AI Has Changed the Software Market: Key Takeaways from the ICONIQ Report

⚡ The gist in 5 seconds - The gist: ICONIQ (a fund with over $80 billion under management) released its September 2026 report on the performance of software companies; Sasha Zhuravlev of Mento VC translated it into Russian and added commentary. - Where to find it: the translation with charts is published on Habr; the original was released by ICONIQ. - Limitation: the sample includes 137 companies — mostly ICONIQ portfolio companies, 11 public ones, and three names are undisclosed. ### 🔍 What was found Over the past year, the S&P 500 software index fell 3% — the worst result among sectors — while chipmakers rose 47%: developing AI requires more computing power. Infrastructure and security are holding up better than others, as AI adoption supports demand for servers, data systems, and protection. Fast-growing public companies (20%+ annual growth) are valued at 9.2x annual revenue versus 6.9x in Q1, while all the rest are valued at just 2.7x. Valuation is still most strongly tied to the Rule of X, where growth is roughly 2.7 times more important than profitability. For the first time since 2024, the link between valuation and FCF margin has strengthened, and in the Rule of 40 the median company gets about 60% from margin and 40% from growth. The share of US companies paying for AI products grew from 7.5% in January 2023 to over 50% by July 2026. The authors divided companies into growth leaders (Pacesetters) and the rest. Leaders must meet two mandatory conditions: fast growth (top 25% by ARR or revenue over three years among companies of the same size) and a strong AI connection of the product — either the product is built on AI models from the start, or the company has embedded AI into its core products and ships separate AI solutions. Specific figures: DigitalOcean's annual AI revenue reached $234 million (+212% year over year), ServiceNow's annual value of AI contracts exceeded $1 billion, and the number of AI agent deployments grew 9x in nine months. Snowflake increased the number of customers using AI from 4,000 to 13,600; Shopify writes more than half of its code with AI without growing its developer headcount; Snowflake avoided hiring about 90 engineers; Samsara's revenue per employee grew more than 30% over three years; CrowdStrike analyzes trillions of events daily. ### 💡 Why it matters The report shows how early-stage investors now evaluate the software market: growth is no longer the only criterion, and business efficiency and free cash flow are becoming increasingly important. At the same time, AI is already generating measurable revenue and enabling more output with the same team, while accumulated data becomes a hard-to-copy advantage — for example, Shopify has been collecting commerce data for 20 years, Samsara has 25 trillion records, and CrowdStrike analyzes trillions of events daily. According to Zhuravlev, for those investing in software or building a company in this industry, the report is one of the most useful materials on the market today. ### 🧩 Context Sasha Zhuravlev is a venture investor and founder of the Mento VC fund, which invests in Seed and Series A stage startups in the US. ICONIQ seem

🔗 Read on habr.com