noted Microsoft, Apple, and NVIDIA together exceed 30% of the benchmark weight. Historical benchmark concentration at start of career: ~3% max per company - She contrasted current concentration with 2004, when no company was above about 3% in the benchmark. Macy's multiple pre-Amazon: 16x to 20x - She used Macy's to show how secular decline can compress valuation multiples over time. Macy's later multiple: 4x to 6x - She cited the post-Amazon decline in Macy's valuation range. Fed last rate hike: July 2023 - She referenced the last rate hike as occurring in mid-2023.

Pivotal Quotes: "When software begins to write software, innovation becomes exponential. It is no longer linear and limited by humans." — Ankor Crawford: Her core thesis on why AI is transformative for productivity, business models, and portfolio valuation. "I fundamentally believe that, you know, first of all, it shook out some weekends, which is, you know, it's a healthy correction, in my opinion." — Ankor Crawford: Her view that the early-August volatility helped wash out weak holders and reflected fear more than fundamentals. "The productivity gains will be significant." — Ankor Crawford: Her response to skepticism about AI’s costs, environmental impact, and real-world usefulness.

Implications: Investors should expect AI to reshape sector leadership, margins, and valuation frameworks, while concentration risk remains elevated in U.S. equities. Bottom-up stock selection still matters, but winners will increasingly depend on who benefits from productivity gains, power demand, and secular change.