Every technology has a shape. Not the shape of the code or the architecture — the shape it takes when it meets the world. The packaging, the form factor, who encounters it first and what it looks like when they do. The container the thing arrives in.
I invest in deep tech at the earliest stages, and I keep noticing something. A founder shows me something genuinely brilliant — a new approach to data synchronization, a novel compute primitive, a breakthrough in how machines perceive their environment — and then they show me how they're planning to ship it. And I can feel it before I can articulate it: the container awkwardly doesn't fit the thing inside it.
I think this points at something the venture world mostly ignores. We talk endlessly about markets and teams and timing. But the actual product architecture — the decisions about what shape a technology takes when it meets the world — gets treated as an implementation detail. Something the founder figures out along the way.
Well — It doesn't get figured out along the way. It gets negotiated.
A packaging choice made under time pressure. A positioning that drifts because a design partner asked for something slightly different. A pitch that didn't resonate well. A larger founding round of a competitor. A form factor inherited from a demo that was never meant to be the product. Each negotiation feels small. But they compound. And undoing a shape that's already been shipped, sold against, and built around is one of the most expensive things a startup can do.
The dominant narrative in venture treats shape as a founder decision. The visionary sees the form, the team executes, the market validates. It's clean. It's also wrong.
Nobody shapes alone. The shape that actually ships is a negotiation between the founder's vision and everything else — the constraints of the technology, the expectations of the first design partner, the platform it runs on, the investor narrative that got the round closed. Some of these aren't even people — they're systems, habits, market structures. But they all participate.
And the negotiation never ends. Every major customer reshapes you. Every platform change reopens the question. The founders who navigate this well aren't the ones who found a shape early and locked it in. They're the ones who got good at the negotiation itself — who developed a feel for when the current shape is holding and when it's starting to crack.
Here's the thing that rarely gets named in board meetings: your product already has multiple shapes. The design partner sees a different product than the dev who found you on GitHub. The enterprise buyer experiences a different thing than the practitioner who adopted you bottom-up. These aren't segments. They're genuinely different shapes of the same technology, enacted differently by different people in different contexts. The question isn't how to collapse them into one canonical form. It's which shapes reinforce each other and which are pulling the company apart.
Let me make this concrete.
Electric sql built something genuinely novel — rich crdts, a synchronization primitive that most of the industry doesn't fully understand yet. Their first shape was oriented around the Phoenix LiveView and Elixir ecosystem — the right community at the right time, one that could appreciate the elegance and stress-test the ideas. That shape did its job. But shapes have seasons.
Then the shape shifted. Same core technology, repackaged around the Postgres ecosystem. Different developers, different use cases, different distribution. Breakthrough traction followed. The technology hadn't changed. The network it was negotiating with had.
And now they're mid-negotiation again. Their latest product, Durable Streams, steps beyond Postgres entirely — making ai workloads robust and stateful. The buyer changes. The urgency changes. A new set of actors enters the room. Three shapes from the same underlying science, each one a different conversation with a different world.
What makes Electric sql remarkable isn't that they found a shape. It's that the founders developed an instinct for when a shape has run its course and a new negotiation is needed. The core technology stayed the same. Everything around it kept moving.
I see the same pattern unfolding with another portfolio company: Embucket. They started with a clear shape: a Snowflake-compatible lakehouse you can run in your own cloud. The network was large enterprises looking to optimize costs. It worked. But the ceiling of that conversation has a limit.
Now a different shape is emerging — isolation and virtualization for agentic workloads. A warehouse per agent. Same underlying technology, but the network shifts: from cost-conscious enterprise buyers to forward-leaning teams experimenting with agent architectures. The icp changes. The energy in the room changes. And there's something fitting about it — the founding team comes from a virtualization background. The new shape isn't a pivot. It's the technology finally arriving at the conversation it was always built for.
Neither company abandoned their technology. They renegotiated its shape — found new actors, new conversations, new contexts where the same capability carried different weight. That's what good shape work looks like. Not finding the answer. Staying in the negotiation.
When everything except judgment gets commoditized — when code and research and market maps all compress toward zero — shape is the thing that compounds. The ability to sit in the negotiation between what's possible and what's adoptable, and feel which configuration holds.
Taste. In the deepest sense of the word.
Venture has built impressive infrastructure on the demand side. Introductions, networks, talent pipelines, customer access — real machinery for pulling companies into the market. And capital itself is a shaping force. Every funding round brings new actors into the negotiation — new board members, new expectations, new constraints on what the company can become. A round doesn't just fund the next phase. It reshapes the company's possibility space. Who you take money from, at what valuation, with what narrative — these are shape decisions, whether anyone names them that way or not.
But the supply side — the work of shaping what the company actually is, how the technology meets the world, what gets built and in what order — is catastrophically underserved. It's left to angels. It's accidental at best. It's not institutionalized as a practice. It's not treated as a discipline. And when it's absent, the damage is quiet but deep. Futures get foreclosed that nobody even noticed were possible. Companies spend years in configurations that were never going to work, and by the time anyone names the problem, the concrete has set.
I've had the privilege of working with co-investors who understand this — who show up in the room and do real shape work alongside founders. I've seen what it looks like when it's done well. I've also seen the opposite, more often than I'd like. The harm isn't dramatic. It's the slow erosion of a company's possibility space, one unexamined shape decision at a time.
This work isn't clean. It's not a three-month program at inception. It's years. It's messy, intimate, sometimes uncomfortable. It's sitting with a founder through the third reshaping when the first two didn't hold. It's being honest when the current container isn't working, even when everyone else in the room has reasons to pretend it is. It's staying in the trouble.
I'm here to make the case that shape work is essential to what it means to add value as a venture investor. Not a nice-to-have. Not a soft skill. The hard thing that actually matters — and the thing almost nobody is doing deliberately.