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

Ben Horowitz founded Loudcloud in the middle of the dot-com bust and sold it for $1.6 billion, then led Andreessen Horowitz from its founding to $46 billion in committed capital. Ali Ghodsi co-founded Databricks, stepped in as CEO during a crisis, and led it to a valuation of over $100 billion. In this episode of “Boss Talk”, Ben and Ali join a16z General Partners Sarah Wang and Erik Torenberg to share founder war stories, how to hire and make deals, how to keep culture intense without burning employees out, and why founders should raise their ambitions even higher. Follow Ali on X: https://x.com/alighodsi Learn more about Databricks: https://www.databricks.com/ Timecodes 00:00 Boss Talk returns 01:01 Why Ali became CEO of Databricks in 2016 09:45 From academic to CEO 16:00 Radical candor feedback and developing high performance 19:10 Scaling intensity and culture with Databricks’ ethos 31:55 The Microsoft deal strategy timing tactics 39:00 Fighting through setbacks and sealing the partnership 42:05 Building vs buying, how Databricks approaches acquisitions 54:55 Turning down acquisition offers and aiming for trillions 1:03:45 Key pivots luck and the Databricks founding team legacy Follow Ben on X: https://x.com/bhorowitz Follow Sarah on X: https://x.com/sarahdingwang Follow Erik on X: https://x.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details, please see a16z.com/disclosure

Ali GhodsiguestErik TorenberghostBen HorowitzguestSarah Wanghost

Oct 14, 20251h 4mWatch on YouTube ↗

WHAT IT’S REALLY ABOUT

Databricks’ CEO playbook: pivots, culture, deals, and talent scaling

Ghodsi explains why he became CEO in 2016: Databricks needed painful internal pivots to differentiate beyond free open source and commit to enterprise go-to-market.
Horowitz highlights Ghodsi’s CEO “superpowers”—deep technical/product competence, fast learning in BD and sales, and decisive pattern recognition that drives big strategic swings like entering data warehousing.
They argue high-performance culture scales through founder example, hiring for intensity via references, organizational design that preserves individual impact, and frequent direct feedback rather than annual-review surprises.
The Microsoft partnership is presented as a template for big-company deals: align to perfect timing, force meaningful “skin in the game,” expect multiple near-deaths, and win through relentless on-the-ground stakeholder work.
On M&A and talent, Databricks prioritizes people/culture and product integration over buying revenue, resists “Excel-sheet” acquisitions, and retains top performers by maximizing growth, learning, and mission—not just compensation hype.

IDEAS WORTH REMEMBERING

5 ideas

Your open-source success can become your biggest competitor.

Databricks faced constant pressure from “just download it” and cloud vendors bundling similar offerings; the company had to create real differentiation (product and GTM) rather than relying on community adoption alone.

The hardest pivot is admitting your instincts stop working outside your archetype.

Ghodsi stresses founders often excel in their native domain (e.g., engineering) but fail in sales/marketing because their intuition is wrong; the remedy is deliberate study, learning from top practitioners, and hiring experts who aren’t “like you.”

Feedback works best when it’s framed as help and delivered continuously.

They criticize annual-review “surprise” feedback and overdone ‘radical candor’; frequent, specific course-corrections normalize critique and prevent the shock of being blindsided or fired without warning.

High-intensity cultures scale through example plus systems that preserve impact.

Working hard at the top sets tone, but Ben notes intensity collapses if org design creates dependency gridlock where effort doesn’t change outcomes; people grind when they feel autonomy and visible impact.

Big-company partnerships require a real trade and real commitment—or they evaporate.

The Microsoft deal worked because each side had something the other needed (portfolio gap vs. distribution), and Databricks pushed for a big enough pre-commit that internal champions had career risk if it failed.

WORDS WORTH SAVING

5 quotes

You can sell, you're gonna make a lot of money, and you'll be super successful in life. But, you know, if you're like me, you're gonna look back the rest of your life thinking, you know, I missed that one shot. That was the one thing. I should have taken it all the way, and now I'll never know how far I could have taken it.

— Ben Horowitz

Actually, your biggest enemy, uh, 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?"

— Ali Ghodsi

Admit that you don't actually know everything about the job. Uh, so A, like, you know, s- first step of Anonymous Alcoholics, like- yeah, admit you have a, admit you have a, admit you have a problem, okay? And then number two, uh, be a student and learn everything you can about it, right?

— Ali Ghodsi

If you're the hardest working person, you know, it kind of everything will take care of itself from there on.

— Ali Ghodsi

Any big deal of that size, you lose at least three times before you win it.

— Ben Horowitz

2016 CEO transition and painful pivotsOpen source monetization and differentiationFrom academia to commercial leadership learning loopRadical candor, feedback frequency, and performance cultureScaling intensity without burnoutMicrosoft Azure partnership: timing, leverage, execution gritBuild vs buy and acquisition integration philosophyTurning down acquisition offers; thinking in trillion-dollar outcomesComp strategy, AI talent wars, and boomerang hiresLuck, timing, and funding near-misses

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.