In Q2, Lazarus grew 340% QoQ, added +$1.7M in net new ARR, and expanded across manufacturing, retail, and professional services.
Customers are using 95% fewer AI tokens, cutting SaaS costs by up to 90%, and speeding up decisions from days to minutes.
The future of enterprise AI is not model lock-in.
It is data ownership, application ownership, and the freedom to keep moving.
Proud of the Lazarus team.
More to come!
@openlazarusai
RAISE the STAKES Top 100 Applicants ANNOUNCED 🚀
+1000 Founders applied, and these top 100 represent the builders, the founders, and the teams turning frontier ideas into companies that ship.
From healthcare to security, agentic AI to climate, quantum to consumer. Every one of them earned their spot.
Our top 100 Startups have collectively raised €110M+ through pre-seed, Seed and Series A.
Congratulations to all 100. We'll be tagging every team below.
Tune in for 24th of July for our top 10 announcement within the tracks of @backblaze@GoogleStartups@GoogleDeepMind
Let’s see who made it to the top 100 👀
#RAISESummit #RAISEtheSTAKES
@Kirkland_Ellis, the highest-grossing law firm in the world, is spending $500m to build their own AI platform. Read that twice. A firm with unlimited budget and unlimited access to every vendor on earth looked at the SaaS market and decided none of it was good enough.
For twenty years, the rational move was to buy. No law firm, no matter how big, would have considered building, because real software required armies of engineers, years of work, and know-how only Silicon Valley had. Everyone outside Silicon Valley couldn't replicate any of it, so they rented.
That era is over. AI changed what one engineer can do. As a 2x founder who codes every day, I can tell you building software is no longer hard. The output of a single engineer has gone up 50x. Silicon Valley doesn't deserve your dollars anymore.
BUY LESS. BUILD MORE. OWN YOUR FUTURE.
And no, you don't need half a billion dollars...
If you want to build but don't know where to start, @openlazarusai is an open-source codebase that gives businesses the foundation to build their own OS.
Lazarus has been invited to the Leaders In AI Summit NYC '26, @InstituteForAi
If you're attending and thinking about reducing SaaS costs through AI — we're offering free sessions to help companies map their platforms and identify where labor inefficiencies live. You walk away with a deliverable whether we work together or not.
For us it's valuable market research as we build our open-source AI agent platform alongside our consulting work.
Come say hi !
https://t.co/KMfN6Tq8jR
Something we've been working toward — Lazarus is going fully open source.
The project will live at https://t.co/7vn1QczyNA and everything will be available on GitHub on May 1st. If you've ever wanted to peek under the hood, contribute, or just follow along, now's your chance.
https://t.co/PsHMmrNEfv
@garrytan careful garry this take got us rejected from yc twice. saas data model ≠ agent economy but yc wanted "saas ai agent for [vertical]." timing is everything though. see you next batch
@garrytan careful garry this take got us rejected from yc twice. saas data model ≠ agent economy but yc wanted "saas ai agent for [vertical]." timing is everything though. see you next batch
We're excited to announce that Lazarus is now part of Mercury Startup Perks.
If you're a Mercury customer, you can activate your perk directly from your dashboard or at https://t.co/gK9YHWbwcb
The hardest part of building an AI agent on Lazarus is deciding which workflow to automate first.
Your first agent is free ツ. No credit card required.
Start building → link in bio
There are two dominant narratives about the future of software:
1. “AI will kill SaaS because every company will build its own tools.”
2. “AI won’t change anything because enterprises won’t maintain what they build.”
Both miss the real shift.
AI doesn’t turn organizations into full software houses.
It simply makes enterprise-tailored software cheap enough to build — the long-tail workflows that never get done because they’re too niche, too specific, or “not worth the budget.”
Those finally become viable.
Hyper-common workflows will still live in SaaS, maintained by the specialists who should maintain them.
But they won’t be used through ten different disconnected vendor UIs anymore.
They’ll be consumed through enterprise-tailored frontends — frontends shaped around how the organization actually works.
Meanwhile, all of the odd, idiosyncratic, business-critical processes that no external vendor will ever productize?
AI pushes those into the realm of cheap, internal feasibility.
And once enterprises start accumulating both:
• their own enterprise-tailored workflows
• and SaaS workflows piped into the same interface
…something bigger emerges:
a true command-and-control layer — what intranets promised for 20 years and never delivered.
A single operational surface over everything the organization runs: internal systems, external SaaS, workflows, and data.
That’s the real productivity unlock.
It simply wasn’t economically possible until now.
The IRONMAN suit for LLMs already exists. it's Linux.
Thousands of hackers built it over 30 years. Runs 90% of the cloud. Every tool an agent needs is already there.
Why are we reinventing this? MCPs could've been CLIs. Skills are just scripts and files. Give agents Linux.
We're bloating context windows prompting LLMs to learn new abstractions every run when their native language is already computer.
I get anthropic's commercial play - create tools to own the community. but two years from now half these tools are abandoned repos.
if you're building longterm agentic workforces check out @thinklazarus.
Every agent would still be an island though. Query Clay, copy-paste, query next tool, repeat.
Lazarus is Dropbox for AI agents, they pass files to each other directly. Everything builds on everything else, fixing this problem.
We applied for W26, would love to show you.
https://t.co/03pZdsFaWa
PS. Intro video pinned to profile