AI Systems of Record
nihal@nihal
+2.5%
since Sep 29
The idea
I believe AI adoption keeps compounding, and over the next two to three years the real money shows up in software revenue, not just chips. So this basket anchors on the systems of record, the companies that already own a company's workflow and data, since their AI features become easy upsells rather than new sales. I add data platforms that make enterprise AI run, a few names already proving AI revenue in earnings, and security platforms with similar data moats. It's concentrated and growth-heavy to match that short window, with some cash held back to buy dips if the thesis wobbles.
What it owns
- These companies already own the workflow and data layer at their customers, so AI features sell as upgrades into contracts that already exist. · ServiceNowIt owns the workflow layer at large enterprises, so AI agents sell as upsells into existing contracts rather than requiring new customer acquisition., CRM, IntuitIt owns the accounting books for millions of small businesses, so AI that automates filing and categorization embeds into subscriptions already in place., SAP SEIt is the resource-planning backbone at the world's largest companies, the deepest system of record that can sell AI layers into established contracts., WorkdayIt owns HR and finance records competitors cannot reach, data that AI agents need, so it can sell AI features as paid tiers into its existing base., Veeva SystemsIt is the system of record in regulated pharma, where switching costs are steep and AI built on trusted data is a straightforward upsell., HubSpotIt is the record-owner for smaller companies, so its AI features go out as upsells to a base that already runs marketing and sales on it.43%
- Companies already showing AI revenue in their reported earnings, which proves the adoption bet is landing and converting to paid growth. · MicrosoftIt is the largest proof AI revenue reaches scale, with Copilot selling into Office and Azure contracts at a level no competitor has matched in earnings., AdobeIt charges for AI inside Creative Cloud subscriptions people already pay for, creating an upsell dynamic where AI tools enrich the base rather than eroding it., AppLovinIt is already collecting revenue from AI driving its ad engine, with the money showing up in earnings today rather than in a promised future., DuolingoIt is proving AI revenue on the consumer side, selling AI-powered tiers to subscribers it already has while raising what each user pays.18%
- Enterprise AI runs on the data these platforms store and organize, so AI adoption pulls their revenue up as consumption grows. · Palantir TechnologiesIt is the data platform already signing operational AI contracts, the infrastructure backbone the take says enterprise AI needs to actually run and scale., DatadogIt records how software actually runs in production, and every AI system deployed adds more infrastructure to monitor and more usage-based revenue., SnowflakeIt is where enterprise data lives, and AI workloads mean more queries, more consumption, and more spending on its compute and storage., MongoDBIt is the database layer new AI applications get built on, so the wave of AI apps needs to translate into growing consumption and spend.16%
- Security platforms hold a similar data advantage, and every new AI deployment creates more threat data and infrastructure to protect. · CrowdStrikeIt holds years of threat data that make its AI better and switching costly, the same moat as systems of record, compounding through module upsells., Palo Alto NetworksIt is consolidating security onto one platform, with AI features making that platform stickier and winning bigger deals as consolidation accelerates., ZscalerIt sits in the traffic path of its customers and holds telemetry data that makes its AI features valuable, positioning it for upsells and billings growth.12%
- Cash11%