Databricks co-founder shares his secrets for winning big enterprises customers

Databricks co-founder Arsalan Tavakoli-Shiraji said enterprise sales depend on people, processes, and distribution as well as strong technology; Databricks partnered with Microsoft early on to gain access to large customers. Databricks is now valued at $190 billion and serves more than 70% of the Fortune 500. Tavakoli-Shiraji said 99% of CIOs he speaks with are at the first or second stage of AI adoption, and identified modernizing legacy systems as an opportunity for startups.

Analysis

The investable signal is distribution and implementation capacity—not proof that enterprise AI is already generating material productivity gains. Microsoft can strengthen Azure’s role as the procurement and contracting route for startups: partner-led access lowers customer acquisition friction and can pull workloads onto Azure. The second-order benefit is ecosystem retention, but it is conditional; if customers use the channel to preserve model choice and move workloads across clouds, distribution may not translate into durable Azure lock-in. Treat this as incremental support for MSFT, not a standalone earnings catalyst.

IBM faces a two-sided read-through. AI-assisted discovery of legacy code could reduce the perceived cost and risk of modernizing mainframe estates, which may expand consulting and hybrid-cloud projects. Conversely, tools that make migration feasible could weaken the “too risky to touch” protection around installed systems. The distinction to monitor is whether IBM captures modernization work and ongoing workloads, or whether third-party agents help customers bypass IBM platforms. The article does not establish which outcome is occurring.

Near term, the claims are anecdotal and unlikely to reset large-cap estimates. Over 1–3 months, watch Azure partner-led workload commentary and IBM services/backlog and hybrid-cloud disclosures for evidence of monetization. Structurally, enterprise adoption is likely slower than AI enthusiasm implies: security, governance, integration, and cost controls extend deployment cycles. That lag can defer startup revenue and cloud consumption even as pilots proliferate. Falsification: sustained acceleration in production AI workloads would weaken the slow-adoption view; deterioration in IBM services demand alongside migration wins by alternatives would strengthen the displacement risk.

Key Decisions for Investors

• Keep MSFT as a modest relative-quality exposure, but do not add solely on this anecdote. Upgrade conviction only if Azure growth or management commentary shows partner-sourced AI workloads converting into production consumption; reassess if Azure growth decelerates or partner activity fails to show workload traction.
• Treat IBM as a monitor, not an automatic short: track services backlog/revenue, hybrid-cloud demand, and evidence that AI code-analysis tools are leading to IBM-led modernization versus migrations away from IBM platforms. A short thesis needs confirmation from operating data, not the existence of legacy systems.
• For the next 1–3 months, favor a watchlist of enterprise software and cloud names with demonstrable production deployments over firms valued mainly on AI pilots. Verify customer conversion, renewal/expansion, and inference-cost economics before underwriting revenue acceleration.
• No event-driven options trade is warranted from the article alone. Revisit after the next MSFT and IBM results; evidence of production deployment and repeatable customer spend is the key catalyst, while continued pilot-heavy adoption would leave the near-term monetization thesis unproven.