Every PULA trap is a default, and a perpetual default never expires
The ten traps in an Oracle PULA are not hidden clauses. They are what the standard order form does when nobody overrides it. A ULA at least gives you a certification day to escape bad language. A PULA removes that day, so the paper you sign is the paper you keep.
Prepared by Redress Compliance · August 14, 2026 · Oracle licensing advisory. Based on 20 to 30 PULA proposals modeled 2024 to 2026.
Executive summary
All ten PULA traps are defaults in Oracle's standard form, not deceptions. Silence on cloud rights excludes cloud. Silence on affiliates excludes the affiliates. Nothing needs to be hidden, because anything left vague at signature resolves in Oracle's favor.
The PULA's defining feature is what it removes: the deadline. In a ULA a drafting mistake expires with the term, because certification forces an exit and a fresh negotiation. A PULA has no certification day, so a drafting mistake made in year one is still billing in year fifteen.
In roughly nine of ten deals we model, the PULA is not the cheapest shape. A structured set of perpetual licenses with standard support costs less over a ten year horizon. The PULA earns its 20 to 40 percent premium only when scope is genuinely volatile and exit is not a priority.
The support stream is where the money concentrates. Support runs at 8 to 12 percent of the license fee per year with a 3 to 5 percent annual uplift by default. At the top of that range the support line doubles inside fifteen years, on an agreement designed never to end.
The negotiation is therefore a drafting exercise, not a discount hunt. Every lever that matters, the product families, the entity tree, the territory, the cloud clause, the renegotiation window, is a sentence in the order form, and each is written or it does not exist.
The ten defaults, side by side
| Trap | Category | What the default does | The line to demand |
|---|---|---|---|
| 1. Fixed product list | Scope | Future Oracle products, options, packs, and acquisitions sit outside the grant | Name product families, not individual SKUs |
| 2. Narrow entity scope | Scope | Subsidiaries and acquired entities fall outside the PULA | Legal parent plus all controlled affiliates, with an M&A and divestiture clause |
| 3. Territory binding | Scope | Country level scope blocks deployment in unlisted territories, including DR sites | Region level scope wherever the business operates |
| 4. Cloud excluded | Scope | AWS, Azure, and Google Cloud workloads fall outside the grant by default | A BYOL clause naming the public cloud providers explicitly |
| 5. Java excluded | Scope | Java SE Universal Subscription sits outside the product list | The Java position in writing, priced separately |
| 6. Support never sunsets | Financial | 8 to 12 percent of the license fee, every year, forever | A support cap and a defined recalculation basis |
| 7. Uplift and reset | Financial | 3 to 5 percent yearly uplift, and Oracle reserves the right to reset support pricing | Cap the escalator, at zero if you can |
| 8. Audit on scope drift | Financial | Deployment beyond the named scope triggers compliance review | Define drift, cure periods, and audit mechanics up front |
| 9. No certification day | Exit | No natural checkpoint ever forces the agreement open again | A scheduled renegotiation window in the order form |
| 10. Cloud lock in | Exit | OCI commitments get entangled with the PULA relationship | Keep cloud spend commercially severable |
Nothing in this table is concealed. Each row is simply the standard form's answer to a question the buyer did not ask in writing. That is what makes the PULA unusual: the risk does not live in exotic clauses, it lives in nouns. The product list, the entity list, the territory list. Whoever writes those lists writes the agreement.
The old advice was to read the order form line by line. The better advice is to treat every line as yours to draft, because on this paper there is no year three correction, no certification exit, no renewal cliff where the mistake surfaces and dies. It just compounds.
The five scope defaults, and what each costs
- The product list is the entire universe of the grant. A PULA written as a list of SKUs excludes every option, pack, and acquisition Oracle ships after signature. Written as product families, the same paragraph covers the estate you will actually run in year five.
- The entity clause decides who is allowed to use what you bought. Acquisitions arrive outside the grant and divested entities leave stranded, unless parent plus controlled affiliates and an M&A mechanism are written in. On a perpetual horizon, corporate change is not a risk, it is a certainty.
- Territory follows the paper, not the workload. Country level scope blocks the disaster recovery region you have not built yet and the market you have not entered yet. Regions, not countries.
- Cloud rights do not exist until written. The authorized cloud environment policy carries its own counting rules for AWS and Azure, and Google Cloud needs explicit treatment. A PULA silent on hyperscalers is a datacenter agreement wearing an unlimited label.
- Java is a separate conversation. Java SE Universal Subscription prices per employee and almost never sits inside the PULA list. If the account team implies it is covered, that sentence belongs in the order form or it is not true.
The PULA exit playbook
The negotiated certification method, the divestiture and M&A clause set, the support stream analysis, and the exit sequencing for estates inside a perpetual agreement.
Get the playbook →A drafting exercise wearing a pricing disguise
Most Oracle negotiations are priced by leverage: the calendar, the alternatives, the compete. The PULA negotiation is different in kind, and treating it like a discount hunt is how the ten defaults survive into signature.
Start with what the perpetual premium buys. Oracle builds the price from your current deployment, a three to five year growth projection, and a 20 to 40 percent premium over a comparable ULA. What the premium removes is the certification day, and with it, every future moment where the two sides must renegotiate.
That removal cuts both ways, and this is the asymmetry the sales pitch does not dwell on. The certification day in a ULA is not only an exit. It is the deadline that disciplines the drafting on both sides, because everyone knows the language will be tested and renegotiated within a few years.
Take the deadline away and the weight shifts entirely onto the signature draft. A scope noun you fail to write in 2026 is not a problem you inherit at the next renewal, because there is no next renewal. It is a problem you inherit annually, at 8 to 12 percent support with a 3 to 5 percent escalator, for as long as the estate runs.
Run that arithmetic once: at the top of the default range the support line doubles inside fifteen years. On a perpetual horizon, the question is not whether you can afford the license fee. It is whether you can afford the sentence you left out.
This is why nine of ten modeled deals favor a structured perpetual license set instead. Not because the PULA price is wrong, but because the PULA's value case depends on scope volatility that most estates do not actually have, while its cost case compounds regardless.
It also reframes what a good PULA negotiation produces. The win is not a lower number on the license line. The win is a shorter list of things that remain unwritten: product families instead of SKUs, the corporate tree instead of a named entity, regions instead of countries, hyperscalers named, Java priced, a support basis defined, and a renegotiation window that restores, by contract, the checkpoint the PULA deleted.
Where the deal shape itself is still open, price the PULA against the ULA it replaces using the PULA pillar analysis, and pressure test the unlimited need against the ULA guide. The wider Oracle position sits in the Oracle practice.
Watch the briefing · 4:30How to Negotiate an Oracle ULA: No Price List, Just Your Business CaseThere is no price list: the fee is a story built from your estate and your growth. Conservative growth answers, a narrow product list, your own breakeven model, and the certification exit negotiated before signature.
- Percentile standing for your exact deal size and industry, from real closed transactions
- Scenario simulation before the call: test alternative terms and see the financial impact of each
- A negotiation playbook, talking points, and a two page executive brief on day one
What we saw across PULA proposals, 2024 to 2026
Across the 20 to 30 PULA proposals we modeled between 2024 and 2026, the same findings kept repeating, and none of them turned on the license fee:
Deals where a structured set of perpetual licenses with standard support beat the PULA over a ten year horizon, once the support stream was priced honestly.
Proposals arriving with all ten defaults intact: SKU level product lists, single entity scope, no cloud clause, no renegotiation window, uncapped uplift.
Three patterns recurred. Product lists written as SKUs rather than families, which quietly excluded the options and packs the estate later needed. Entity clauses with no M&A or divestiture mechanism, which turned corporate activity into license events. And silence on the public clouds, which converted every migration conversation into a new Oracle negotiation.
The buyer side move in every case was the same: draft first, price second. The wider library sits in the Oracle practice.
Your first five moves
- Model the structured perpetual alternative before responding to the PULA proposal, because nine of ten deals price better that way and the comparison is your strongest anchor even if you still choose the PULA.
- Rewrite the three lists before discussing the number: product families instead of SKUs, parent plus controlled affiliates instead of a named entity, regions instead of countries.
- Write the cloud and Java sentences yourself. A BYOL clause naming AWS, Azure, and Google Cloud, and the Java position priced separately in writing, close the two scope gaps that surface most often after signature.
- Cap the support stream at signature: an escalator cap, at zero if you can get it, and a defined recalculation basis, because 8 to 12 percent with a 3 to 5 percent uplift doubles inside fifteen years at the top of the range.
- Restore the deadline the PULA deleted. A scheduled renegotiation window and a divestiture and M&A clause put a checkpoint back into a perpetual agreement. The Oracle practice drafts and negotiates these terms with you.
Frequently asked questions
What are the ten Oracle PULA contract traps?
Five scope defaults: a fixed product list, an entity clause narrower than the corporate tree, territory binding, cloud rights excluded, and Java excluded. Three financial defaults: a support fee that never sunsets, a renewal price reset, and audit exposure on scope drift. Two exit defaults: no certification day and cloud spend tied to the agreement. Each one is standard form language, not a hidden clause.
Why are PULA traps called defaults rather than tricks?
Because every one of them is simply what Oracle's standard order form does when the buyer does not override it. Nothing needs to be concealed. Silence on cloud rights excludes cloud. Silence on affiliates excludes the affiliates. Anything left vague at signature gets resolved in Oracle's favor, and in a perpetual agreement there is no later checkpoint that forces the language open again.
How does a PULA differ from a ULA on contract risk?
A ULA ends with certification, so a badly drafted clause has an expiry date: you certify, exit, and the mistake dies with the term. A PULA has no certification day, so the same drafting mistake is permanent. That is why the drafting bar at signature is higher for a PULA than for any other Oracle commit shape.
Is a PULA cheaper than perpetual licenses over ten years?
Rarely. In roughly nine of ten deals we model, a structured set of perpetual licenses with standard support costs less over a ten year horizon. The PULA earns its price only when scope is genuinely volatile and exit is not a buyer priority, and it carries a perpetual premium of 20 to 40 percent over a ULA of similar scope.
How does Oracle build a PULA price?
From your current processor and named user position, a three to five year scope projection, and a perpetual premium of 20 to 40 percent over a comparable ULA. Support then runs at 8 to 12 percent of the license fee per year with a 3 to 5 percent annual uplift by default, and that support stream is where the long term cost concentrates.
Is Java included in an Oracle PULA?
Almost never by default. Java SE Universal Subscription is a separate product priced per employee, and it sits outside the PULA product list unless it is written in. Account teams sometimes imply otherwise in conversation, so the Java position belongs in writing, priced separately, before signature.
Can an Oracle PULA be renegotiated after signature?
Only if the order form carries the language that forces a checkpoint open: a scheduled renegotiation window, a divestiture and M&A clause, and a defined support recalculation basis. Without that language there is no natural moment where Oracle must return to the table, which is precisely why the levers have to be drafted in at signature.
