Where the money is, Oct 2026
Building got cheaper. The hard part is picking a customer who already has the problem. Here's what still looks like upside.
Earlier this month I wrote that I'm looking for a job someone will still pay for after a $20 chat can do the boring version. This is the next question. If that's true, where is the money?
I don't believe the answer is another general app. I do believe it starts with a real customer who already has the problem.
The short version:
- Start with low cost and high upside. One industry. One fixed offer. One niche you can talk into.
- Energy, trades, and health grew fast this past year. The reasons behind them look durable. They cost more to enter, so they come later.
- People are buying less often and spending more when they do. Trust beats cheap volume. Results beat hours.
- Skip the traps. Dropshipping courses, "AI agency" courses, crypto schemes, and apps nobody asked for.
Start here
One industry
Software for one specific industry is almost free to build now. It can sell for a long time, and it can grow without you sitting in every deal. The chat made the boring general tool cheap. The industry tool still has to know how that business actually works.
Goldman Sachs published a survey of small business owners in March. Most of them already use AI. Only about one in seven has it running through the core of the business. That gap is the product. They don't need another chat. They need the job done inside their industry.
A fixed offer
Productized services beat custom hourly work. You sell one thing at one price. "AI front desk for dentists, $X a month." The buyer knows what they get. You know what you deliver. You can improve the same offer over and over instead of reinventing every week.
Media and community
A newsletter, a channel, or a community in a niche you already know is not the whole business. It's how the next thing finds buyers. When you know who reads you, you stop guessing who to sell to.
Worth more later
Some markets ask for more money up front. I still think they matter, because the trends behind them are strong.
Energy and electricity
Abrigo looked at company sales data for the year ending in March. Electric power was among the fastest-growing U.S. industries, at about 20% median sales growth. The median across all industries was about 4%. That is not a tip. That is demand. AI data centers are one reason people keep talking about power, but Abrigo measured the growth, not the cause.
Trades and building services
Building-equipment contractors and other trades are growing and often hard to staff. One real path is buying a small, profitable trades business, often with an SBA loan, and running it better with AI. It's boring. It throws off cash. It's underrated next to another SaaS pitch.
Health and aging
An aging population keeps money moving into healthcare, wellness, and care. Offices of other health practitioners also grew strongly in the same 12 months. People get older every year. That demand doesn't need a hype cycle.
How people buy
Fewer, better purchases
Across about 2,300 brands tracked by impact.com in the first half of 2026, transactions fell about 7% while average order value rose about 16%. People bought less often and spent more when they did. Cheap volume plays are getting harder. Quality and trust are getting easier to charge for.
Pay for results
On impact.com's platform, brands put about 90% of their partner marketing spend into deals tied to performance. Pricing on results instead of hours makes you easier to buy. It also forces you to pick a job where the result is clear.
Traps to avoid
Dropshipping and "AI agency" courses. Crypto get-rich schemes. Apps nobody asked for. Every business above starts with a real customer who already has the problem. If you can't name that person, you're building the boring layer again.
What I'd do
If I were starting this month, I'd pick one industry. For me that's aviation, because I already care about it. I'd talk to ten businesses in it and find the one task they all hate. That task is the first product.
I could be wrong about which industry wins. I'm more sure about the order. Talk to the customer before you build the tool.