How Edo Liberty Built Pinecone: 40 VC Rejections to a $750M Vector Database

August 17, 2026

TL;DR: Pinecone is the fully managed vector database that became core infrastructure for AI applications — the place developers store and search the embeddings that power semantic search, recommendations and RAG. It was founded by Edo Liberty, who left AWS in 2019 after building SageMaker, at age 40, with a four-year-old and one-year-old twins at home. He was rejected by roughly forty VCs and had six months of runway left before one rewritten pitch — thrown together in a Google Doc at 10pm the night before a meeting he almost cancelled — closed a $10 million seed led by Peter Wagner at Wing. Pinecone went on to raise a $28 million Series A led by Menlo Ventures and a $100 million Series B led by Andreessen Horowitz in April 2023 at a $750 million valuation — $138 million in total. Along the way, its own users had to tell the company what its product was called.

Edo told the full story on The Product Market Fit Show.

What does Pinecone do?

Pinecone is a managed vector database. When a model processes an image, a document or a sentence, the output is an embedding — a long list of floating point numbers that encodes meaning. Pinecone stores those, organizes them, and lets you search across millions or billions of them in production.

Edo's own explanation of why embeddings are strange and powerful is one of the clearest on record:

"So you start in a sentence in Arabic and you end in a sentence in French. In the middle of that network, there is a numerical representation that knows nothing about the original sentence, but somehow it can translate it into French. So clearly the meaning is hidden inside that representation somehow." — Edo Liberty, Pinecone

"It might be opaque to you and me when we read this thing, but for models, it's everything. It has the actual meaning, it doesn't care about the tokens, frankly doesn't even care about the language." — Edo Liberty, Pinecone

Crucially, the category did not exist when he started:

"The term vector database didn't exist, it just didn't exist. Nobody talked about it this way." — Edo Liberty, Pinecone

Key stat: more than 5,000 customers have shipped AI applications on Pinecone, per the company; it employed roughly 128 people as of 2026.

Who founded Pinecone?

Edo Liberty started in physics at Tel Aviv University and took a computer science minor purely because he figured he'd be a bad physicist if he couldn't code. He did his PhD at Yale, bounced across data science, functional analysis, graph theory, high performance computing and computer vision, and fell in love with machine learning under brutal hardware limits — 512MB of memory, three images counting as big data.

"That's where I found my passion on marrying the machine learning aspects of it and the optimization functions you're trying to solve with a very limited resources at hand." — Edo Liberty, Pinecone

He ran data platforms at Yahoo, then joined AWS in 2016 as one of three people tasked with building AI services, ended up managing roughly a hundred scientists and engineers, and built SageMaker. Vector search kept showing up inside every system he touched — ad servers, recommenders — always rebuilt from scratch, never named.

"I had built the internal workings of vector databases, God knows how many times, 3, 4, 5, 6 times in different products." — Edo Liberty, Pinecone

Leaving was, by his own account, the wrong decision on every axis except conviction:

"I was 40, I had already had a pretty successful-, I was an adjunct professor and a director and this and that. I was already in a place in life where going back to walking in the desert and working my ass off, not seeing my kids and making less than an intern at Facebook is not necessarily the wisest decision" — Edo Liberty, Pinecone

The kitchen-table conversation with his wife — then working full-time at a hospital, later in the Covid ward at Stanford — came down to a spreadsheet and one condition: keep paying the mortgage.

"My suggestion is to not factor in financials into what a good outcome looks like." — Edo Liberty, Pinecone

How much has Pinecone raised?

Friends-and-family round: a small amount to fund a team of five or six engineers plus one generalist who called himself "chief of everything else."
Seed — $10 million, completed after ~40 rejections, led by Peter Wagner of Wing, who had previously backed Snowflake and knew data infrastructure.
Series A — $28 million, led by Tim Tully at Menlo Ventures. Tully was formerly CTO of Splunk and had worked with both Edo and Pinecone CTO Ram at Yahoo, where they'd built this exact kind of system internally. Revenue at the time was somewhere around or under $100K.
Series B — $100 million, April 2023, led by Andreessen Horowitz with ICONIQ Growth, Menlo Ventures and Wing, at a $750 million valuation.

Key stat: $138 million raised in total across three rounds; $750 million valuation as of the April 2023 Series B.

Where Pinecone sits in 2026: Edo Liberty moved from CEO to Chief Scientist in 2025, with Ash Ashutosh appointed CEO. Calcalist has reported that the company has been weighing a sale amid rising competition in the vector search market.

The pitch that worked after 40 that didn't

Every investor in the Bay Area took the meeting. His profile — the SageMaker guy leaving AWS at the peak of the MLOps hype — read as automatically fundable. Then he'd explain what he actually wanted to build.

"Correct. But then I would explain what I want to do." — Edo Liberty, Pinecone

"They're like, I have no idea what this shit is. It's got nothing to do with what I thought you were going to do. Frankly, I don't get it." — Edo Liberty, Pinecone

"I swear to God there's not a third rate VC in the Bay Area that I have not pitched to with everything I got and got the door slammed in my face." — Edo Liberty, Pinecone

"I remember towards the end of the journey I was, I was giving up, we had six more months of runway and effectively failed to raise the seed round." — Edo Liberty, Pinecone

The last meeting on the calendar was one he planned to cancel. It fell on the first day of his son's first grade during Covid, with his wife working the hospital's Covid ward. At 10pm the night before, he threw out the deck they'd spent months polishing.

"I'm going to just rewrite the whole thing in a freaking whatever, Google Docs. I'm just going to do the most bare bones thing, but explain it in a completely different way." — Edo Liberty, Pinecone

"And I went on the call and I said, Hey, forget everything that I told you. We're just building a different kind of database. That's it. It's a different kind of database. This is what it does really well. This is what it doesn't do. I think it's big and here's why that completely changed the story." — Edo Liberty, Pinecone

The seed closed within weeks. The old deck had tried to explain the value by walking through every application you could build on a vector database — which left investors unable to say what the company was. The new one gave them a category they already understood the economics of.

Letting customers name your product

Pinecone shipped V1 as a self-serve service with a strong core engine and almost nothing else — no billing, no metering, no autoscaling. They had a Slack alert fire whenever anyone so much as opened the console.

They also had no working name for what they'd built. So they asked the ten or twenty teams using it:

Never miss a founder's PMF story

Subscribe to The PMF Show

"We told you what the API is, but we never told you what this thing is called." — Edo Liberty, Pinecone

"Well, what do you call this internally, 7/10 would say, we just call it a vector database. It's like that's where we have our vectors, we have our embeddings. You guys are the database and we just call it that." — Edo Liberty, Pinecone

Then came the positioning epiphany, around 2022, after hiring Greg Cogan as VP of marketing purely to explain what the thing was. Every marketing instinct they'd been given was wrong:

"It's insane, insane. Every investor, every business leader, every marketing guru, everybody gave us the absolute worst advice." — Edo Liberty, Pinecone

"If you don't already experience that, you're not our customer." — Edo Liberty, Pinecone

"It's not what you're going to build. It's not why it's important for your business. It's not how it's going to propel your application forward. It's not about the time you'll save. It's not about your developers, it's nothing. It's an API. It's a vector database. Here's the API, here's how much it's going to cost you. Here's the performance. If you like it, great. If you don't like it, go to hell." — Edo Liberty, Pinecone

"All the metrics, all the adoption, everything just completely was like a discontinuity. It went from essentially nothing to something that looks like adoption." — Edo Liberty, Pinecone

His reasoning for why stripping the marketing worked is a rule for anyone selling to engineers: a professional-grade tool should read like a spec sheet, not a lifestyle brand. If you're building in this category, see product-market fit for developer tools.

ChatGPT, and rebuilding the database mid-flight

Pinecone ended 2022 with roughly $2 million in revenue and a few hundred customers. Then RAG became the default way to get your own data into a model, and Pinecone was, in his words, the only game in town.

"We started getting thousands of signups a day. We would spin up more and more and more machines. We got to the point where we would saturate whole regions in AWS and GCP. We just couldn't meet the demand. It was insane." — Edo Liberty, Pinecone

"We didn't sleep for three months. We would wake up in the morning and revenue's up 5%. It was like, okay, what happened? We don't know. Let's go look at the logs." — Edo Liberty, Pinecone

The Series B was the exact inverse of the seed:

"It was I was invited for VCs to pitch to me why I would be, I swear to God, I remember it as a cultural shock. VCs would send me their pitch deck on Pinecone about why it's a great investment." — Edo Liberty, Pinecone

Key stat: Pinecone 5x'ed its paying customers between signing the Series B term sheet and closing the round, as Edo shared on the show.

Because they ran the managed service, they could see production workloads roughly eighteen months before the rest of the market — and what they saw said the architecture wouldn't hold at 100x. So they rebuilt the whole database while serving live traffic, separating reads and writes, moving storage to blob, building in multi-tenancy. Serverless launched at the beginning of 2024.

"For a database to go rebuild what it is mid-flight and keep serving all these workloads was a massive strategic bet" — Edo Liberty, Pinecone

Key stat: the rebuild cut costs enough to drop prices roughly 10x — customers paying $3,000–$4,000 a month started paying about $200, as Edo shared on the show. Related reading: when to rebuild your startup's product and startup data moats.

Key lessons from Edo Liberty's playbook

1. Pitch the category, not the applications. Forty decks explaining what you can build with a vector database failed. "We're building a different kind of database" worked in one meeting. 2. Ask your users what your product is. Seven out of ten said "vector database." That became the category name for an industry. 3. Marketing to experts means removing marketing. API, price, performance. If the buyer doesn't already feel the pain, they aren't your buyer and no copy will change that. 4. Being the managed service is an information advantage. Seeing real workloads early is what told them to rebuild before the market forced it. 5. Know what you control. His closing advice:

"I think the one thing that founders tend to misjudge is how much control they have over anything in both directions." — Edo Liberty, Pinecone

"You also overestimate your ability to influence the market and the timing. One of the things that we have had amazing good fortune with is ChatGPT and AI shooting up to the sky. That's a market thing. We didn't do that." — Edo Liberty, Pinecone

The best moment in the episode is an investor who rejected him, meeting him again during the Series B: I looked up my notes and I now understand that what you were pitching that you were going to do, you actually did.

"And I said, yeah, I mean isn't that how it usually works? He was like, no!" — Edo Liberty, Pinecone

Listen to the full interview: He got rejected by 40 VCs & had 6 months of runway — 2 years later, he raised $100M from a16z | Edo Liberty, Founder of Pinecone

FAQ: Pinecone

Q: What is Pinecone? A: Pinecone is a fully managed vector database. It stores and searches the embeddings produced by AI models, powering semantic search, recommendations, deduplication, anomaly detection and retrieval-augmented generation (RAG) at production scale.

Q: Who founded Pinecone? A: Edo Liberty, formerly of Yahoo and AWS, where he built SageMaker. Ram is the company's CTO and previously worked with Edo at Yahoo. Edo moved from CEO to Chief Scientist in 2025, when Ash Ashutosh was appointed CEO.

Q: How much funding has Pinecone raised? A: $138 million across three rounds: a $10 million seed led by Wing, a $28 million Series A led by Menlo Ventures, and a $100 million Series B led by Andreessen Horowitz in April 2023.

Q: What is Pinecone's valuation? A: $750 million, set at the April 2023 Series B. No newer valuation has been announced publicly as of August 2026.

Q: Who are Pinecone's investors? A: Andreessen Horowitz, Wing Venture Capital, Menlo Ventures and ICONIQ Growth.

Sources: Listen to the Full Founder Story

Edo Liberty, Founder of Pinecone — how the man who built SageMaker got rejected by forty VCs pitching a category that didn't have a name yet, then had to rebuild his database mid-flight when ChatGPT arrived.