Why Artificial Intelligence (AI) Won't Destroy Software Companies, According to This Microsoft Executive

Microsoft executive Rajesh Jha argued that AI agents could increase rather than reduce software demand, since embodied agents may create additional 'seat opportunities' for per-seat licensed software. The article also notes that software stocks have fallen sharply, with the iShares Expanded Tech-Software Sector ETF down nearly 20% and Microsoft trading at 26x trailing earnings versus nearly 40x last year. Overall tone is constructive on software valuations and Microsoft’s pullback, though the piece is largely commentary rather than new company-specific financial results.

Analysis

The market is still pricing software as if AI is a pure substitution shock, but the more important second-order effect is account expansion: agentic workflows create more authenticated identities, more audit trails, and more concurrency than a single human user. That favors vendors with seat-based monetization, identity controls, and usage-linked upsell paths, while punishing point solutions where AI can compress the number of human operators without increasing system touchpoints.

The key nuance is that this is a timing trade, not a structural blank check. In the next 1-3 quarters, the biggest upside likely comes from sentiment mean reversion and multiple expansion, not immediate fundamental inflection; customers will test AI workloads in small pilots first, then broaden licenses only if governance and security are solved. That means the best software longs are the names where AI increases active usage per customer and where pricing power is already embedded in the install base.

For Microsoft specifically, the asymmetry is better than the market is implying because the downside from “AI displacement” is already well-discounted, while the upside from copilots/agents as incremental authenticated users has not been fully underwritten. The contrarian miss is that AI may shift spend from labor budgets into software budgets faster than seat counts fall, which is bullish for large platform vendors and a headwind for labor-light discretionary software spend elsewhere. The biggest risk is a slower enterprise procurement cycle if CFOs decide to wait for a clearer ROI on agent deployments, which would keep the de-rating in place for another 2-3 quarters.

Key Decisions for Investors

• Add to MSFT on weakness over the next 2-6 weeks; expect the stock to outperform the software complex as the market rotates from headline AI fear to install-base monetization. Risk/reward is favorable for a 6-12 month horizon because the multiple has already compressed, while agent-driven seat expansion is a plausible earnings-upside catalyst.
• Pair trade: long MSFT / short IWM or a basket of lower-quality software names most exposed to seat compression and pricing pressure. The thesis is that platform winners capture the agent layer while smaller vendors face slower renewal cycles and higher churn risk over the next 2-4 quarters.
• Use call spreads rather than outright equity for MSFT if entering near-term: buy 3-6 month at-the-money calls and fund with higher-strike sales. This expresses a mean-reversion view with defined downside if enterprise AI adoption stalls.
• Reduce exposure to software names with weak net retention and limited workflow ownership; those are the most vulnerable if AI reduces human seats without increasing authenticated usage. Best expressed as shorts over 3-9 months if valuation remains sticky while growth decelerates.
• For investors already long software, hedge with a small short in the broader tech-software ETF over the next 1-2 quarters. That isolates the winners from the market’s current over-discounting of the entire group while protecting against a longer digestion period.