Ali Ghodsi

Co-founder and CEO of Databricks

Mentions

Appeared in

Discussed in

Key positions and views

Artificial General Intelligence (AGI) has already been achieved according to the definition accepted by the AI community for decades, and the goalposts are now being moved.

Large Language Models (LLMs) are a commodity where price is the main differentiator, and most economic value in the AI stack will accrue to the application layer, not the model layer.

There is a significant bubble in the current AI market, evidenced by startups with zero revenue achieving valuations in the tens of billions.

A company's talent pool is its most critical asset, to the point that an otherwise strategically sound acquisition should be vetoed if the target's employees would dilute the existing talent quality.

The 'scaling laws' approach of simply building ever-larger models, pursued by frontier labs, may not be the correct path to achieving more advanced AI, a view he attributes to prominent scientists.

Podcast consensus on Ghodsi

Points of consensus

▶Ali Ghodsi consistently emphasizes a pragmatic, business-focused approach to building Databricks, highlighted by the strategic partnership with Microsoft for distribution, a talent-first M&A strategy, and learning from early failures like the Product-Led Growth model.Apr 2026

▶Across discussions, Ghodsi expresses a contrarian view on the AI market, asserting that there is a significant bubble, that Large Language Models (LLMs) are already a commodity, and that most long-term value will be captured at the application layer, not the model layer.Apr 2026

▶He repeatedly grounds his analysis in specific enterprise use cases, citing examples from Merck, 7-Eleven, and Royal Bank of Canada to illustrate how AI is delivering tangible business value today, moving beyond theoretical hype.

▶Ghodsi's perspective on AI development is shaped by his engagement with prominent AI scientists, referencing figures like Yann LeCun and Rich Sutton to support his skepticism about the 'scaling laws' approach and to frame the timeline for more advanced AI.Apr 2026

Points of debate

▶Ghodsi presents a conflicting view on AGI, stating unequivocally "I think we have AGI," yet also citing respected AI scientists who believe true advanced AI is a "20-year problem," suggesting a personal definition of AGI that differs from the one he attributes to others.Apr 2026

▶He is simultaneously bullish and bearish on the AI industry, predicting revenue growth for companies like OpenAI and Anthropic while also declaring an "AI bubble" defined by startups with massive valuations and zero revenue.Apr 2026

▶There's a tension between his view of LLMs as a commodity, like gasoline, and the significant strategic and financial resources Databricks and the industry are pouring into developing and acquiring proprietary model capabilities.Apr 2026

▶Ghodsi highlights the immense potential of AI to automate complex tasks, such as equity research at RBC, while also cautioning that its potential is overhyped in other areas like customer service and that attempts to replace RPA with generative AI have largely failed.Apr 2026

Key themes

▶Contrarian AI Market PhilosophyApr 2026

Ghodsi consistently argues against the prevailing market hype. He posits that Large Language Models are becoming a commodity, that a significant bubble exists in AI valuations, and that the ultimate economic value will accrue to the application layer rather than the foundational model providers.

This perspective suggests that Databricks' strategy is likely focused on building an indispensable platform and application ecosystem around commoditized AI components, rather than competing to build the largest foundational model.

▶Pragmatic Enterprise AI ImplementationApr 2026

Ghodsi focuses on the real-world application and challenges of AI in large organizations. He provides concrete examples of success (Merck's drug discovery, RBC's automated reports) and failure (attempts to replace RPA, initial automation efforts at Databricks), emphasizing that organizational and human factors are often bigger hurdles than the technology itself.

Investors should note his focus on tangible ROI and integration challenges, indicating a mature understanding of the enterprise sales cycle for AI, which is often slower and more complex than consumer tech adoption.

▶Hard-Learned Lessons in Company BuildingApr 2026

Ghodsi reflects on critical moments in Databricks' history, including the failure of its initial Product-Led Growth strategy, the competitive threat from its own open-source project, and the strategic necessity of the Microsoft partnership. He also details a disciplined M&A strategy that prioritizes talent quality above all else, even vetoing an otherwise sensible deal over talent concerns.

This theme reveals a leadership style that values adaptability, strategic partnerships for distribution, and an obsessive focus on maintaining a high-quality talent pool as the primary competitive advantage.

▶The Shifting Human-Computer InterfaceApr 2026

Ghodsi makes bold predictions about the future of user interfaces, driven by advances in AI. He speculates that speech interaction will become so effective that it will completely eliminate the need for keyboards and suggests that companies like Zoom are well-positioned to become the hub for automated data entry from conversations.

This forward-looking view indicates he sees the next wave of disruption not just in analytical capabilities, but in fundamentally changing how users interact with all software, placing a premium on conversational AI and data capture at the source.

Source episodes

Sentiment over time

Not enough data for timeline

Changes over time

2013

Databricks is founded, a date Ghodsi considers critical to its success due to market timing for cloud and AI.

Early Years (post-2013)

The company's open-source project, Apache Spark, achieves worldwide success but also becomes its biggest competitive challenge as cloud vendors offer it for free.

2016

Amidst operational challenges and a search for a new CEO, Ali Ghodsi is appointed to the role.

Post-2016

Databricks pivots its strategy after a Product-Led Growth (PLG) model fails, capping at just $3M ARR. The company forges a crucial, mutually beneficial partnership with Microsoft to gain enterprise distribution.

Recent Past

To compete with 'FAANG' companies for talent, Databricks shifts its compensation strategy to pay at the 95th percentile for engineering roles. The company also receives and presumably rejects a major acquisition offer valued at six times its valuation at the time.

Suggested prompts

How does Ali Ghodsi's definition of AGI, which he believes is already achieved, differ from the one he attributes to scientists like Yann LeCun and Rich Sutton? &nearr;Given Ghodsi's assertion that LLMs are a commodity, how does Databricks intend to build a durable competitive advantage in the AI platform space? &nearr;What specific human and organizational challenges did Ghodsi observe that caused initial AI-driven software engineering automation to fail at Databricks? &nearr;How does Ghodsi's M&A philosophy of prioritizing talent quality over revenue, as seen in his veto of a deal, shape Databricks' long-term product roadmap and culture? &nearr;

Key concepts

Databricks Business Strategy 2 ep AI Market Analysis 1 ep Enterprise AI Use Cases 1 ep M&A and Valuation 2 ep Artificial General Intelligence (AGI) 1 ep LLM Commoditization 1 ep Talent and Compensation 1 ep Competition and Partnerships 1 ep Future of Technology Interfaces 1 ep

Notable quotes

“I think we have AGI. I think we have artificial general intelligence. We really have it.”

Ali Ghodsi · AI Enterprise - Databricks & Glean | BG2 Guest Interview

“I think the LLM is a commodity. People are not saying that, but it is a commodity. Like, you can get gas from this gas station, you can get gas from that gas station, it doesn't matter. Just compare price.”

Ali Ghodsi · AI Enterprise - Databricks & Glean | BG2 Guest Interview

“The acquisition offer was on the table. It was six times bigger than the valuation we had at the time”

Ali Ghodsi · Ben Horowitz and Ali Ghodsi: How to Run a $100 Billion Business

“So your biggest enemy is your open source project. The main thing you have to fight in the market is, hey, why can't I just download the open source version? Amazon is offering it, the cloud vendors are just offering it, I'm just going to use that. So this was the biggest challenge that Databricks had at the time.”

Ali Ghodsi · Ben Horowitz and Ali Ghodsi: How to Run a $100 Billion Business

Report last updated: Apr 21, 2026

Get started free

Back to Entities Intelligence Report