What’s 🔥 in Enterprise IT/VC #432

Only the Paranoid Survive - all moats are temporary until they aren't - OpenAI coming for you

I’m in my 29th year as a venture capitalist, investing at the first check stage, now called the Inception round, in technical founders in the enterprise sector. In all my time, I have never seen our industry move so fast and with such conviction before the next big thing comes and goes. Just recently, all the talk in the newsletter was about DeepSeek, and now we’ve already shifted to Mistral’s Le Chat, Gemini, and whether the increasing CapX is all worth it. It's crucial in these times not to react to everything, but to think deeper and longer term while staying tuned in to what’s happening. Writing this every Saturday for you gives me that space, and I hope it does the same for you.

One big discussion this past week was around the idea of moats in AI. What are defensible moats, how defensible are they, and well, are there really moats. Here’s where I come out on them:

All moats are nice but only temporary. All software can and will be copied with time and money, just depends how long it takes. Sure, some moats like your underlying data or network effects and community may be stronger, but constantly being paranoid and continuing to innovate with core R&D is one of only ways to be first and stay ahead.

Here’s my commentary:

Regarding the last mile, there’s a reason why OpenAI is hiring for Forward Deployed Engineers as it knows that solving problems and building solutions will lead to large enterprise deals and feed knowledge back to the mother ship to be productized later. Here’s the rec:

Diving deeper into the job rec, OpenAI confirms that this FDE will spend at least 50% of time on-site with customers and former founder experience is super helpful!

Are willing to travel up to 50% and work on-site with customers to build strong relationships and deeply understand their needs.

Former founder, or early engineer at a startup who has built a product from scratch is a plus

This is 100% the same playbook from Palantir’s early days - I’ve written extensively about the Palantir model from FDE to BLG, Bootcamp Led Growth which you can check out here.

It’s also working in a huge way! Here’s Palantir 2 quarters later 🤯

You can read the entire Palantir earnings presentation here.

Now back to moats in AI, Semil Shah also sparked a healthy discussion on this which you should check out 👇🏼

OpenAI understands that the foundational model layer is commoditizing rapidly which is why it continues to keep shipping new models to stay ahead but also move up the stack into the application layer and build their own applications. All of those Forward Deployed Engineers building custom applications in the enterprise will also help tweak the product and platform to build even faster. Founders, just remember, stay paranoid as OpenAI will be coming for you - no one in the app layer is immune, if it works and it's big, then prepare for battle.

So what’s a founder to do if those moats that everyone talked about are only temporary? Just keep running fast 🏃🏼‍♂️ and hope that focus, learnings from building, and continuing to innovate will keep you far enough ahead of the competition to grow your customer base.

Also remember that your current success can also vaporize tomorrow - as easy as it is to have great “vibes” early, it can also just as easily disappear tomorrow.

Here’s one new way to create that early momentum - building in public from the very beginning. Multitime founder Jeff Haynie (Appcelerator, Pinpoint Software) and now Agentuity (a boldstart portfolio co) is going to post every week or two on the progress of the company as it builds the Heroku for AI agents - check it out here.

And wow, talk about shipping code:

It’s this mentality from Jeff and also👇🏼 that keeps the best technical founders from Inception and scale always ahead.

As always, 🙏🏼 for reading and please share with your friends and colleagues.

---

Scaling Startups

#calling all founders in the Boston area - this is going to be 🔥 from my boldstart partner Ellen Chisa - sign up form in comments and due 2/14

#Peter Walker with another data banger - image from Peter’s LI post here and commentary on my post below

#💯

#for VCs and LPs (investors in VC funds)

---

Enterprise Tech

#Glean - enterprise AI search is real, crossed the $100M ARR mark in 3 years 😲 - full thread on LinkedIn if you click graphic

and that last mile…need to get so much right like…

🔹 Built for the enterprise from day 1: Glean has been designed for the enterprise since its inception, handling the tough challenges of security, permissions, compliance, reliability, and scale.

#later in week - got to update that $100M chart above and now add Cursor who did it 1 year!!

#not to be outdone for the week and speaking of vibes, Mistral’s Le Chat is here - over 5M views

#what’s 🔥 for AI use cases in the enterprise

#this is pretty darn good - idea to app in minutes from an iOS app

#the new dev pipeline to build applications - from research to spec to build

#on OpenAI deep research

#💯 all about the vibes

#AI is working, and in many cases, better than humans

#🤔

#confirmed - more agents deliver less hallucinations…

#I must say that my colleague Ron Miller, formerly TechCrunch, has just been crushing it with our new boldstart publication, FastForward - if you’re interested in hearing from the leading IT minds in the largest enterprises, then please subscribe here. Not only do we have leaders like Belinda but also the CIO of Salesforce, CTO of Intuit, SVP Consumer Engineering at Paypal…all sharing their priorities and how they partner with startups.

#we live in a multimodel world and making these AI systems work for more complex workflows requires expertise

#🐶 years for AI - LFG!

---

Markets

#😲

#big fundings - Island secure browser to raise at $4.5B and OpenAI at $300B post while Safe Superintelligence founded by Ilya, formerly OpenAI, to raise at 4x last round at $20B valuation despite zero revenue…

#when is that moment for AI, don’t get me wrong, massive upside ahead but we will experience a series of mini-bubbles on way to 📈