the software opportunities based on the Barron's piece (Adam Levine, July 17, 2026):
SOFTWARE / AI: WHERE THE BARGAINS ARE
Thesis: AI is eating software's golden gross margins (70–90%) as agents and consumption-pricing erode seat-based subscriptions. Selling has stopped being indiscriminate — the market is now sorting winners from the buried. Hardware won first (PHLX Semis +108% YoY), but the profit eventually shifts up to the application layer.
ALREADY WINNING — AI tailwind, but pricey (fwd P/E 37–148x) The data + cyber names that AI needs. Own the proprietary-data and security layers that fuel agents.
- CRWD / PANW / OKTA — security expands as AI creates new attack surfaces (prompt injection, agent auth). Stocks +95% / +112% / +105% in 3mo.
- SNOW — data lakes for agents; already on consumption pricing (66% GM). +88% in 3mo.
- DDOG — monitors AI infra, avoids seat pricing (80% GM).
- PLTR — operating layer linking people/AI/data (84% GM).
- Caveat: great businesses, expensive stocks — limited value here.
THERE'S STILL HOPE — the actual value plays Cheap, deeply embedded incumbents that can adapt if leadership sacrifices margin and cuts S&M.
- MSFT — fwd 20x, owns ~25% of OpenAI + cloud rental as offsets.
- SAP — fwd 18x, <40% of sales from per-seat licenses; sticky in regulated/gov.
- CRM — fwd 12x (was 20x YTD), hustling hardest: Agentforce, consumption pricing, $8B Informatica deal, "headless" agent-only product.
- WDAY — fwd 13x, cloud-native HR.
- Best risk/reward for value hunters — embedded, cheap, adapting.
PERMANENTLY CHALLENGED — avoid No lock-in, disruptable by cheap AI. ADBE (a $70 Creative seat vs a $20 ChatGPT), DOCU, and point solutions ASAN / MNDY. CHEG (~$0.78) flagged as the market's first AI casualty.
Bottom line: Winners already priced richly; the real bargains sit in the "Still Hope" bucket — MSFT, SAP, CRM — where cheap multiples meet enough embeddedness to survive the transition.