Your Technology Refresh Cycle is Backwards. Here's the Right One

The average practice owner is running a $4,000 server from 2019 to drive a $110,000 CBCT unit from 2024, and calling that a technology plan. It isn't. It's a slow-motion write-off, and the part that breaks first is never the expensive box.

Here's the claim most owners will push back on: your big-ticket clinical hardware is almost never the thing holding your practice back, and you should stop budgeting for it first. The constraint is the cheap layer underneath it (computers, network, software licenses) plus the staff certifications that make the equipment legal to point at a patient. Buy the glamour item first and you get a $110,000 asset throttled by a five-year-old workstation and an assistant who technically can't take the radiograph. That's the backwards cycle, and almost every practice I look at is running it.

The Depreciation Schedules Already Told You the Answer

The IRS has been handing out the correct refresh order for decades and nobody reads it as strategy. Most dental equipment (chairs, units, sterilization, digital imaging systems) falls under 7-year MACRS depreciation, while computer equipment and software qualify for 5-year depreciation. Same source, same page, the practical lifespans diverge even harder: digital X-ray systems run 10 to 12 years, computers and servers run 4 to 5.

Read that as an operating instruction. Over the life of one imaging system you will replace the computers behind it two to three times. Not maybe. Structurally. The imaging sensor is the slow-moving asset and the computer is the consumable, and yet the consumable is the line item that gets deferred every single budget cycle because it's boring and nobody's rep is calling about it.

So when an owner tells me they're "saving up for the next scanner," my honest answer is that they're saving up for the wrong thing on the wrong clock. Refresh the 5-year layer on a 5-year schedule, in writing, funded monthly, and let the 7-to-12-year layer come when it comes. Do it the other way and the expensive asset spends a third of its life underperforming.

43.3% Are Already Using AI and Only 26.4% Plan To

This is the number that made me want to write the piece. The ADA Health Policy Institute found that 43.3% of responding dentists are already using AI for at least one task in their practice, while only 26.4% plan to adopt it in the future.

Sit with the shape of that. Live usage is higher than stated intent. In a normal technology adoption curve those numbers are reversed: intent runs ahead of usage, because people plan, then buy, then deploy. Here, adoption has already outrun the planning. The tools arrived inside software people had already bought, or a hygienist started using one on her own, or the imaging vendor pushed a feature into an update, and it got used before anyone put it on a capital plan.

That's not a story about AI specifically. It's a diagnosis of how technology actually enters a dental practice in 2026: through the software and workstation layer, sideways, unbudgeted. The stuff you scheduled and financed and signed a five-year lease for shows up in the operatory. The stuff that actually changes how the day runs shows up in an update. If your refresh plan only covers the things a rep sells you, it's not covering most of what's happening.

Compare it to intraoral scanners, which are the classic capital purchase: an ADA Clinical Evaluators Panel survey of 369 members published in the August 2021 JADA found 53% of dentists use an intraoral scanner. That took a decade of financing to get to half the profession. The software layer moves in quarters.

Colorado Will Sanction Your License Over a Training Module

This is the failure mode nobody puts on the purchase order. The Colorado Dental Board's revised rule requires any dental assistant exposing radiographs, including digital X-ray and CBCT, to complete one of three approved training paths (a CODA-approved program, a board-approved on-the-job module, or DANB certification), with noncompliance risking sanctions against the dentist's license.

So the actual gate on your imaging upgrade is not financing approval. It's whether the three people who will operate it hold a credential. You can have the unit installed, calibrated, and networked, and still be looking at a compliance exposure on your own license because a module went uncompleted. Colorado is one state board making this explicit, and if you're operating anywhere else the right assumption is that your board has its own version and you should go read it before you sign, not after.

The general principle: every piece of clinical technology has a staffing prerequisite, and the prerequisite has a lead time that nobody quotes you. Equipment arrives in six weeks. Certification arrives when the person finishes the coursework, which is a human timeline you don't control.

What Most Owners Get Wrong: Treating the Rep's Catalog as the Refresh Plan

The specific mistake, and it's most common in solo and two-doctor owner-operated practices where the owner is also the de facto IT director: the technology plan is whatever the equipment reps have been pitching, ranked by how much the owner wants it. There's no line for workstations. No line for the network. No line for staff certification hours. No line for the software subscriptions that quietly rose 40% over three years.

It happens because reps sell the 7-year assets and nobody sells the 5-year ones. There's no commission on replacing four aging front-desk computers, so no one books a lunch to talk about it. The owner ends up with a capital plan that covers maybe 60% of actual technology spend and none of the spend that gates the rest of it.

The second version of this mistake is subtler and I see it in group practices that should know better: buying clinical technology as a staffing substitute without doing the labor math. Sometimes that works. Usually the equipment adds a workflow step that someone has to be trained and paid to run, and the "savings" evaporates into an hour a day of somebody's time.

A $110,000 CBCT That Costs $154,000 in Year One

Run it properly. Say you're buying a CBCT unit at $110,000 and you're doing it the backwards way: equipment first, everything else later.

The four workstations that touch imaging are five years old, which means they're at the end of the 4-to-5-year practical lifespan for computers. Replacing them properly, imaging-capable machines plus monitors, runs about $3,200 each, so $12,800. Network and server work to move the file sizes CBCT generates, call it $9,000. That's $21,800 of infrastructure that has to exist for the $110,000 asset to run at full speed.

Now the staff side. Three assistants need certification training. Using the Bureau of Labor Statistics figure for dental hygienists, a median annual wage of $98,100 in May 2025, the loaded hourly cost of clinical staff time is roughly $47 before benefits. Twenty-four hours of training per person across three people is 72 hours. At $47 that's $3,384 in paid training time, plus roughly $1,500 in course and exam fees for three people, so about $4,884.

Add it up: $110,000 plus $21,800 plus $4,884 is $136,684 for the real cost of the thing you thought cost $110,000.

Here's the part that stings. Do it backwards and the infrastructure doesn't get bought in year one, it gets bought in year two after eight months of slow renders and a support call every week. In those eight months the unit runs at reduced throughput. If CBCT was supposed to support six additional scans a week at a $250 fee and you're capturing four, that's two scans a week of lost revenue for 35 weeks at $250, which is $17,500 left on the table. Sequence it properly and you spend the same $136,684. Sequence it backwards and you spend $154,184 in effect, because the $17,500 is real money you didn't collect.

Same equipment. Same total capital. A $17,500 difference created entirely by the order of the purchase orders.

What I'd Actually Do With the Budget

I'd fund the 5-year layer first and I'd fund it as a standing monthly line, not a capital request. Four workstations at $3,200 on a five-year cycle is $12,800 every five years, which is $213 a month set aside. That's less than most practices spend on coffee and flowers, and it's the difference between equipment that performs and equipment that limps.

I wouldn't finance a major clinical purchase until three things are true in writing: the workstations and network that serve it are under three years old or already ordered, the staff who will operate it have their certification either complete or scheduled with a date, and I've written down what specific procedure volume the thing needs to produce to pay for itself. If any of the three is missing, I'd delay the purchase a quarter. A quarter of delay costs you almost nothing. A year of a throttled asset costs you five figures.

And I'd stop treating the software layer as an operating expense you review at renewal. Given that adoption is already running ahead of intent across the profession, the tools are entering your practice whether you planned for them or not. Better to know which ones, run by whom, and at what cost, than to find out during a billing dispute.

On the labor substitution question: the BLS projects 8% employment growth for dental hygienists from 2025 to 2035, with about 225,300 jobs held in 2025. That's a growing occupation, not a shrinking one. I wouldn't buy any piece of technology on the theory that it replaces a hygienist. Buy it because it makes the hygienist you're already paying $98,100 more productive, and the math works. Buy it as a headcount play and you'll be disappointed twice.

Do This Week

Pull every computer, server, and monitor in the building into a spreadsheet with its purchase date. Anything past four years goes on a replacement schedule with a dollar figure and a month, and you start funding it monthly starting this one. Then pull your state board's radiography rules and check each assistant's certification status against them, today, not at renewal. If you have a clinical purchase in the pipeline, hold it until both of those lists are clean.

What I'm tracking next: whether that 43.3%-versus-26.4% gap closes from the intent side or widens from the usage side. If stated intent catches up, this was a one-time surprise and planning cycles will absorb it. If live usage keeps outrunning intent, then capital planning in dentistry is structurally broken for this class of technology and the answer isn't better forecasting, it's a standing infrastructure budget that doesn't require a forecast at all. That's where my money is. I'd change my mind if the next HPI read shows intent above usage.

If you want to know whether your technology spend is actually the problem or your overhead is hiding something bigger, our free overhead calculator breaks your spend into its real categories and shows you where equipment and infrastructure sit against practices your size. Run your own numbers and see whether the 5-year layer or something else is eating your margin.