Live every tech company, developer first/infra companies took a beating in the private and public markets over the last couple of years, and unlike those benefitting directly from the AI tailwinds, it has taken more time for some of the old school 🦄 to bounce back. However, this week, what caught my eye is that two of these companies announced that they surpassed the $100M ARR mark and also raised new financing.

Just 2 years ago I posted this…

Since that post, Hashicorp dropped to a market cap of around $3.9B, changed their open source license due to competitive threats, and recently sold to IBM for $6.4B. Gitlab has traded back up and now has a market cap of almost $9B. Don’t get me wrong, these are all enormous outcomes for developer first and infrastructure startups but not the take-over-the-world values that they once commanded. Because of the success of these companies, many a VC threw 💰 at any open-source project with a bunch of stars leading to an incredibly overfunded dev tools sector. That being said, there are still some winners, quietly growing, and it’s refreshing to see both Vercel and Harness announce that they surpassed the $100M ARR mark and raised new capital!

Let’s look at Vercel first. Vercel is the leading frontend cloud platform and just announced a $250M Series E funding at a $3.25B valuation led by Accel, slightly up from its last round in 2021 at a $2.5B valuation. According to Accel, when it led its Series A four years ago, “300,000 developers used nextjs” and now “it’s used by over 1 million monthly active developers and automates the infrastructure for some of the world's most important companies.” It also crossed $100M ARR - simply incredible growth 📈.

“Vercel’s Frontend Cloud has been adopted by a growing number of enterprises globally to build more performant personalized web experiences”

Vercel has won over developers and enterprises by automating infrastructure, enabling companies like Under Armour, Unity, Sonos, and Nintendo to focus on the application layer and user experience. With strong momentum behind its Frontend Cloud platform, Vercel recently surpassed $100M in annualized revenue. Vercel's Next.js technology, with over 1 million monthly active developers, powers the world's largest web applications—positioning itself at the forefront of the Generative AI revolution.

This new funding will help Vercel deliver the combined power of its Frontend Cloud, AI, and enterprise-grade security to every developer, user, and company. With AI transforming the way developers build and deliver software, Vercel will invest in further building v0, its groundbreaking Generative UI product that makes product creation as simple as describing ideas via prompts. Additionally, the company will invest in its Vercel AI SDK, providing developers and teams with a comprehensive framework for building AI applications and products.

Next up is Harness, and while the round was a not a growth equity round, the company did announce that it crossed the $100M ARR mark last year as it also took a $150M line of credit to continue growth.

* Experienced rapid Annual Recurring Revenue (ARR) growth, going from $1M to more than $100M over the past five years.
* Customers executed more than 44 million code deployments with the Harness platform in 2023, more than double the number of deployments completed in 2022.
* Broadened and evolved product offerings by adding 2,800+ new features and enhancements across the Harness platform.
* Introduced AIDATM, a new generative AI assistant integrated directly into all aspects of the Software Delivery Lifecycle.

More here

BTW, I’d call both of these companies the new school incumbents, new enough to disrupt the prior generation of startups, but not born in the AI native world like all of the recent developer copilot fundings. Vercel was started in November of 2015 so it has taken over 9 years to cross $100M ARR - one has to be patient in dev tools and it doesn’t always work out!!! Harness started in 2017 so it too took 7 years to cross $100M ARR, but it also had a more focused model to sell to larger enterprises as a devops infra provider from the very beginning.

IMO, both companies would not be successful without also layering in their AI pixie dust into the product lifecycle - a must-have for any enterprise software company now. Vercel is continuing to invest in v0, “its groundbreaking Generative UI product that makes product creation as simple as describing ideas via prompts. Additionally, the company will invest in its Vercel AI SDK, providing developers and teams with a comprehensive framework for building AI applications and products.”

Harness just “introduced AIDATM, a new generative AI assistant integrated directly into all aspects of the Software Delivery Lifecycle.” The website is also updated with new messaging, the single Platform for all your Software Delivery needs empowering software engineering teams with AI-infused technology for seamless software delivery.

There will be a lot more pain to come for open source and dev tooling/infra companies in the next 12 months, but it’s awesome to also see real winners emerge crossing the hard to crack $100M ARR mark.

Vercel and Harness will both tell you, despite the rumors of it being hard to make money in dev tooling/infra, some old school incumbents sprinkled with some AI pixie dust can do just that.

As LL Cool J said, don’t call it a comeback, we been here for years.

Switching gears, the one area in dev tooling that has been on 🔥 has been any company that has to do with code generation. Replit definitely found its groove early, but while building buzz and acquiring lots of developers is super hard, it’s even tougher to build a long-term business with solid retention. To that end, Replit announced this past week a transition of its GTM team to enterprise and a 20% layoff to go along with it. Here’s Replit CEO Amjad Masad email to staffers.

Team,
Over the past year, we've been working to make Replit a platform that empowers anyone in a company to become a programmer. While the opportunity is clear, I've come to realize that to serve businesses, we need a different mix of roles.

As a result, we decided to part ways with 30 of our exceptional colleagues. You'll shortly receive an email to let you know if we no longer have a role for you...

I know this news may come as a surprise, given our strong financial position and the amount of work ahead. Usually, at this stage of a startup's life, the company dumps fuel on growth, hiring as many people as possible-not laying people off. At Replit, we've always done things differently and achieved much with a smaller team. Smaller teams can be more focused, enabling us to move faster while giving us the space we need to realize our potential.

And now is the time to move fast because the opportunity for Replit has never been clearer. Replit and Al make it possible for almost every knowledge worker to create software. Businesses want Replit, and users at more than half of the Fortune 500 are already experimenting with us. We now need to make it even easier to build great software, fast.
While the technological trends are in our favor, success is not guaranteed. We are building a new go-to-market team and will go after what makes us useful to people working at companies and helping those companies discover - and pay for - what we offer...

For those remaining, the road ahead will be challenging, but it will be a chance to do the best work of our lives. I'd like to thank you in advance for your effort. Later today, we'll meet to discuss this change. Tomorrow, instead of our usual Weekly Wins meeting, I'd be happy to take questions and discuss our plan to make an even greater Replit.
Thank you,
Amjad

While developer led motions are awesome, all roads eventually lead to steak 🥩 dinnahs and hard-core enterprise sales to pay the bills! Rising above the noise is tough and the definition of success with $15B+ outcomes may no longer exist, but that still doesn’t mean that those who can make the transition like Harness and Vercel will still create a ton of value for all.

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

Scaling Startups

#👇🏼👌🏼

#Founders and investors, do you cheer, fear or 🤔 with each new foundational model upgrade - 👇🏼 from Aaron Levie - this is the way

OpenAI just made their new GPT4 model 50% cheaper and 2X faster for developers. This is an insane level of improvement for anyone building in AI right now.

#🤣 how to calculate TAM and get later stage VCs excited

#How We Saved 10s of Thousands of Dollars Deploying Low Cost Open Source AI Technologies At Scale with Kubernetes - from OpenSauced (a portfolio co)

However, as soon as you need to scale up your generative AI operations, you'll quickly encounter a pretty significant obstacle: the cost. Once you start generating thousands (and eventually tens of thousands) of texts via GPT-4, or even the lower-cost GPT-3.5 models, you'll quickly find your OpenAI bill is also growing into the thousands of dollars every month.

Thankfully, for small and agile teams, there are a lot of great options out there for deploying low cost open source technologies to reproduce an OpenAI compatible API that uses the latest and greatest of the very solid open source models (which in many cases, rival the performance of the GPT 3.5 class of models)...

Enterprise Tech

#👇🏼 this is simply huge - democratizing agents for all. Agentic workflows are the future and huge congrats to João Moura, founder of CrewAI (a boldstart Inception investment), on this epic new course brought to you by Andrew Ng! No easier way to learn to build your own crew of AI agents working 24/7 for you. Here’s another tweet on why I’m so excited…

#🤯 our dystopian future is here…wait till those real time deep fake videos start proliferating…

#Worth a read from Alexander Wang, founder of Scale AI, on OpenAI and Google AI announcements

1/ Some thoughts on the recent OpenAI and Google announcements, and what it indicates about what's next in AI.

Hint: post-training is REALLY important...

THREAD here:

#🤣 ❤️ the sniping between OpenAI and Google - this is from a research engineer at Google Brain

#Super interesting discussion Ask HN: Disillusioned after AI?

Maybe it’s me but the advances in AI are just leaving me feeling disillusioned as a builder. There’s an odd feeling that whatever I am going to build will just get gobbled away by some big tech company. The demos just become more cringey, the messaging more duplicitous and fake-authentic. I’m wondering if anyone else has wrestled with similar feelings or maybe I am just going through a rough patch...

015a 3 days ago | parent | prev | next [–]

I think you bring up some good points.
One thing I'd add: If we enter a world where the ability to build some product, movie, song, novel, whatever is so widespread and cheap that anyone can do it (and I do emphasize the word "if", because I think its debatable that this will ever happen at a level that is cost-accessible):

The strongest way businesses will be able to differentiate is Taste. Human-ness; novelty, Design, things that AI isn't actually all that good at. AI may generate an episode of Law and Order, but it probably couldn't generate, say, Oppenheimer; because nothing in the world knew that it wanted what Oppenheimer delivered. If everything else goes to zero, then that's all that remains.

The issue is: AI is tremendously bad at working with Humans who actually have taste. You generate an AI image with the subject on the left; you ask for a revision, same image, but move the subject to the right side of the image; it changes too much, or generates new images entirely rather than iterating. Its not-obvious that this will get better because it is somewhat antithetical to the way the AI works...

Rest here:

#How Simo built Folk, the next gen CRM, into 100k users and 2k paying customers in 4 years - Congrats Simo, reader of What’s 🔥 - this is just so awesome to see!

I sat down with Simo to hear the behind-the-scenes story – folk’s four year journey from zero to 100,000 users. Here’s the TL;DR:

Attracted a waitlist of 10,000 people before writing a single line of code

Spent 18 months building before entering beta, and another six months before launching on Product Hunt

Doubled prices every six months to land on the right pricing strategy 

Turned to influencers and affiliates to stand out in the most competitive software market on the planet

Manually onboarded hundreds of users (Superhuman-style) before moving to self-serve

Doubled free-to-paid conversion after experimenting with a new user onboarding flow

More here from Kyle Poyar's Growth Unhinged

#A deep dive from The Information: How Lacework Went From Cybersecurity Stardom to Fire Sale Talks - the below is just one of the bangers in the article and another case in point - more 💰 does not solve problems, constraints breed creativity

By mid-2023, the company’s annual recurring revenue had barely grown over the prior two years, hovering below $100 million. Lacework’s effort to expand beyond its main customers—small and medium-sized businesses in healthcare, finance and other categories—to court large businesses mostly stalled. Throughout 2022 and 2023, it laid off roughly 300 employees, mostly salespeople. Some customers, including the Winklevoss twins’ Gemini crypto exchange, were ditching it for competitors. And Lacework was still incinerating cash: In all, it burned roughly $200 million in 2023 to achieve a measly $2 million in net new revenue...

The Lululemon giveaway was just a sample of Lacework’s efforts to win over customers. Around the same time as the Lululemon gathering, the startup authorized its salespeople to give away items like Nintendo Switches ($300), stainless-steel Solo Stoves ($400) and AirPod Pro earbuds ($250) to anyone who took a meeting with them, regardless of whether that person showed any interest in buying the startup’s security software, according to a former Lacework salesperson.

#The tech pendulum always swings from centralized to distributed - here’s Kelsey Hightower on AI

This big AI push is taking us back to the mainframe days. Only a few will have systems big enough to participate while the rest will be time-sharing.
----------

Open source software and open hardware have never been more important.

#Speed is magical - Mitchell Hashimoto, co-founder of Hashicorp

I've had "4166" written on a piece of paper on my desk for a couple months. Many will recognize the value as the number of microseconds I have per frame for a 240Hz refresh rate. A few months ago a full screen redraw took 36000 microseconds 🤮. Today, it takes around 1900us. 😍

#Phil Venables Google Cloud Security on “Securing theAI Software Supply Chain”

#The 50 most promising Israeli startups - 2024 (CTech) - quite a diverse list with lots of cool startups

We could have made things much easier for ourselves and built a list entirely composed of cyber companies, considering Israel's enormous relative advantage, and whose reputation in this area remains surprisingly unaffected by the security and technological failures of October 7. But Israel’s startup ecosystem extends far beyond cyber, and, in fact, the biggest Israeli exits mostly came from chip and hardware companies. Cyber should not be underestimated, especially not Israel’s unique capabilities, but we also want to foster companies in foodtech, cleantech, and AI, the latter of which was, from the investment world's perspective, the big revolution of 2023.

#😲 our wireless networks are compromised and being used to track and spy on us (@mattjay)

Kevin Briggs, a senior advisor at CISA, has publicly revealed ongoing vulnerabilities in U.S. telecom networks.

TL;DR - He has evidence vulns in teleco's are being used to track and spy on U.S. citizens.

Buckle up, here's what we know: 🧵

Markets

#25 different ways to calculate NDR??? (Mostly Metrics)

#Snowflake in talks to buy Reka, creator of LLMs, for over $1 Billion (Bloomberg)

Reka AI makes large language models — software trained on massive swaths of the internet that can be used for a range of tasks, such as captioning images or as a customer-supported chatbots. The startup was founded in 2022 by researchers from Alphabet Inc.’s Google and Meta Platforms Inc. It was valued at about $300 million in a 2023 investment round, which included funding from Snowflake’s venture arm, according to a Reuters report.

#from the Milken conference in LA last week (FT)

As snaking lines of enthusiasts queued to hear former US president Bill Clinton, Argentine president Javier Milei, and hedge fund tycoon Ken Griffin, would-be dealmakers barely looked up as they laid the groundwork for new mergers, acquisitions, buyouts and debt deals.

“The consensus from Milken is that 2025 is going to be a great year. The question is how much of that activity gets pulled forward in 2024,” said John Miller, executive co-chair of global investment banking at Jefferies. “Announced M&A is a leading indicator and it’s flashing green right now.”

#SIEM…Exabeam and LogRhythm merge in UEBA space and Palo Alto Networks buys QRadar from IBM (CNBC) and signs broader partnership - end of an era post Cisco acquisition of Splunk - from Forrester Research:

This makes IBM the second legacy SIEM player (the other being LogRhythm) this week to attach itself to a newer, more innovative vendor. These moves come on the heels of Cisco’s completed acquisition of Splunk. All legacy SIEM players are facing increasing competition from tech titans (aka hyperscalers) as well as extended detection and response (XDR) vendors that are aggressively positioning as SIEM alternatives.