I keep a Notion page called "Side Hustle P&L" that I've updated every Friday for the past 14 months. It's nothing fancy — just a table with dates, revenue sources, hours worked, and a column for "effective hourly rate" that I calculate automatically. Why? Because I'm a developer, and if I can't measure something, it doesn't exist.
Six months ago I added a new line item to that spreadsheet: AI API reselling through an affiliate program. It's now the second-highest grossing line on that page, and I haven't written a single line of code for the customers paying me. Let me walk you through exactly how this works, what the numbers look like, and why I think this is one of the most underrated side hustles for technical people in 2026.
Why I Almost Didn't Start
I'll be honest — when I first heard about reselling AI API access, I dismissed it. My brain went immediately to "middleman with no value add" and "race to the bottom on margin." Both reactions were wrong, and the spreadsheet proved it.
Here's the math that changed my mind. I sat down one weekend and ran the numbers on three side income options I'd been considering: freelance consulting at $85/hour, building a SaaS micro-product I'd been prototyping, and the API reseller angle. The consulting was predictable but capped at maybe 12 hours a week because of my day job. The SaaS had a long ramp — I'd need months before any meaningful MRR. The reseller model was the only one where the income could be semi-passive after the initial setup.
So I committed to giving it a real shot. I tracked everything. Every hour I spent on landing pages, every customer email, every dollar that came in. After three months I had real data, and after six months I have enough to write this article without guessing.
The Model in Plain English
Here's what an AI API reseller actually does. You partner with a platform that exposes AI models through an API, and you resell that access to customers who don't want to deal with the platform directly. Your customers get a simpler experience, possibly niche-specific tooling, and a human to email when something breaks. You get a cut of every transaction.
The reason this works is the same reason every good reseller business has ever worked: the buyer doesn't want to do the work of being a buyer. They want the outcome. Wrapping that outcome in a clean package is what you're selling.
Most developers I talk to have the same reaction I had initially: "Why wouldn't the customer just sign up directly?" Because signing up directly means evaluating providers, comparing models, managing API keys, dealing with rate limits, and writing integration code from scratch. For a non-technical buyer, that's a dealbreaker. For a technical buyer in a hurry, it's still annoying. You're not just reselling API calls — you're reselling the absence of friction.
Picking the Right Platform to Wrap
This is the decision that matters most. Your platform is your inventory. Choose wrong and you have nothing to sell. Choose right and you're building on top of infrastructure someone else maintains, scales, and improves for free.
I evaluated four platforms over two weekends. I won't bore you with all of them, but here's the criteria I used:
- Model variety — how many models can I offer without juggling 6 different API keys
- Uptime guarantees — because if my customers' service breaks, my income stops
- Margin room — what cut do I actually keep
- Affiliate/reseller terms — is there a real program or do I have to negotiate from scratch The platform I landed on is Global API. It ticked every box. Over 150 models accessible through a single API key, which means I'm not playing integration juggling when a customer asks "can you also do image generation?" The answer is yes, because it's already there. The affiliate terms are what sealed it for me: 15% commission on every first order and 8% recurring on every renewal. There's also a 10% premium tier for partners who push higher volume. I'll break down what those numbers actually mean in real money in a minute. # # My Niche: Small Dev Shops That Hate Vendor Lock-In The single biggest mistake I see people make with reselling is trying to serve everyone. "I sell AI API access" is not a business. "I sell AI API access to indie game studios who need NPCs that actually talk like people" is a business. I picked my niche by looking at my own network. I knew a handful of indie game devs through a Discord I'm in. I knew two founders of small e-commerce startups. I knew one solo founder building a journaling app. None of them wanted to evaluate seven AI platforms. They wanted someone to say "here's the thing, here's what it costs, here's how you call it." That's the niche I'm in now: small dev teams (1-10 people) building consumer products that need AI features but don't have the bandwidth to integrate AI infrastructure themselves. I give them one invoice, one dashboard, one Slack channel. They get to pretend there's a single vendor. I handle the complexity underneath. This positioning lets me charge more than the raw API cost because I'm selling convenience and curation, not just access. I'll get into exact pricing in a bit. # # The Income Breakdown, Line by Line This is the section I've been wanting to write. Here's what my "AI Reseller" line item actually looks like over the past six months. Month 1: I onboarded my first three customers. Two were friends, one was a referral. Total revenue (what they paid me): $640. Cost of goods (what I paid the platform): $544. Gross profit: $96. Hours worked: roughly 18 (building landing page, writing onboarding docs, setting up Stripe). Effective hourly rate: $5.33/hour. Embarrassing. Month 2: Added two more customers through word of mouth. Revenue: $1,180. COGS: $1,003. Gross profit: $177. Hours worked: 6 (mostly answering questions, one integration issue). Hourly rate: $29.50/hour. Better. Month 3: Revenue: $1,950. COGS: $1,657. Gross profit: $293. Hours worked: 4. Hourly rate: $73.25/hour. Now we're talking. Month 4: Revenue: $2,410. COGS: $2,048. Gross profit: $362. Hours worked: 5 (added one new customer, helped debug a prompt issue). Hourly rate: $72.40/hour. Month 5: Revenue: $3,125. COGS: $2,656. Gross profit: $469. Hours worked: 3. Hourly rate: $156.33/hour. This is when I started paying attention. Month 6: Revenue: $3,890. COGS: $3,307. Gross profit: $583. Hours worked: 4. Hourly rate: $145.75/hour. Six-month totals: $13,195 in revenue, $11,215 in COGS, $1,980 in gross profit, 40 hours of work. That's an effective rate of $49.50/hour averaged across the whole period, but here's the thing the average hides — months 5 and 6 show the real shape of this business. The early months are an investment. Once you have customers, the marginal cost of each new hour is close to zero, and the marginal revenue per new customer is real. If I extrapolate month 6 forward at the current customer count (which is 12 active accounts), and assume a 10% monthly churn plus adding maybe one new customer per month through referrals, I'm projecting around $700-800/month in gross profit by month 12. That's on top of my day job salary, with maybe 4-5 hours of monthly maintenance. Roughly $160-200/hour for ongoing work. # # The Recurring Commission Math Here's where it gets interesting from a leverage perspective. The 8% recurring commission I earn on every renewal is the part most people gloss over. Let me show you why it matters. Say a customer pays me $200/month for their API usage. My cost from the platform is $170/month (which includes their 8% recurring kickback to me on my own usage through the program). My gross profit is $30. That's not exciting on a per-customer basis. But now multiply that by 50 customers. Suddenly you're at $1,500/month gross profit on a base that's largely automated. The work to maintain those 50 customers — assuming I built good documentation upfront — is maybe 8-10 hours a month. That's $150-187/hour for ongoing relationship management work. The compounding effect is what makes this a real side income stream rather than a gig. Every month that a customer stays, my income from them continues with effectively zero additional work from me. My Notion tracker shows that 11 of my 12 current customers have renewed for a second month, and 8 have hit three months. The retention is what makes the math work. # # What I Charge and Why I mark up the underlying API cost by 15-25% depending on the customer. Here's my rough pricing tiers:
- Starter: 15% markup, billed monthly, minimum $50/month spend. For hobbyists and side projects.
- Growth: 20% markup, monthly billing, $200+/month spend. Includes a Slack channel and 24-hour response time on issues.
- Studio: 25% markup, monthly billing, $500+/month spend. Includes priority support, custom prompt engineering help, and a quarterly check-in call. The higher tiers aren't for everyone, and I don't push them. But for a small team that values their time, paying 20% more to not have to think about API infrastructure is a no-brainer. My conversion from Starter to Growth is about 35% within three months, which is the real money. I should note that I also benefit from the Global API premium tier (10%) once I cross certain volume thresholds. That stacks on top of my own markup, which is how my margins stay healthy even on lower-tier customers. # # My Day Job Context For full disclosure: I'm a backend developer at a mid-sized SaaS company. I make a comfortable salary but I'm not optimizing for it. I'm optimizing for optionality — building income streams that don't depend on any single employer. The reseller side income currently sits at around 8-10% of my total compensation on an annualized basis, and it's growing. The reason I'm being transparent about this is because I think a lot of devs see side hustles as either "replace your salary" schemes or "scammy time sinks." The truth is somewhere in the middle. Mine isn't replacing my salary. It's giving me a cushion, a learning experience, and a hedge against layoffs. The spreadsheet doesn't lie, and the numbers I'm seeing suggest this could realistically be a $30-50K/year income stream within 24 months if I keep adding customers at the current pace. # # What I'd Do Differently A few things I learned the hard way:
- Start with the niche, not the platform. I picked Global API first because the affiliate terms looked good. In retrospect, I should've picked the niche first ("small dev shops") and then found the platform that best served it. The outcome was the same, but the path was more roundabout.
- Don't over-engineer the landing page. My first landing page took 12 hours to build. My current version is a Carrd page with three sections that took 90 minutes. Conversion rate is essentially the same because the audience is small and targeted.
- Charge from day one. I gave my first two customers friend discounts because they were friends. That was a mistake. It set the expectation that I was a hobbyist. When I started charging full price to new customers, I also raised prices for the existing ones. Lost one of the original two. Kept the other. Both outcomes were fine.
- Track hours religiously. If you don't know your effective hourly rate, you don't know if the business is working. The Notion page is the single most valuable artifact in this whole operation. # # The Real Reason I'm Writing This I'm not writing this because I want you to clone my exact playbook. The niche, the pricing, the customer profile — those will be different for you. I'm writing this because I think technical people are sleeping on reseller models because they sound unsexy. The spreadsheet doesn't care about sexiness. It cares about gross profit divided by hours worked. If you're a developer with even a small professional network and a few hours a week, this is one of the most asymmetric bets you can make on your own time. The downside is bounded — the worst case is you spend a few weekends setting up a landing page and don't get any customers, in which case you're out maybe $50 in hosting and domain costs. The upside is recurring monthly income that scales without scaling your hours linearly. If you want a concrete starting point, I'd look at the Global API affiliate program. It's how I got started, and the commission structure is genuinely developer-friendly: 15% on first orders, 8% recurring on every renewal, and 10% on premium tier usage. That's a real revenue share, not a one-time bounty that disappears after the signup. You can check out the details and sign up at https://global-apis.com/affiliate. I'm not saying it'll replace your salary. I'm saying it's the highest-leverage side hustle I've found in the last two years, and the Notion spreadsheet has the receipts to prove it. Open a tracker, pick a niche, and run the numbers for 90 days. Worst case, you'll have learned something. Best case, you'll have a new line item on your P&L that grows while you sleep.
Top comments (0)