I currently pay $104.99 per month across seven app subscriptions. The total annual cost is $1,260. The 2010 equivalent of these seven services cost me, on average, $0 per month, because five of the seven did not exist and two were one-time purchases. The shift from one-time purchase to monthly subscription has, in the last fifteen years, been the most important change in consumer software. The shift is, by every measure, deliberate.

Below is the list of what I pay for, what each one cost in 2010 dollars, and what each one would cost in 2026 if I tried to buy it as a one-time purchase. I have changed my mind about several things I used to believe about paying for software. I have opinions on what to skip. There is a brief history of how the subscription model went from Adobe's Creative Cloud gamble in 2013 to the default expectation of every consumer app in 2026. And there is, at the end, an honest summary — because anyone who tells you "subscriptions are the future of software" has not, in fact, tried to cancel lately.

The list: what I pay

1. ChatGPT Plus, $20/month. The most-used subscription on my list, by hours. I have been on ChatGPT Plus since December 2022, when the service launched at $20 per month. The closest 2010 analog — Wolfram Alpha Pro — cost $4.99 per month. The current ChatGPT Plus is, by every measurement, a different product category. Cost in 2010 dollars: $0 (did not exist). Cost in 2026 dollars if one-time: approximately $1,000.

2. New York Times digital + Games, $25/month. The second most-used subscription. I have been on the Times digital since 2011. The current combined price is $25 per month ($17.25 for digital + $6.95 for Games, then $5 for the bundle discount). The 2010 equivalent — a print + digital subscription — cost $20 per month, or $240 per year. The premium, in nominal dollars, is $60 per year. The premium, in real 2010 dollars, is approximately $0 (the inflation adjustment is $84). Cost in 2010 dollars: $20/month. Cost in 2026 dollars if one-time: impossible.

3. Spotify Family, $20/month. The third most-used subscription. I have been on Spotify since 2011, when the service launched in the United States at $9.99 per month for Premium. The current Family plan is $20 per month for up to six accounts. The 2010 equivalent was iTunes, where I bought individual songs at $0.99 and albums at $9.99. The total Spotify Family spend for 6 people in 2026 is $20/month, or $3.33 per person, which is, by my standard, less than the iTunes spend per person in 2010. Cost in 2010 dollars: $9.99/month. Cost in 2026 dollars if one-time: impossible.

4. 1Password Family, $5/month. The fourth most-used subscription. I have been on 1Password since 2015. The current Family plan is $4.99 per month for up to five users. The 2010 equivalent was a notebook of passwords ($5 one-time, replaced every 18 months). The premium, in nominal dollars, is $4 per month. The premium is worth it because of the breach risk. Cost in 2010 dollars: $0.28/month. Cost in 2026 dollars if one-time: $100 (lifetime 1Password license, which used to exist).

5. iCloud+ 2TB, $10/month. The fifth most-used subscription. I have been on iCloud since 2014. The current 2TB plan is $9.99 per month. The 2010 equivalent was an external hard drive ($100 one-time, replaced every 5 years). The premium, in nominal dollars, is $8.33 per month (the amortized hard drive cost). The premium is worth it because of the automatic backup and the Find My integration. Cost in 2010 dollars: $1.67/month. Cost in 2026 dollars if one-time: impossible.

6. Amazon Prime, $15/month. The sixth most-used subscription. I have been on Prime since 2011, when the service was $79 per year. The current monthly equivalent is $14.99 per month (or $139 per year annually). The 2010 equivalent was free 2-day shipping on orders over $25, which did not exist. Amazon Prime was, in 2011, a new category. Cost in 2010 dollars: $0 (did not exist). Cost in 2026 dollars if one-time: impossible.

7. Apple TV+, $10/month. The seventh most-used subscription. I subscribed in 2022 at $6.99 per month. The current price is $9.99 per month. The 2010 equivalent was, by every measure, cable TV, which cost $50-$80 per month for the channels I actually watched. The Apple TV+ library is small but, in 2026, every show is high quality. Cost in 2010 dollars: $0 (streaming did not exist). Cost in 2026 dollars if one-time: impossible.

Total monthly cost: $104.99. Total annual cost: $1,259.88. Total over five years (2021-2026): approximately $6,300.

The list: what 2010 cost

For comparison, here is what I paid for the closest 2010 analogs to my current subscriptions.

  • Wolfram Alpha Pro, $4.99/month (closest to ChatGPT Plus). Canceled 2015 when Google search got good enough for most queries.
  • New York Times print + digital, $20/month (closest to NYT digital). Canceled 2018 when I moved to digital-only.
  • iTunes, $9.99/month average (closest to Spotify Family). Songs at $0.99, albums at $9.99, occasional $0.99 TV show.
  • Notebook of passwords, $5 every 18 months (closest to 1Password Family). Canceled 2015 when I moved to 1Password.
  • External hard drive, $100 every 5 years (closest to iCloud+). Canceled 2014 when I moved to iCloud.
  • Free 2-day shipping on $25 orders, $0 (closest to Amazon Prime). Did not exist. Used USPS.
  • Cable TV, $65/month (closest to Apple TV+). Canceled 2022 when I moved to streaming.

Total 2010 monthly equivalent: $99.98 per month. The 2026 monthly cost is $104.99. The monthly cost has, in nominal dollars, increased by $5. The monthly cost has, in real 2010 dollars, decreased by $75 (after inflation adjustment). The "everything is more expensive now" narrative is, in my arithmetic, wrong for software.

What I have changed my mind about

A few beliefs I held in 2010, when I first subscribed to Spotify, that the last fifteen years have killed.

I used to think subscriptions were a scam. They are, in 2026, the default. The 2010 one-time-purchase model is, in 2026, an exception. The shift happened, by every measure, because subscriptions give companies predictable revenue. Predictable revenue allows companies to invest in continuous updates. The 2010 model — buy Photoshop CS5 for $699, get no updates for 3 years, then buy CS6 for another $699 — was, by every measure, worse for users in the long run. The subscription model is, in 2026, a fair trade.

I used to think one-time purchases were always cheaper. They are, by my standard, usually more expensive over a 5-year horizon. Adobe Photoshop CS5 was $699 in 2010. Adobe Creative Cloud is $660 per year. The Photoshop subscription is, by my arithmetic, cheaper after year 2 of ownership. The "one-time purchase is cheaper" framing is, in most cases, a framing that ignores the 5-year horizon.

I used to think I owned the software I bought. I did. I also, by every measure, paid for the privilege of not getting updates. The "ownership" was, in retrospect, a license to use a specific version. The license, in 2026, is the same — the difference is that the license now includes updates. The "I own it" framing was, in retrospect, a framing that made the lack of updates feel like a feature rather than a bug.

I used to think the free tier was enough. It was, in 2010, sufficient for most apps. It is, in 2026, an entry point to the paid tier. The free tier of Spotify, the free tier of YouTube, the free tier of every freemium app is, by every measure, designed to make the paid tier feel necessary. The "free tier is enough" framing was, in 2010, accurate. The "free tier is enough" framing is, in 2026, wrong.

I used to think annual plans were a scam. Annual plans are, in 2026, a 16% discount over monthly plans. The lock-in is, by my standard, real. The discount is, by my standard, worth it. I have, in 2026, switched every subscription I can to an annual plan. The annual plan I avoid is the one that auto-renews at the end of the year with no reminder. Every other annual plan, in 2026, is worth the discount.

I used to think the App Store was a fair marketplace. It is, in 2026, a 30% tax on every in-app purchase. The Apple-Google duopoly on mobile app distribution is, by every measure, a $21 billion (Apple) + $24 billion (Google) market in 2023. The 30% cut is, by my standard, the largest tax on consumer software in modern history. The alternative — the EU's Digital Markets Act, which forced Apple to allow sideloading in 2024 — is, in my opinion, the right response.

What to skip when subscribing

If I were advising a friend who was about to set up their first app subscriptions in 2026, I would tell them to skip these categories entirely at the start.

Streaming bundles. The Disney+/Hulu/ESPN+ bundle was, in 2022, "the only way to watch Andor, The Bear, and college football." It was also, by 2023, redundant. The bundle is, by every measure, a way to charge you for services you do not use. Subscribe to the individual services. Skip the bundle.

"Pro" tiers of consumer apps. The free tier of most consumer apps — photo editors, video editors, note apps — is, in 2026, more than sufficient. The "pro" tier is, by my standard, a way to charge for features that used to be free. Skip the pro tier until you have actually used the free tier enough to know what you need.

Anything with a free trial that requires a credit card. The free trial that requires a credit card is, by every measure, a setup for an automatic charge. The free trial that does not require a credit card is, by my standard, the only honest free trial. Subscribe with intention. Skip the auto-trial.

Anything with "Premium" or "Plus" in the name under $10/month. The Premium or Plus tier under $10/month is, by every measure, the entry point to a higher tier. The service wants you to upgrade within 90 days. The Premium or Plus tier is, in 2026, a marketing artifact, not a product.

Annual plans with auto-renewal. The annual plan is, by my standard, worth the 16% discount. The annual plan that auto-renews at the end of the year with no reminder is, by my standard, a setup for an automatic charge. Set a calendar reminder 30 days before the renewal. Skip the auto-renewal.

Calculator and weather subscriptions. The calculator app is, by every measure, a built-in iPhone feature. The weather app is, by every measure, a built-in iPhone feature. The 2024 wave of "premium calculator" and "premium weather" apps at $2.99/month is, by my standard, the most cynical subscription category in the modern app economy. Skip them. Use the built-in app.

A brief history

The subscription economy for software is, by every measure, a 21st-century invention. Until 2005, the standard model for consumer software was a one-time purchase. Microsoft Office cost $399 one-time in 2005. Adobe Photoshop cost $649 one-time in 2005. The "subscription" model for software was, by 2005 standards, mostly limited to enterprise software (Salesforce, founded 1999) and Adobe's Acrobat ($15/month, launched 2004).

The consumer subscription economy was, in large part, built by three things: Adobe's move to Creative Cloud (2013), the rise of streaming video (Netflix streaming 2007, Hulu 2008, Spotify 2008, Disney+ 2019), and the App Store's introduction of subscriptions (2011, expanded 2016). Each of these shifted, by every measure, the default expectation from "buy once" to "pay monthly."

The 2020s subscription economy is, by every measure, in retreat. The "subscription fatigue" backlash of 2023-2024, the rise of one-time-purchase alternatives (Affinity, Capture One, DaVinci Resolve), and the EU's Digital Markets Act (2024, forcing Apple and Google to allow sideloading) have, by every measure, restored the choice. The seven subscriptions I pay in 2026 are, by my standard, the seven I actually use. The seven I would have paid in 2010 are, by my standard, more expensive than the seven I pay now.

The Apple App Store subscription revenue was, by Apple's own accounting, $21 billion in 2023. Google Play subscription revenue was, by Google's own accounting, $24 billion in 2023. The combined $45 billion in 2023 is, by every measure, more than the entire global recorded music industry. The subscription economy is, in 2026, the largest single category in consumer software.

The honest summary

The shift to subscriptions is, in the end, a shift in who pays for the software. The 2010 one-time-purchase model required the user to pay for development up front. The 2026 subscription model requires the user to pay for development continuously. The trade-off is, by every measure, real — users pay less per month, but they pay forever. The companies get predictable revenue. The developers get ongoing support. The trade-off is, in my opinion, fair when the service is worth using.

What works for me: an AI subscription, a news subscription, a music subscription, a password manager, a cloud backup, a shipping service, and a video service. The total spend is $104.99 per month. The total value, by every measurement, is incalculable.

What does not work for me, and may not work for you: the idea that subscriptions are always bad. The 2010 model was, by my standard, worse for users in the long run. The subscription model is, in 2026, the only way to get continuous software updates. The subscription model is also, in 2026, the only way to lose track of what you are paying for. The honest summary is, in the end, this: the industry is selling recurring revenue. The seven services, when chosen correctly, do not.

Six years in, I am paying $104.99 per month for seven services. The 2010 equivalent cost me, on average, $100 per month for six different services. The nominal cost has increased by $5. The real cost has decreased by $75. The convenience has increased by a factor I cannot measure. The honest answer to the question "why does every app want a subscription" is, in the end, the obvious one: because it works, for them.