Ali Ghodsi 1.9 10 ideas

Ghodsi explicitly stated that AI will kill SIEM systems this year, citing inefficiencies in legacy systems such as high costs, hard data ingestion, and manual labor. Splunk, Crowdstrike, and Palo Alto are major providers of SIEM solutions; disruption to the SIEM product category directly threatens their core business segments and revenue streams. These companies face significant obsolescence risk unless they adapt quickly, making them unattractive investments in the short to medium term due to potential decline in SIEM demand. The companies successfully pivot to new models like the open security lakehouse, diversify their offerings, or leverage partnerships to mitigate disruption.

There is unprecedented capital expenditure ($50B–$100B) flowing into hardware, data centers, and energy. While the AI trend is real, the current build-out creates a risk of "overbuilding." The market is pricing in perfection, but physical constraints (energy) and ROI questions remain. The sheer volume of capital chasing hardware, combined with "circular" funding deals in the AI startup ecosystem, mirrors the pre-crash vibes of 2000. If AI adoption accelerates faster than hardware supply, these stocks will continue to run despite valuation concerns.

Chinese models and open-source alternatives are catching up to US closed models rapidly. Models like "Kimi" and "DeepSeek" are performing nearly as well as top-tier US models but at a fraction of the cost (or free). This creates a "race to the bottom" for pricing power among US model providers. Databricks' largest customers are offloading workloads to cheaper Chinese/open models for cost efficiency. Geopolitical regulation or chip bans could stifle the progress of Chinese models.

Databricks raised $7B despite being cash flow positive. There are "2000 vibes" (Dot-com bubble) in the market. If the bubble bursts, capital markets will freeze for 3-4 years. Having a massive cash pile allows a company to survive a wipeout and acquire distressed assets while competitors conduct layoffs. Startups with zero revenue are raising capital at multi-billion dollar valuations, a classic bubble signal. If the bull market continues uninterrupted, holding excessive cash drags on returns (opportunity cost).

Chinese models (like Kimi and DeepSeek) and open-source models are creating a price ceiling for US tech. These models are "good enough" (slightly behind US models) but significantly cheaper or free. This forces US hyperscalers to lower prices to compete, preventing them from maintaining massive margins on pure compute/token costs. Large Databricks customers are offloading high-volume tasks to Chinese models to save money. Geopolitical regulations could ban the use of Chinese models by Western enterprises.

AI agents and "vibe coding" are taking over software creation. 80% of Databricks' new databases are created by AI, not humans. AI agents write software faster than humans. Every piece of software needs a database to store information. Therefore, the volume, usage, and optimization of databases will increase significantly over the next few years. Databricks internal data shows a massive spike in AI-generated database creation. The optimization must shift from catering to human administrators to catering to AI agents; legacy databases that cannot adapt to "agent-native" interaction may fail.

Despite the AI disruption narrative, legacy giants like SAP are insulated. SAP possesses a "Data Moat." They are the "system of record" for global enterprises. It is incredibly difficult to replace them because they hold the critical operational data. There is no viable alternative to rip and replace their infrastructure, and they are successfully integrating AI into their existing stack. Ghodsi asks rhetorically, "What would you replace SAP with?" noting there is no real alternative. Slower execution on AI integration compared to agile startups.

Databricks is delaying its IPO despite being free cash flow positive. The company prefers private capitalization to weather potential public market volatility (fearing a repeat of 2022). They want to invest in long-term R&D (5-10 years) without the quarterly pressure to cut costs or layoff staff if the stock drops. Raised Series L funding specifically to stay private longer. Prolonged delay could fatigue early employees or investors seeking liquidity.