For a few months I've been reading the market for buying and selling online businesses. Not to buy anything. Just out of curiosity, to understand which digital models actually hold up. Listings and broker reports are a good X-ray of the internet economy, and one pattern keeps showing up: the market is active but picky.
Buyers pay premiums for recurring revenue, profitability, low owner dependence, clean records, and defensible customer acquisition. Weak or speculative businesses get discounts and earn-outs, where the seller only gets paid if the business survives. That split, premium versus earn-out, tells you most of what the market rewards.
How the categories stack up
Here's the map I've built from listings, broker reports, and my own reading of where each type is heading.
| Business type | Market / trend | Potential & opportunity | Main risk | Outlook |
|---|---|---|---|---|
| SaaS | Largest buyer market; 68% of Acquire.com deals in its 2025 report | Improve retention, pricing and sales; strong recurring revenue | Churn, crowded tools, outdated code, AI disruption | Strong |
| AI | Very hot, but increasingly crowded | Vertical AI with proprietary data/workflows | Easy replication, model dependence, rapid obsolescence | High upside / high risk |
| Agency | Healthy demand for profitable, systemised agencies | AI automation, productised services, niche specialisation | Founder/client dependence, staff turnover | Good if transferable |
| Marketplace | Attractive after network effects are proven | Expand categories/geographies; monetise both sides | Hard to reach liquidity; platform leakage | Strong but difficult |
| Shopify app | Active buyer niche with recurring revenue | Cross-sell, pricing optimisation, app portfolio roll-up | Shopify policy/API dependence and competition | Good |
| Mobile app | Selective; subscriptions outperform ad-only apps | Improve monetisation, retention and app-store optimisation | Platform rules, declining engagement, copycats | Moderate |
| Ecommerce | Still liquid, 10% of Acquire.com deals, but buyers are cautious | Brand extension, subscriptions and operational improvement | Inventory, ads, tariffs, thin margins | Mixed |
| Content site | Depressed by AI search and Google volatility | Buy cheaply; add products, community or email revenue | Traffic concentration and falling ad/affiliate clicks | Opportunistic |
| Newsletter | Attractive when audience engagement is genuine | Paid tiers, sponsorships, products and community | Creator dependence, weak subscriber quality | Moderate–good |
| Digital products | Simple, high-margin acquisitions | Bundling, localisation, evergreen funnels | Piracy, trend decay, paid-ad dependence | Moderate |
| Crypto | Highly cyclical and difficult to finance | Infrastructure, compliance and recurring B2B tools | Regulation, token exposure, banking/security | Speculative |
| Other / services | Currently one of the strongest SMB segments | Digitisation, AI efficiency, consolidation/roll-ups | Labour intensity and owner dependence | Strong |
Where the value concentrates
Value clusters around four kinds of business:
- Profitable vertical SaaS with low churn.
- Niche agencies with recurring contracts and management already in place.
- Shopify apps or newsletters with diversified customer acquisition.
- Traditional service businesses that technology can improve.
Where the market gets nervous
What AI does to each of these
The lens I keep coming back to is one question. Does AI make a business's customers need it more, or need it less? That line separates a tailwind from an extinction event, and it runs straight through this table.
- SaaS: AI guts thin tools that wrap a form or a prompt. It strengthens software with proprietary data and an embedded workflow. Retention becomes the whole game. A tool people can rebuild in a weekend has no future.
- AI businesses: The wrappers get eaten first. What survives is vertical AI with proprietary data or distribution you can’t copy. A lot of "AI startups" on the market are a feature dressed up as a company.
- Agencies: AI collapses production cost. Good, systemised agencies get more profitable. Generic ones lose their floor. Value moves to judgment, relationships, and repeatable systems, the parts AI can't hand off.
- Marketplaces: AI barely touches the real moat, which is liquidity and network effects. The risk is discovery. If AI answers the buyer before they reach the platform, it leaks.
- Shopify and mobile apps: AI raises what a solo developer can ship overnight, so competition floods in. Platform dependence stays the core risk. Recurring revenue and switching costs are the defence.
- Content sites: The clearest casualty. AI search answers the query without the click, so traffic and ad revenue fall. Cheap prices only matter if the remaining traffic can be converted into an owned audience like email, community, or products, and done fast.
- Newsletters: Trust is the moat, and AI can't fake a genuine human voice at scale. It also floods the top of the funnel with cheap content, which makes real engagement more valuable, not less.
- Digital products: AI makes creation almost free, supply floods, and prices compress. The only protection is brand and distribution, the funnel rather than the file.
- Crypto: Mostly separate from AI, driven by regulation and the cycle. Speculative either way.
- Services and SMB: The quiet winner. These are labour-heavy businesses AI can make cheaper to run, and sellers aren't pricing that upside in yet. The value sits in digitising them and widening the margin.
So where's the demand, and the mispricing?
On the buyer side, everyone is chasing "AI-native". That inflates anything with an AI label and leaves boring but durable service and SaaS businesses underpriced. My read is that the real mispricing sits in unglamorous businesses AI quietly makes cheaper to run, not in the businesses AI is supposed to disrupt. The disruption is already in those prices. The efficiency upside isn’t.
The pattern underneath all of it is simple. AI raises the value of what it can’t replicate, like trust, proprietary data, embedded workflow, and real relationships. It destroys the value of what it can, like generic content, thin tools, and undifferentiated production. The market underpays for the first kind and overpays for the second.
What I take away from it
I'm not shopping. I'm reading the market as a map of which digital businesses are built to last. The takeaway is consistent. Durability comes from what AI can’t replicate, and the tools on top keep changing while the fundamentals of a good business stay the same.
