Danny Postma makes $300,000 a month. No employees. No investors. No office. He works from Bali, mostly in shorts and a t-shirt, and runs one of the most profitable solo AI businesses on the internet.
His product, HeadshotPro, does something deceptively simple: it generates professional headshots from selfies using AI. You upload a handful of casual photos of yourself, wait a couple hours, and get 120+ studio-quality professional portraits. LinkedIn headshots, corporate team photos, creative headshots -- all generated by AI, all indistinguishable from what you'd get at a $300 photo studio.
I've been following Danny's journey since before HeadshotPro existed, and what makes this story worth studying isn't just the revenue number. It's the decisions behind it. Because Danny didn't stumble into this. He engineered it. And the playbook he used is something any developer can replicate.
The Problem That Turned Out to Be Worth Millions
Here's something most people don't think about: almost everyone needs a professional headshot, and almost no one has a good one.
Think about the last time you updated your LinkedIn photo. If you're like most professionals, your profile picture is either a cropped group photo from someone's wedding, a selfie with weird lighting, or a headshot from 2019 that doesn't look like you anymore.
Getting a proper headshot the traditional way is a surprisingly painful process. You have to find a photographer, which means reading reviews, comparing portfolios, and sending inquiry emails. Then you schedule a session -- typically 2-3 weeks out because good photographers are booked. You drive to a studio, spend an hour being positioned and photographed, then wait 5-7 days for edited photos. Total cost: $150-500 for an individual session, $2,000-10,000 for a corporate team shoot.
For that price and hassle, most people just... don't bother. They keep the bad photo. They know it's hurting their professional image, but the friction is too high.
Danny saw this gap and realized AI could collapse the entire process -- the scheduling, the travel, the photographer, the editing -- into a 2-hour automated workflow that costs the customer $29-59 instead of hundreds.
How HeadshotPro Actually Works (and What It Costs to Run)
The technical process is interesting because it reveals just how good the economics are.
When a customer signs up, they upload 8-12 casual selfies. These don't need to be good photos. Bathroom mirror selfies, random snapshots from your camera roll, whatever. The AI just needs multiple angles of your face.
Behind the scenes, HeadshotPro trains a custom AI model on those specific photos. This process takes about 30 minutes and uses fine-tuning techniques on image generation models (Stable Diffusion, and more recently, Flux-based architectures). The model learns what you look like -- your facial features, skin tone, hair -- so it can generate new images of you in settings you've never been photographed in.
Then the system generates 120+ images. Different backgrounds (office, outdoor, studio), different outfits (suits, business casual, creative), different angles and lighting setups. The customer picks their favorites and downloads them in high resolution.
Now here's where the money part gets interesting. The compute cost to train a custom model and generate 120 images is roughly $2-5 per customer, depending on the infrastructure setup and model used. The customer pays $29 for the basic package or $59 for the premium.
Run those margins: at $29 revenue and $3 average compute cost, gross margin is about 90%. At $59 with the same compute cost, it's 95%. Danny is reportedly doing $300K/month in revenue, which means roughly $270K/month in gross profit.
For a solo operator. Let that number sit with you for a second.
Why Headshots Specifically? Danny's Selection Framework
Danny Postma didn't wake up one day and think "headshots." He's a serial builder. Before HeadshotPro, he created a string of AI products -- some successful, some not. He built AI art generators, AI interior design tools, and several smaller experiments. What he was really doing was testing a thesis.
The thesis was this: find an expensive human service, replace it with AI, charge 10% of the original price, and serve 100x more customers. The revenue math works because you're converting a small, high-ticket market into a massive, low-ticket market. It's the single most repeatable playbook in solo AI right now — I've catalogued a dozen more openings that fit it exactly in AI SaaS ideas for solo developers in 2026.
Not every service fits this pattern. Danny reportedly evaluated dozens of options before landing on headshots, and the selection criteria were specific.
The human service had to be universally needed. Headshots check that box -- every professional needs one. The human service had to be inconvenient. Photo studios require scheduling, traveling, and waiting. The AI replacement had to be indistinguishable from the original. This is crucial. If AI headshots looked obviously fake, nobody would use them for LinkedIn or corporate websites. By late 2023 and into 2024, AI image quality crossed the threshold where generated headshots genuinely looked professional.
The purchase had to be a one-time decision, not a recurring subscription. People need headshots occasionally -- when they change jobs, update their profile, join a new company. A $29-59 one-time purchase is an impulse buy. You don't need to convince anyone of ongoing value.
And the market had to be massive enough that even a small capture rate generates serious revenue. There are roughly 900 million LinkedIn users worldwide. If even 0.1% of them buy an AI headshot in a given year, that's 900,000 customers.
The B2B Play That Most People Miss
Individual headshots at $29-59 are the visible part of HeadshotPro's business. But the real scale engine is B2B.
Companies need headshots for their entire team. New hires, website updates, annual refreshes, team pages. A company with 100 employees paying a photographer for team headshots is looking at $5,000-15,000 plus the logistics nightmare of scheduling everyone.
HeadshotPro offers team packages where each employee uploads selfies on their own time, and the company gets consistent, professionally styled headshots for the entire team. One corporate deal can be worth $5,000-50,000 depending on team size and package tier.
This is where Danny's revenue really accelerated. The B2B flywheel works like this: an HR manager or marketing director discovers HeadshotPro because they personally used it for their own headshot. They think "wait, we could use this for the whole company." They buy a team package. Once the team photos are on the company website, other companies see the quality and ask "who did your team photos?"
The viral loop is built into the product. Every headshot that goes on a LinkedIn profile or company website is a free advertisement. People compliment the photo, ask where they got it done, and the answer is a direct referral to HeadshotPro.
Why Most Copycats Fail (and What They Get Wrong)
After HeadshotPro's success became public, dozens of copycats appeared. AI headshot generators are everywhere now. Most of them make very little money. Here's why.
The obvious reason is that they're late. HeadshotPro built brand recognition and trust early. When someone Googles "AI headshots," HeadshotPro ranks #1 organically and has thousands of reviews and testimonials. Copycats are fighting for scraps of a market that already has a dominant player.
But the less obvious reason is quality and consistency. Generating AI headshots that look professional is harder than it appears. The common failure modes are: skin texture that looks plasticky, eyes that are slightly off, lighting that doesn't match the background, and clothing that has weird artifacts.
Danny invested heavily in training data curation, model fine-tuning, and post-processing pipelines. The difference between a "pretty good" AI headshot and a "holy crap I can't tell this isn't a real photo" headshot is the difference between a product people recommend and a product people abandon.
Most copycats use off-the-shelf models with minimal customization. They compete on price -- "AI headshots for $9!" -- and deliver quality that makes people feel ripped off. They churn through customers once and never see them again, while HeadshotPro builds a brand that companies trust with their corporate identity.
The Economics of Solo AI Businesses in 2026
HeadshotPro is the highest-profile example of a pattern that's repeating across dozens of niches. The economics are specific and worth understanding.
API costs for AI inference have dropped roughly 90% since 2023. What cost $50 in compute to generate a set of images in early 2023 costs $2-5 today. This margin expansion means that solo founders can capture dramatically more profit per customer than they could even two years ago — and it's the same force making the one-person company a genuinely viable business model in 2026.
Infrastructure costs are similarly minimal. Next.js on Vercel, Supabase for the database, Stripe for payments, a queue system for processing -- the entire stack runs for a few hundred dollars a month until you're doing serious volume, at which point it's a rounding error against revenue.
The customer acquisition math is also favorable for this specific type of product. HeadshotPro gets a massive amount of organic traffic because "AI headshots" is a high-intent search term. People searching for it want to buy, not browse. Organic SEO plus word-of-mouth referrals plus the viral nature of the product itself create a customer acquisition machine that costs close to zero per customer.
Compare that to a typical SaaS business spending 30-50% of revenue on customer acquisition, and you start to see why HeadshotPro's margins are so extreme.
What You Can Actually Learn from This
Let me be direct about what's transferable here and what isn't.
What's transferable is the framework: expensive human service, replaced by AI, at a fraction of the price, serving a much larger market. This pattern works across many categories.
Think about product photography. E-commerce brands pay $50-200 per product image. AI product photography tools like Pebblely are already proving the model works. Logo design, resume writing, voice acting, interior design visualization, real estate virtual staging -- each of these follows the same pattern of high-cost human service that AI can replicate at 10-20% of the price. If you want the fastest possible test of one of these, the build-an-AI-SaaS-in-7-days approach is exactly how you'd validate one this month.
What's NOT easily transferable is the specific timing and market dynamics. Danny shipped HeadshotPro at exactly the right moment -- when AI image quality crossed the "good enough" threshold but before the market was saturated. The headshot market specifically was large, underserved, and perfectly suited to AI replacement.
If you try to launch an AI headshot product today, you're competing against an entrenched leader with years of brand equity. That doesn't mean you can't succeed, but the playbook changes from "be first" to "be 10x better in a specific niche" -- like AI headshots specifically for real estate agents, or specifically for medical professionals, or specifically for actors and models.
The bigger lesson is about speed and focus. Danny didn't spend six months building a perfect product. He shipped fast, learned from real customers, and iterated aggressively. He didn't diversify into ten different AI products. He went all-in on headshots -- one product, one market, relentless execution.
$300K/month. Solo. From Bali. It sounds like a fantasy, but the financial mechanics are straightforward: 90%+ gross margins on a product with near-zero customer acquisition cost, serving a market of hundreds of millions of potential buyers. The APIs exist. The infrastructure is commodity. The demand is real and growing.
The only question is whether you're willing to pick a specific niche, build the product, and commit to making it excellent.
---
Keep Reading
If the solo AI business model resonates, these go deeper into different angles:
- 10 AI SaaS Ideas That Actually Make Money in 2026 -- More opportunities following the same "replace expensive human service with AI" pattern
- How to Make Money with Claude AI -- Practical approaches to building revenue with AI tools as a solo developer
- AI Startup Ideas Worth Building in 2026 -- Larger-scale versions of the same playbook, for when you're ready to think bigger
- How AI Is Killing Middle-Class Jobs (and Creating New Ones) -- The macro view of what happens when AI replaces professional services at scale
