Jonathan Lehr

Enterprise VC @Work_Bench | We Lead Seed Rounds with $2-4M Investments | Organizer @NYETM, @KauffmanFellows 19 | My better half is @MichaelaLehr | 305 Native

[Jonathan Lehr @fendien]

🚨 Big news: We’ve raised $160M for @Work_Bench Fund IV to back Seed-stage founders with massive enterprise ambitions.

Enterprise is in our DNA and with Fund IV, we’re quadrupling down.👇

[Jonathan Lehr @fendien]

@doppenhe @authzed has an answer for that!

[Jonathan Lehr @fendien]

@Hadley @Bonecondor dang don't hate on Olive Garden! it's a @Work_Bench team favorite

how can you ever go wrong with unlimited salad and breadsticks 🤔

[Jonathan Lehr @fendien]

🗽🗽🗽

[David Haber @dhaber]

The best founders don't end up in New York by accident. They come because this is where the hardest problems, the biggest customers, and the highest stakes are. The city demands you build something real.

I've always believed the best opportunities live between fields of expertise — not inside any single industry, but at the intersections. No city in the world has more intersections than New York. Finance and media. Healthcare and technology. The customers who need AI to actually work — banks, hospitals, newsrooms, asset managers — are all here, within a few miles of each other.

AI adoption is accelerating faster than any cycle I've seen. Per @PitchBook, AI is now 65% of NYC venture deal value, up from 16% six years ago. Early-stage funding hit $8.9B in 2025, up 50% in a year. Active unicorns: 32 in 2019, 159 today.

The infrastructure of a generational tech hub is being built in real time.

We're still early. The biggest deals haven't happened. The companies that define applied AI haven't been built.

The founders who build them will be ambitious. They'll want to be where the problems are.

They'll be here.

NY @Techweek_ starts today. Come build 🗽

[Jonathan Lehr @fendien]

@SierraPlatform @btaylor @btaylor we'd love to host you at a founders dinner with local AI founders if you're in NYC on 6/23 or 6/25.

We did one with @mitchellh and his AMA was a hit.

You'd bring super valuable insights for the group!

https://t.co/Tjs76f6vDy

[Jonathan Lehr @fendien]

What an absolutely amazing NYC Infra leaders dinner last night with our special guest @mitchellh of @HashiCorp

Mitchell was so thoughtful and open in sharing lessons from his journey founding and scaling Hashicorp

And the quality of folks in the room led to great contributions to the conversation

If you're building in infra and want to join for a future session like this, hit me up

cc @kelleymak @bobbytables @JacobMoshenko @yevgenypats @chr_norm @DonnyGreenberg @kegilpin @asklyndon @zshrm @MerrillLutsky @ajwerner

[Jonathan Lehr @fendien]

Read @ychernova's full article in the @WSJ

https://t.co/Obb42xwwza

[Jonathan Lehr @fendien]

I appreciated speaking with @ychernova of the @WSJ for a story she wrote about the NYC and SF tech ecosystems.

Pre-COVID, roughly 70% of our @Work_Bench portfolio was NYC-based. Today it’s closer to 60% by company count. That shift isn’t because we have less conviction about NYC - quite the opposite. The NYC enterprise ecosystem has become dramatically stronger over the last decade. There are simply more technically ambitious founders, more repeat founders, deeper operator talent, and far more enterprise customers willing to engage early.

That said, our job is ultimately to find the best enterprise founders wherever they are. The biggest change over the last few years is that AI has reshaped where different categories of companies tend to emerge.

For vertical AI applications and a lot of enterprise software broadly, NYC continues to be an incredible place to build. The proximity to large financial institutions, media companies, healthcare organizations, retailers, and other Fortune 500 buyers creates a huge advantage around customer discovery, design partners, and early go-to-market execution. We’ve always believed enterprise go-to-market is massively underrated at Seed, and in today's world where distribution becomes your moat, NYC remains one of the best places in the world for that.

Infrastructure is a different story. You can’t ignore San Francisco right now, particularly around AI infrastructure, developer tooling, and frontier model ecosystems. There’s an intensity and density there across researchers, infrastructure engineers, ex-founders, talent from the major labs and hyperscalers that creates a unique environment. Having said that, we're excited for what the next wave of infrastructure talent will build in NYC over the next few years as people who have scaled AI-first companies like Modal, Coreweave, Cursor, and others eventually leave to launch their own companies.

In cybersecurity we continue to see exceptional technical talent coming out of Israel, and we often partner with Israeli pre-seed and seed funds where we can serve as the US Seed lead and help companies navigate early enterprise go-to-market.

What hasn’t changed is how we think about winning. Whether a founder is in NYC, SF, or Tel Aviv, the common thread is usually that they want a highly engaged Seed partner who can materially help with early enterprise traction across customer introductions, ICP development, pricing, positioning, and building a repeatable go-to-market motion. That has always been the core of the Work-Bench strategy, and in many ways it matters even more in today’s market.

[Jonathan Lehr @fendien]

@ProbyShandilya @Work_Bench lfg!!

although @DanielChesley and I may be rooting for different teams 🙈

@ProbyShandilya the question is who you'll root for...

[Jonathan Lehr @fendien]

@largedatabank @CockroachDB Congrats on such an amazing ride and kudos to the incredible impact you made during your time at @CockroachDB 🙌🙌

[Augustine Mavor-Parker @MavorParker]

The unix terminal is the natural interface for agents to get work done on a computer but how well can agents actually use unix?

Claude Code. Codex. Devin. Every frontier agent ships as a terminal tool.

With unix-ctf, Vmax is using setters and solvers to measure Unix competence.

[Jonathan Lehr @fendien]

"The app companies that will work are not wrappers around models. They will be systems wrapped around metrics. A wrapper exposes a capability, but a product takes that capability and makes it show up in a number the customer already tracks.

That is where the data, workflow, feedback loops, integrations, evals, and deployment model matter. A thin interface to a model gets eaten. A system wrapped around a metric has a chance."

Excellent read 🎯

[Rory O'Driscoll @rodriscoll]

Does Corporate America want to buy 1Trillion of token value on a wholesale basis, direct from the Frontier labs or intermediated through AI Apps built for specific vertical and horizontal use cases?

There will be some build-your-own, but as we are seeing in real time, that is harder and more expensive than it looks. Our belief is that the AI Apps business will be just as vibrant as the prior SaaS apps business, and we are investing accordingly. Thoughts from @siddharthvader_

[Jonathan Lehr @fendien]

@RMB as long as you don't mind not getting greeted by a guy in a suit it's a deal 😂

[Jonathan Lehr @fendien]

We loved having you @RMB 🙌

[Rob Bailey @RMB]

Working from @Work_Bench today! They are one of my favorite early stage investors and awesome to work with. The community they have built is super helpful for anyone scaling in AI/B2B. DM me if I can help you with an intro.

[Proby Shandilya @ProbyShandilya]

Few insights are truly as earned as Vas's on Forward Deployed Engineering. His clarity on how AI will be metabolized by organizations is special, and anyone thinking about this should read the post below. Sharing some of my favorite parts:

"The FDE is a highly skilled engineer who can understand the customer's problems very deeply, write code into a code base they've potentially never seen before, and communicate the business impact to a non-technical decision maker to close the deal. This is a million-dollar hire." The role is truly multidimensional!

"Don't overuse AI when building your agents. Most automation tasks can be done with a series of tool calls and just one call to an LLM as an orchestrating layer. Too much AI leads to unneeded token costs (which compound at scale) and often a lower-quality output."

"Evals prove value to the customer. While everyone says they want AI in their company, there are still many who are skeptical of whether it works. A good agent evaluation is what an executive needs to trust the agent will provide ROI."

I'm excited to meet the future Varick FDEs whose curiosity for the role emerged from reading this post :)

[Aaron Levie @levie]

Great post on FDEs. Everyone should read it if you’re interested in this job category. This is a job that is going to be around as long as AI keeps changing rapidly, which it inevitably will.

People often wonder why isn’t this like just deploying other forms of technology in the past, like cloud.

Because something like cloud adoption affected a fairly concentrated set of users (developers and IT), and generally didn’t require a fundamental change to the workflows of employees to get the benefits of the new service being delivered on the cloud. At best you went to one training session and you were done.

With agents, the work to implement them is not only highly technical, but they directly impact the underlying workflows that people participate in. This means there’s a ton of technical work and change management that comes with it.

Further, the pace of change of cloud wasn’t nearly as quick, so there was a lot more time for best practices to propagate. Now, every model change means either something new can be done that wasn’t possible before, or some piece of scaffolding is now redundant or holding you back.

This is why it’s commonly easier for a vendor or partner that’s seen the implementation hundreds or thousands of times help do the work, even with internal support from the customer.

So, this job isn’t going away any time soon, and will be a great path for a lot of technical talent, especially early career.

[Jonathan Lehr @fendien]

"At Vmax, that is the long-term ambition: AI systems that keep generating harder, broader, and more useful challenges for their own improvement. Moreover, these methods have the potential to not just match, but exceed humans by discovering radically new ways of performing work in the domains we train them in."

Amazing progress by the @VmaxAI team towards their huge vision

[Augustine Mavor-Parker @MavorParker]

This work represents @vmaxai 's first steps towards breaking through the human data bottleneck, and stabilizing self-play.

This work was led by @creus_roger; my DMs are open if you're interested in collaboration and we are hiring: https://t.co/wvjEAHhRrK

[Kelley Mak @kelleymak]

The Vmax team just released PopuLoRA, their latest research which uses asymmetric self-play between populations of teacher and student models, creating an adaptive training loop where the curriculum evolves with the models themselves.

Big step in the direction of self-improving AI. Congrats to the team!

[Jonathan Lehr @fendien]

@nrmehta Love this tactic!

[Jonathan Lehr @fendien]

@nikunj Such a fun game! Get a coast to coast 😎

[Jonathan Lehr @fendien]

"Bad speed is impatience that pretends to be ambition. It optimizes for novelty over depth and doesn’t let you sit long enough inside the unglamorous stretch where the real edge is formed. Because the truth is that the frontier rarely yields itself in the first six months"

Excellent read from @adityaag

[Aditya Agarwal @adityaag]

https://t.co/AoyPtuI9jx

[Jonathan Lehr @fendien]

@ProbyShandilya My first online experience was on Prodigy!