Four-time founder. The investor who failed at AI in 2011 - and came back for 55x.
Yan-David "Yanda" Erlich · B Capital Group
In 2011, Yan-David Erlich built an AI assistant startup and watched it die. Not because of bad execution. Because the models weren't good enough yet. That moment - building for the right idea at the wrong time - became the lens through which he'd eventually deploy hundreds of millions of dollars into the companies that would remake what software means.
He goes by Yanda. Born in Paris, relocated to Austin, Texas at age 9, educated at Rice University with a double degree in Computer Science and Electrical Engineering (Phi Beta Kappa), then Stanford GSB (Arjay Miller Scholar). The pedigree is real. But the more interesting part is the fourteen years of operator calluses he built before most VCs were taking AI seriously.
His first company, Mogad, was acquired by iSkoot (a Qualcomm division) in 2008. His second, ChoiceVendor, went to LinkedIn before LinkedIn itself went public in 2011. That early LinkedIn acquisition is a footnote now, but it told you something about his instinct for timing markets that were about to matter. The failed AI startup that came next? Same instinct, wrong window.
By 2013 he had co-founded Parsable - a software platform to digitize industrial workflows. He served as CEO until 2017, when he made a rare call: he recognized the company needed a different kind of leadership for its next phase. "This company needed a college professor," he said later. "I was the kindergarten-to-middle-school guy." Parsable was eventually acquired by Computer Associates International in September 2024.
After Parsable, he went to Coatue Management, where he built the enterprise venture practice from scratch and raised a $700M first-stage fund. He sourced deals not by throwing brand weight at founders but by actually showing up as a person they could talk to. In a market where every firm has the same logo hoodie, he built a pipeline on empathy. His portfolio from that era includes Weights & Biases, Covariant, Abacus.AI, Infinitus, Impira, and Raycast - names that read like a who's-who of AI infrastructure circa 2017-2021.
Then he stopped being a VC and became an operator again. He joined Weights & Biases as COO and CRO in 2021. Over four years he scaled revenue from $3M to $50M ARR and grew the enterprise customer list from zero to more than 800 - Meta, Samsung, NVIDIA, OpenAI. When CoreWeave acquired Weights & Biases in early 2025, his original angel investment had returned 55x DPI. It was a number most GPs post as a career highlight. For Yanda it was one line in a longer story.
In January 2025, B Capital Group announced him as General Partner and Head of Tech Investment Practice. The firm is a multistage global venture fund with deep roots in Southeast Asia and a portfolio spanning early-stage through growth. His mandate: lead technology and AI investments with checks ranging from $5M to $50M. His first announced bets at B Capital include Writer, Poolside, and Axiom - all firmly in the AI infrastructure and enterprise AI layer he has been building conviction around for years.
"AI is evolving from tools that assist (help me) to agents that execute (do this for me) to co-workers that own outcomes alongside you (own this with me)."
- Yan-David Erlich, B Capital investment thesisErlich has mapped a clear progression for how AI integrates into organizations. He's betting heavily on the third stage - the one most investors haven't started pricing in yet.
Across personal investments, Coatue, and B Capital - the AI infrastructure layer, built brick by brick.
"As an entrepreneur, the hardest move is taking the first step. After that, people will almost magically come out of the woodwork to help you."
- Yan-David Erlich on entrepreneurship"Most founders don't have someone they can talk to about the hard stuff. That's why I encourage each founder we back to work with a coach: someone outside the swirl who can help navigate the emotional, interpersonal, and strategic challenges."
On founder support"I actually think if we look over a 10-, 20-, or 50-year horizon, society is going to be reconstructed. The economy is going to get reconstructed. The types of jobs we're doing today are going to seem ridiculous."
On AI's long arc"Most significant improvements in human quality of life have come through the reduction of the cost of the goods that we need to have happy lives."
On economic impact of technology"Failing to meditate is meditation. It's not about pushing unproductive thoughts away, though that's part of it."
On mindfulness - posted to 10K+ followers on XThe cleanest thing about Yanda's career is the feedback loop. He builds companies, they work or they don't, and he takes notes. At Coatue he won deals not because Coatue was the biggest name in the room - it wasn't always - but because he had been the person sitting across the table. Founders trusted him because he'd earned the credibility of someone who had actually run out of runway, made the executive hire that didn't work, and given the board a number he couldn't hit.
The Weights & Biases chapter is the clearest example. He had backed the company as an investor at Coatue. Then he went inside it as COO and CRO. He rebuilt the go-to-market from scratch, expanded the enterprise customer base from essentially nothing to 800 of the most recognizable names in AI, and did it while the AI landscape was reordering itself faster than any playbook could track. The company was acquired by CoreWeave in early 2025. His angel position - the bet he'd made years earlier as a conviction investor - returned 55x.
What makes the B Capital role interesting is the mandate. B Capital isn't a pure-play AI fund or an early-stage-only shop. It operates across stages and geographies, which gives Erlich room to follow companies through the full arc he's seen from the inside - from the first $5M check to a growth-stage position. His focus on AI co-workers reflects a specific view: that the category creating the most value in the next decade won't be AI tools (that market is already crowded) but AI systems that take on accountability - that own outcomes, not just generate outputs.
He's not religious about sectors. He'll back legal AI, finance AI, customer service AI, consumer AI - wherever the co-worker model can take root. His filter is closer to: does this company understand the difference between a feature and a workflow, and can it defend a moat once the underlying models commoditize? Those are the questions he knows how to ask because he's lived them.
The regret-minimization framework he talks about publicly isn't just a quote. It explains every major career decision: staying at Parsable until the company needed a different gear, going inside Weights & Biases when most investors wouldn't, joining B Capital in January 2025 when the AI investment market is simultaneously the most crowded and the most important it's ever been. Each move made sense as the choice he'd least regret at 80 - not the safest or the flashiest, just the one that would hurt most to have avoided.