Atlassian raised cloud prices for Jira and Confluence by 3% to 10% on 13 October [1][2]. Twelve days earlier, OVHcloud raised rates on active servers, saying the memory chips inside now cost six times more than last year [3]. One OVHcloud server example renews 41.2% higher [3]. One hike hits the cost to buy software, and the other hits the cost to build it.

The big picture:

A software plan used to act as a hedge. Vendors bought the parts and took the risk when costs moved. Buyers got one flat number for their budget.

OVHcloud says this flat-rate era is over [7]. What strikes me is how fast the flat part shrank.

The firm now caps how much automation each plan includes. A Standard plan carries 400 steps per user per month, Premium 750, and Enterprise 1,000 [1]. Its AI tool, Rovo, is billed at one cent per credit, or $10 per 1,000 [1]. A support request that an agent closes is billed by the result [1].

Microsoft makes a similar move on 1 December. New Microsoft 365 Copilot Business licenses bought through a Cloud Solution Provider (CSP) — a firm that handles Microsoft billing — will default to pay-as-you-go [6]. That date was moved up from an earlier plan for 2 November [6]. The meter is now the default.

OVHcloud took a different path. It applied new rates to servers already running in its data centers [3]. The vendor blames the rising cost of DDR5 memory chips; storage is close behind [3][7].

TrendForce, a memory research firm, sees no quick relief. It expects memory prices to climb another 10% to 15% this quarter [5]. Flash storage will rise 15% to 20%, pushed by fast enterprise solid-state drives [5].

Demand is why. Gartner expects AI spending to reach $2.7 trillion this year, up 49.5%, and calls it “inelastic to pressures from memory-related pricing increases” [8].

By the numbers

  • 3% to 10% — Atlassian list rise: The jump on new buys, renewals, and upgrades starting 13 October, with the top step on plans above 5,000 seats [2].
  • 41.2% — OVHcloud contract spike: The jump for a server with 128 gigabytes of memory and four solid-state drives, rising from €204.99 to €289.49 [3].
  • Six times — Memory cost jump: OVHcloud’s reported cost rise for the memory chips inside its servers over the past year [3][7].
  • 10% to 15% — Memory price forecast: TrendForce’s expected jump for memory contracts this quarter, with storage rising even faster [5].
Your Jira Bill Rose 10%: Memory Rose Six Times Figures as stated in this article's own numbers section (verified figures, %) Atlassian list rise 3% ([2]) OVHcloud contract spike 41.2% ([3]) Memory price forecast 10% ([5])

What I’d watch:

The teams closest to this shift are reading contracts, not price lists. Four details catch my eye.

  • The free tier: A meter only bites once free use runs out. The firm pools these limits, so the extra use is the number to model [1].
  • The invoice clock: Price changes land on the invoice date. OVHcloud’s hike arrives when the contract renews [3]. The firm honors quotes made before 13 October until they expire [2].
  • The old hardware gap: OVHcloud charges 127% more for memory on new orders, but only 20% to 40% more on machines already in place [3]. Old invoices carry old hardware costs forward.
  • The own-boxes option: Some teams are moving work back to their own servers, arguing that the memory tax is paid once [4]. I read that as an argument about risk, not about price.

I will watch the first quarter of meter data to see if October was just a price hike or a true shift in scope.

The catch

The big 41.2% rise is one vendor’s example for one setup [3]. OVHcloud applied it only to machines using the newest memory chips; older hardware keeps its prices [3]. My read: the direction is clear, but the size is not.

The Atlassian figures come from partner agencies. The firm renders its own prices live in the browser, so a single menu figure is one example [2].

TrendForce’s numbers are a forecast, and shortage forecasts have missed in both directions before [5]. AI made it cheaper to write code. The bills to run that code did not drop.

At a glance

  • The Big Shift: Both sides of the build-versus-buy choice got more expensive in October. Atlassian raised cloud prices 3% to 10%, while OVHcloud hiked rates on active servers by up to 41.2% when they renew [1][2][3].
  • Why It Matters: Flat plans used to shield buyers from hardware costs. Now, vendors are shrinking what is included and passing rising memory and storage costs straight to users [3][5][7].
  • What I’d Watch: How usage meters act in the real world, and how long old contracts hold up.
  • The free tier: The firm pools usage limits, making total extra use the key metric [1].
  • The contract date: OVHcloud’s changes land when a contract renews, spreading the hit out over time [3].
  • The hardware gap: Memory costs 127% more on new OVHcloud orders but only 20% to 40% more on installed machines [3].
  • The Catch: The 41.2% hardware spike applies only to the newest memory chips [3]. The Atlassian data relies on partner copies of live price tables [2].

Related reading

Sources

[1] Atlassian, “Usage-based pricing” — https://www.atlassian.com/resources/usage [2] Adaptavist, “Atlassian Cloud price updates, effective October 2026” (28 August 2026) — https://www.adaptavist.com/blog/atlassian-price-updates-effective-october-2026 [3] OVHcloud, “Dedicated servers: price change in the second half of 2026” (Octave Klaba) — https://blog.ovhcloud.com/en/posts/dedicated-servers-pricing-update-h2-2026/ [4] Hacker News, “Tell HN: OVH price increase for dedicated servers” (2 October 2026) — https://news.ycombinator.com/item?id=49931175 [5] TrendForce, “AI Server Demand Sustains Memory Contract Price Increases in 4Q26” (30 September 2026) — https://www.trendforce.com/presscenter/news/20260930-13258.html [6] Microsoft, “October 2026 announcements”, Partner Center (1 October 2026) — https://learn.microsoft.com/en-us/partner-center/announcements/2026-october [7] OVHcloud, “Public Cloud: the end of all-inclusive pricing” — https://us.ovhcloud.com/resources/blog/public-cloud-the-end-of-all-inclusive-pricing/ [8] Gartner, “Gartner Forecasts Worldwide AI Spending to Grow 49.5% in 2026” (16 September 2026) — https://www.gartner.com/en/newsroom/press-releases/2026-09-16-gartner-forecasts-worldwide-ai-spending-to-grow-49-point-5-percent-in-2026