Mitchell Sorkin·July 24, 2023

Building an ATM Empire | Mitchell Sorkin Interview

Mitchell Sorkin is a former e-commerce and SaaS entrepreneur (co-founder of Stack Influence) who grew disillusioned with the VC funding treadmill and pivoted to buying ATM routes in the Los Angeles area alongside his brother. He entered the industry through a creative carve-out — convincing a seller to split off just three machines from an 80-machine route for $36k — and scaled to 140 machines by the time of the interview through a combination of broker relationships, conference networking, and Twitter-sourced private debt capital. The operation generates roughly $37-38k/month in top-line revenue with 47% average EBITDA margins, and one seller-financed deal achieved over 100% cash-on-cash return. The core bull case is that ATMs serve an underbanked population that credit cards and Venmo do not displace, that the industry will persist far longer than tech-pessimists assume, and that the highly fragmented market is ripe for roll-up by a nimble operator who can act faster than processors and PE firms. The main challenges are the near-impossibility of bank or SBA financing (forcing all-cash or private promissory notes), the operational complexity of deal-by-deal syndication, theft risk, and the difficulty of exiting to a strategic buyer at scale.

Deal facts

purchase price: First carve-out: $36k (3 machines); subsequent acquisitions ranged from $30k to $45k per route; total portfolio not stated
multiple: First acquisition at ~10x monthly net (e.g. $30k purchase doing $3k/month net); no aggregate multiple stated
sde ebitda: ~$22k gross profit per month at time of interview (~47% EBITDA margins average); annualizing ~$264k
revenue: ~$37-38k/month top-line (~$456k annualized) at time of interview; dispensing ~$1.1M/month
financing structure: All-cash for first four acquisitions; private debt (promissory notes sold to network/Twitter followers) for later ones; one deal with 40% down and 60% seller financing at 9% over 6 years; no SBA or traditional bank financing
notes: 140 machines total at time of interview; 33% (approx. 47 machines) are vaulting-only (cash loading for other operators); started with 3-machine carve-out from an 80-machine route; raised capital via SPVs syndicated deal-by-deal; planning to raise a fund

Why this business

Mitchell was burned out on the VC/SaaS treadmill and wanted a genuinely profitable business that compounded reliably. He and his brother were looking at everything from gas stations to vending routes on BizBuySell and stumbled into ATMs almost by accident. After a vending deal fell apart, they met an ATM seller willing to carve off just three machines, giving them a low-risk entry. Mitchell fell in love with the predictable cash flows, minimal labor, high margins, and simple operations — cash has no feelings and doesn't quit.

What's working

Extremely high EBITDA margins averaging 47% from day one, with low capex requirements (~3% of gross)
Predictable, recurring cash flows at each location — transaction history closely predicts future performance with only ~15-20% monthly variance
Minimal labor: only one W-2 employee (field loader) in California; cash-heavy model means no personnel drama
Machines have very long useful lives (some 20-26 years old still running), making IRS depreciation schedules far more conservative than reality
Three-pillar revenue model: own machines + vault yourself, vault-only for others, and processing/interchange income — provides multiple ways to layer revenue
Growing deal flow through broker relationships (ATMbrokerage.com), conference networking, and Twitter presence; off-market deals from aging owners (e.g. 80-year-old seller) are available
Roll-up thesis: industry is highly fragmented, too small for PE, so strategic aggregation is possible; planning to raise a dedicated fund
One acquisition achieved 100%+ cash-on-cash return after seller financing (put $50k down, generates $53-54k/year)

What's hard

Cannot get SBA or traditional bank financing for ATM routes — likely due to AML concerns and banks' discomfort with cash-heavy collateral; forces all-cash or private debt deals
Syndicating deal-by-deal via SPVs is exhausting: simultaneously finding deals, convincing sellers to wait, and scrambling to raise capital from LPs
Geographic concentration required — buying a route far away (e.g. Midwest) is very difficult without a local network for vaulting and technicians; not recommended for first-time buyers
Theft is a real operational risk — machines get stolen and cash loaders can be robbed; requires cash-in-transit insurance (Marshall & Sterling) and/or self-insuring as a cost of doing business
ATM sellers can be elusive and paranoid (carry large amounts of cash); many are Blue Collar operators unfamiliar with M&A processes
Exit options are limited — market is too small for PE buyers; must rely on strategic buyers or simply harvest cash flow indefinitely
Early BizBuySell acquisitions were somewhat overpriced; better deals come through networking and proprietary outreach
California labor laws make 1099 classification nearly impossible — even a single field worker must be W-2

Notable quotes

The amazing thing about ATMs is vast majority of the parts on the machine are just like a few screws and a plug so anyone who's decent with their hands and has any sort of any even like an ounce of technical Savvy can kind of pull up a YouTube video and figure out how to fix them.

Cash doesn't have feelings. You're not gonna have anybody quitting on you.

People kind of like grossly overestimate how quickly technology can actually disrupt these really established industries. Fax machines last year did a billion in revenue. How does that make any sense?

The math is really simple right — you do like 35 a year so you know every three years you break even, so in six years you've doubled your money. So if the entire industry dies in six years, goes to zero, completely nobody uses an ATM ever again in six years, that's already you doubled your money in six years which is better than the market.