Zero VC, $1B Revenue. AI Is Breaking the Startup Game.
Scale AI's biggest competitor is a company you've never heard of. And they make $1B in revenue every year. Also, a vibe-coded app sells for $80M to Wix after 6 months.
Scale AI’s biggest competitor is a company most people have never heard of.
It brought in $1B in revenue last year, surpassing Scale’s $850M.
The best part of all: Surge AI, an SF-based company whose data labelling platform serves model makers like Google, OpenAI and Anthropic, is entirely bootstrapped. It hasn’t raised a single dollar of venture capital.
AI isn’t just going to change how software is built. It will change venture as we know it today.
In the last 2 years since the arrival of chatGPT blew open the AI cycle, we’ve seen a dramatic shift in startup dynamics that’s upended traditional assumptions in venture.
All of a sudden:
$5M in ARR in a few months seems achievable
You can raise $0 in VC funding and hit hypergrowth scale
Fielding acquisition offers in days or months is entirely possible
In my view, there are a few distinct archetypes that have emerged in the last two years. I’ve been using this working taxonomy to break it down:
The Vibe Solos ($1 - 5M ARR)
Solo founders or ultra-lean teams building with agentic tools like Cursor or Lovable. Typically bootstrapped or lightly funded. Products show early breakout traction and are often acquired within 6 to 12 months for talent, distribution, or product love.
Looks like:
Base44, a prompt-to-code app acquired by Wix for $80M this week. It was generating $3.5M ARR, and reportedly grew to 250,000 users in 6 months. At this stage, the Solo startups tend to be high-growth, low-profitability: Base44 netted out at $189,000 in profit in May on its $3.5M ARR, after covering high LLM token costs.
The Fast-Climbers ($20M - $100M ARR)
Startups that have quickly captured the imagination and attention of the early adopters. Many developers consider them as core to the new software stack. Tends to be highly collaborative within the ecosystem. Fast-growing, but early revenue may not be an indication of long-term user retention as early-adopters can be flaky in loyalty.
Looks like:
Supabase ($20M ARR), Vercel ($20M ARR), Genspark ($36M ARR), Gamma ($50M ARR), Lovable ($75M ARR), Harvey ($75M ARR), Glean ($100M ARR)
*numbers are estimates from what I could find online.
** There are a few in-betweeners like Hugging Face ($130M ARR) and Windsurf ($100M ARR) who could be on their way up to the next tier.
The Lean Giants ($500M - $1B ARR)
Companies operating at scale with surprisingly small teams. Some are heavily venture-funded; others, like Surge, are entirely bootstrapped. They’ve locked in real enterprise traction and are not slowing down any time soon. Typically lean teams, with remarkably high revenue per employee.
Looks like:
Cursor ($500M ARR; ~50 person team; raised >$1B VC funding)
Surge ($1B ARR; ~100 person team; raised $0 VC funding)
The Anchors ($3 - $10B ARR)
These are the foundational players driving the AI cycle. They’re building the infrastructure like models, platforms, protocols, and products that others build on.
Looks like:
OpenAI ($10B ARR), Anthropic ($3B ARR), Snowflake ($3.8B ARR), Databricks ($2.8B ARR)
The Titans ($150B - $350B ARR)
Tech giants dominating capital, distribution, and compute. They’re integrating AI into everything, from chips to cloud to consumer.
Looks like:
Alphabet ($350B ARR), Apple ($400B ARR), Microsoft ($245B ARR), Meta ($170B ARR), Nvidia ($150B ARR).
The technology industry grew by about $5 trillion dollars over the past 2 decades, about 5-6x growth. Where do you think the next 20 years will net out?
Have a great weekend - ttyl.
The Download —
News that mattered this week
Midjourney launches its first AI video generation model, V1: Midjourney announced the launch of its much-anticipated AI video generation model, V1. Users can upload an image and V1 will produce a set of four five-second videos based on it.
OpenAI signed a $200M deal with the Defense Department: The US Department of Defense has awarded OpenAI a $200 million contract to develop "prototype frontier AI capabilities”. The contract is a pilot program and the first partnership in the new OpenAI for Government initiative, through which the company aims to put its AI tools in the hands of "public servants across the United States."
Groq became an official inference provider on opensource model platform Hugging Face: It also announced that it now supports Alibaba’s Qwen3 32B language model with its full 131,000-token context window, a technical capability it claims no other fast inference provider can match.
Nvidia nears record highs with $3 Trillion market cap: And now, investors are betting that Nvidia is the most important company in tech. Nvidia has seen its stock rally nearly 7% over the past month and gain more than 10% over the past year. The company reported Q1 2025 revenue of $44.1 billion, up 69% year-over-year, and earnings per share of 96 cents, beating analyst expectations.
Six-month-old, solo-owned vibe coder Base44 sells to Wix for $80M cash: Maor Shlomo sold his 6-month-old, bootstrapped, vibe-coding startup Base44 to Wix for $80 million in an all-cash deal. In its first six months, Base44 reportedly grew to 250,000 users, hitting 10,000 users within its first three weeks.
Zuck continues to build out his AI dream team. Next on his list seems to be AI investors and founders Daniel Gross and Nat Friedman, who run their own VC firm NFDG. Gross is also the CEO of Safe Superintelligence, which he co-founded with former OpenAI chief scientist Ilya Sutskever last year.
Double Click —
Links to reads we found interesting
Claude explores Confidential Inference via Trusted Virtual Machines. (Anthropic)
Penn State researchers have built the world’s first working CMOS computer entirely from atom-thin 2D materials and free of traditional silicon. (Science Daily)
Danish has deployed four uncrewed robotic sailboats (known as "Voyagers") for a three-month trial. Powered by wind and solar energy, these sea drones can operate autonomously for months at sea. (AP News)
An international team of astronomers has trained a neural network with millions of synthetic simulations and AI to learn more about black holes, revealing the one at the center of our Milky Way is spinning at nearly top speed. (Science Daily)
Anthropic's recent research reveals how AI systems can engage in covert sabotage while appearing to be compliant, raising significant ethical concerns (Anthropic)
The ‘OpenAI Files’, created by two nonprofit tech watchdog organizations, breaks down how Sam Altman’s company works (OpenAI Files)
MIT researchers find that over-relying on LLMs can restructure your brain (MIT)