Google has repackaged Workspace AI three times in fourteen months, moving base prices 16 to 22% and withdrawing a purchased SKU mid-term, and almost no enterprise contract prices that risk
Between January 2025 and July 2026, Google folded the $20 and $30 Gemini add-ons into base tiers with a $2 to $4 per seat increase, replaced promotional AI access with a paid AI Expanded Access SKU, then withdrew AI Ultra Access from accounts entirely. Every one of those moves was contractually free for Google because standard Workspace paper contains no repackaging protection. The clause you table at this renewal decides whether the fourth wave costs you nothing or another 20% per seat.
Prepared by Redress Compliance · August 28, 2026 · Google Cloud and Workspace advisory. Renewal and repackaging engagements 2024 to 2026.
Executive summary
The Gemini fold-in was a price increase disguised as a feature grant, and it cost buyers 16 to 22% at the base tier.
Business Standard moved from $12 to $18 and Business Plus from $22 to $28, which is a $6 per seat annualised hit of $72, delivered in exchange for capability most organisations had not asked for and could not decline.
Google withdrew a SKU customers had already paid for, and the only remedy on standard paper was proration to the cancellation date.
AI Ultra Access stopped being sellable and began disappearing from accounts on July 7, 2026, with no credit for the parity gap and no compensation for the features, including Flow, that did not survive the migration into AI Expanded Access.
Substitution pushed buyers into a separately priced product line, which is the real cost and the one no proration covers.
Customers relying on Antigravity, Gemini CLI or Gemini Code Assist were redirected to Gemini Enterprise Agent Platform or Gemini Enterprise Standard and Plus at $30 to $50 per seat, converting a bundled entitlement into a new purchase order.
A repackaging clause is cheap to win at signature and impossible to win after the announcement, and buyers with $10M combined GCP and Workspace spend have threaded both lines for 15 to 20% better overall rates.
Ask for no net price increase on repackaging, feature parity on migration, and a right to revert or exit the affected line, and price the ask against your commit rather than as a standalone legal favour.
What Google actually did, and why standard paper let it
Read the three events as one playbook and the shape becomes obvious. In January 2025 Google stopped selling the $20 Gemini Business and $30 Gemini Enterprise add-ons, folded the capability into base tiers, and recovered the revenue through a $2 to $4 per seat base increase. That was not a discount.
Base list moved roughly 16 to 22% (Starter $6 to $8, Standard $12 to $18, Plus $22 to $28), and there was no opt-out: on Standard and above you cannot buy Workspace without Gemini.
Thirteen months later, February 2026, the AI Expanded Access SKU arrived and the promotional Nano Banana Pro and Veo 3.1 access that customers had been using became a paid line, with a March 1, 2026 deadline to purchase or lose it.
Five months after that, on July 7, 2026, Google withdrew AI Ultra Access, a SKU customers had bought with money, from accounts entirely, redirecting Flow, Antigravity, Gemini CLI and Code Assist users toward Gemini Enterprise Standard at $30 or Plus at $50.
Each move was contractually free because standard Workspace paper grants Google the right to modify service composition, and buyers signed it. The shift from opt-in to opt-out purchasing is the whole story: your consent was engineered out of the transaction.
| Wave | What Google did | What the contract permitted | What the buyer paid | Buyer's actual option |
|---|---|---|---|---|
| Jan to Mar 2025 | Add-ons discontinued, Gemini folded into base | Unilateral service composition change | +$2 to $4 per seat, 16 to 22% list move | Accept at renewal or exit |
| Feb 2026 | AI Expanded Access created; promotional Nano Banana Pro and Veo 3.1 access ends Mar 1 | Promotional features carry no entitlement | New add-on line, or capability loss | Buy the SKU or lose function |
| Jul 2026 | AI Ultra Access withdrawn from accounts | No continuity obligation on purchased SKUs | Prorated refund only, no parity credit | Auto-migrate to Expanded Access, or cancel |
The staging sequence is the part of this the table cannot show, and it is the part that matters at the table. New customers absorbed the 2025 increase on January 16, flexible-plan customers on March 17, and annual contract holders only at their next renewal.
Google repeats that pattern because it minimizes churn: it hits the least protected populations first and gives contracted customers a runway. That means your renewal date, not Google's announcement date, is the moment your protection either exists or does not. Two operational consequences follow.
First, if you are on flexible billing you have no runway at all, which is a reason to move to a commit for protection rather than price alone.
Second, usage data and log events for removed AI Ultra licenses stay available for only 180 days, so your evidence of what you actually consumed under the withdrawn SKU has a hard expiry. Capture it before you argue parity.
Pricing the harm: what repackaging costs a 2,000 seat estate
The arithmetic is unglamorous and that is precisely why it goes unmodeled. Take 2,000 Business Plus seats. The 2025 fold-in moved list from $22 to $28, a $6 per seat per month delta, which is $144,000 a year of new spend before anyone opens a single Gemini prompt. That number stands alone.
Layer AI Expanded Access on top for the subset that lost promotional Nano Banana Pro and Veo 3.1 access, then account for the population Google redirected out of Ultra: Flow, Antigravity, Gemini CLI and Code Assist users now land in Gemini Enterprise Standard at $30 or Plus at $50.
A separately priced product line with its own commit terms.
If 300 of your 2,000 seats are in that developer and creative population, Standard at $30 adds $108,000 annually and Plus at $50 adds $180,000.
On a $528,000 base estate, the repackaging cascade plausibly adds 25 to 60% to total Workspace-adjacent spend inside eighteen months, and none of it appeared as a renewal price increase you could contest.
Two things make this harm invisible until you build the model. Enterprise tiers are quote-based, so there is no published list to anchor a percentage complaint against; market-reported bands of $23 to $36 per seat are planning figures, not negotiating evidence.
And the 12-month commit versus flexible monthly choice swings per-seat price by roughly 20% on its own, which means a repackaging increase can be partially masked, or badly compounded, by a billing-model decision made separately by a different person.
Model those variables together, not sequentially.
A price hold with a genuine uplift cap only protects the line items it names, so name the repackaging exposure explicitly and quantify it in the same spreadsheet as your seat count, or you will negotiate hard on a $528,000 base while a $288,000 substitution cost walks in unpriced.
What Gemini for Workspace adds to your bill
What Gemini for Workspace really costs as a Workspace add on: named user licensing, bundling pressure, and the buyer side levers that cap the spend.
Get the white paper →The three-part clause: no net increase, migration parity, right to revert
Draft this as three limbs, in this order, because the order determines what survives the deal desk. Limb one, no net increase: any Google-initiated repackaging, bundling, unbundling, tier consolidation.
Or withdrawal of a SKU you have purchased shall produce no net increase in effective per-seat cost for the affected population, measured against the blended pre-change rate, for the remainder of the term plus one full renewal cycle.
Blended is the operative word. The 2025 move raised base tiers 16 to 22% (Starter $6 to $8, Standard $12 to $18, Plus $22 to $28) while removing $20 and $30 add-ons, so a naive comparison of add-on line to base line reads as savings.
Only the blended per-seat figure across your whole estate tells you what happened. Limb two, migration parity: a replacement SKU must deliver materially equivalent functional scope, and where it does not.
The delta is credited against the affected line or supplied at no charge from the adjacent product line for the balance of the term.
The AI Ultra to AI Expanded Access transition is the worked example to cite verbatim in your redline: from July 7, 2026, AI Expanded Access no longer includes Flow, and Google directs Antigravity users to Gemini Enterprise Agent Platform and Gemini CLI users to Gemini Enterprise Standard or Plus.
Both separately priced lines starting at $30 per seat.
That is not a substitution, it is a cross-sell dressed as one. Limb three, right to revert: a 90-day window from written notice to revert to the prior commercial position, or to terminate the affected line at no penalty, with no disturbance to the balance of your commit.
Add the eligibility rider: replacement SKUs must remain available on every tier the original supported. AI Ultra Access ran on Business Base, Business Starter, Business Standard, Business Plus, Enterprise Starter, Enterprise Standard, and Enterprise Plus.
AI Expanded Access dropped Base and both Starter tiers, which means the remedy for those seats is a tier upgrade you did not budget. Pair this with your Workspace price hold and uplift cap so the two clauses reference the same blended baseline.
The limb buyers forget is the evidence one. Usage data and log events for removed AI Ultra licenses stay visible for 180 days after removal, then disappear.
If you intend to claim a parity credit, you need consumption evidence captured inside that window, which means your clause should carry an obligation on Google to preserve or export affected usage telemetry for the full remedy period, not the default retention.
Also fix the default. Taking no action transitions licenses automatically to the replacement SKU, and cancellation produces a prorated final bill with no credit for the parity gap. Flip that: silence should preserve your position, not migrate it.
Why Google will concede parity long before it concedes price
Repackaging is not a pricing accident. It is the mechanism Google uses to move Workspace revenue per seat without holding a price negotiation, because a bundle change is announced rather than agreed.
Three waves in fourteen months, each one delivering a revenue effect that would have required a signature if it had been presented as a rate increase, and none of them requiring one. Any vendor that finds a lever with that yield and that little friction will pull it again.
Understand where the money is going. Google is buying share on the infrastructure side, where a three-year $10M GCP commitment routinely closes at 28 to 35% effective discount, deeper than the comparable AWS or Azure structures.
That discount has to be recovered somewhere, and Workspace, with its captive seat base and its absence of a real switching event, is where.
So when your account team faces a choice between conceding on Workspace rate and conceding on functional commitments, the rate concession hits a forecast line that someone senior is personally carrying. The functional commitment does not appear in the model at all.
That asymmetry is your entire opening. The parity limb and the revert limb cost Google nothing in modelled revenue. Parity says: if you take Flow away, give me the equivalent or credit the delta. Revert says: if you change the deal, I get 90 days to decline. Neither reduces the contracted number.
Both are therefore approvable at a materially lower authority level, often inside the account team with legal sign-off, without the deal desk escalation that any rate concession triggers. The no-net-increase limb is the opposite.
It caps a revenue mechanism the company is actively using across the base, which means it goes up the chain and comes back with conditions.
Buyers who lead with the price limb lose all three. Here is the sequence, and I have watched it run at least a dozen times. You table no-net-increase first. It escalates.
It comes back declined, or approved only for the current term with carve-outs for "changes to underlying third-party components." By the time that round closes, the negotiation has hardened, the clock has moved.
And parity and revert are now bundled into the rejected package rather than standing on their own merits.
You end with nothing but a slightly better discount.
Buyers who lead with parity and revert bank two limbs early, at low friction, often in the first redline exchange. Now the conversation changes shape.
You are not asking for price protection in the abstract, you are asking Google to close the one remaining gap in a structure it has already substantially agreed to. And you have a fallback that costs you nothing: if no-net-increase is genuinely unapprovable, take a capped uplift instead.
A ceiling of 5% on any repackaging-driven increase, measured blended, is a realistic landing zone at 1,000 seats and above, and it converts an open-ended exposure into a budgetable one. Our Google Cloud contract terms guidance sets out where those caps typically hold.
The structural point is the one worth internalising. This clause is not really price protection. It is a re-opener. What it does is convert a unilateral vendor action into a negotiation you get to attend, at a moment when you still hold a live contract and a functioning alternative.
A re-opener is worth more than a discount, because a discount protects one number for one term while a re-opener protects your position against every move the vendor has not thought of yet. Google has run three repackagings in fourteen months. Price the fourth.
What the vendor says back, and the counters that work
Table the three-part clause and you will hear four rehearsed responses, in roughly this order. First: we cannot contractually freeze the product roadmap. Nobody asked Google to freeze anything. The clause governs price and access, not engineering.
Google is free to fold, rename, split, or retire any SKU it likes, provided the per-seat cost of the functionality you are actually consuming does not rise as a consequence. Say that sentence out loud in the room and the roadmap objection dies, because it was never an answer to the ask.
Second: the repackaging delivered added value, so no compensation is owed. Added value is a defense only where the buyer had a choice. There is no opt-out and no way to buy Workspace without Gemini on Standard and above, so the transaction was a compulsory purchase, not an upsell you accepted.
Value delivered under compulsion does not settle a price question; it just relocates it.
Third: standard terms are non-negotiable at your spend level. Google's own conduct contradicts this. It ran a three-stage rollout in 2025 (new customers on January 16, flexible plans on March 17, annual contracts at next renewal) precisely because contract paper differs by cohort.
It also issued a grace window on AI Ultra Access, protecting pre-May 5, 2026 purchasers through July 7. Both are bespoke commercial accommodations. Ask which desk approved them. Fourth: proration is the remedy.
Cancellation prorated to date, with credit overages unchanged and no compensation for the parity gap, is not a remedy.
It is an exit dressed as one, and it is worth less than nothing when the withdrawn feature (Flow, for instance) reappears inside a separately priced Gemini Enterprise line at $21 to $30 per seat.
On leverage: Google deal desks recognize and discount pricing-only competitive threats. An M365 alternative that has been operationally evaluated (pilot cohort, migration cost modeled, sponsor named) lands deals 5 to 10 points above the median band.
Pair the clause ask with a genuine price hold and uplift cap and treat them as one package, not two concessions.
Evidence base and the patterns we see repeat
Starter moved $6 to $8, Standard $12 to $18, Plus $22 to $28 in early 2025 while the $20 and $30 Gemini add-ons were withdrawn.
Usage data and log events for removed AI Ultra licenses stay available for 180 days, which sets your evidence-capture deadline.
Across engagements the sequence repeats with enough regularity to plan against. Rollouts are staged so the increase lands hardest at annual renewal, which means the cohort with the least room to react absorbs the full delta.
Migration defaults favor Google: taking no action moves licenses automatically to the replacement SKU, and cancelling requires affirmative action before a published deadline (July 7, 2026 on flexible plans, next renewal on annual).
Eligibility narrows on the way through: AI Ultra Access covered Business Base, Starter, Standard, Plus and Enterprise Starter, Standard, Plus, while AI Expanded Access covers only Business Standard, Business Plus, Enterprise Standard and Enterprise Plus.
So Starter and Base populations lose the upgrade path entirely and get pushed up a tier to regain it.
Then the log window closes at 180 days. Capture consumption evidence inside it or you argue the next repackaging on assertion.
One live discrepancy to resolve before you quote anything internally: published 2026 Workspace list pricing splits between sources reporting Business Standard at $14 on annual commit and sources reporting $18. Verify against Google's own billing page, not a third party, and not us.
Enterprise has no list at all, so the $23 to $36 band is a planning figure only.
Achievable outcomes, from our engagement data: 15 to 30% off list at 250 to 1,000 seats, 25 to 35% above 1,000, renewal discounts typically 10 to 20%, three-year commits adding 7 to 10 points and five-year another 3 to 6.
Read those bands alongside the wider Google commit and Workspace redlines, because a discount won without repackaging protection is a discount Google can reclaim in one product announcement.
- 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
Your first five moves
- Pull 180 days of usage and log evidence before the window shuts, because Google retains usage data and log events for removed AI Ultra Access licenses for only 180 days after removal, and without that export you have no factual basis to price the parity gap when the deal desk asks you to quantify harm.
- Model the fold-in and the substitution on your actual seat mix, not on list, stacking the $2 to $4 per seat base increase against the AI Expanded Access line and the redirect of Flow, Antigravity, Gemini CLI, and Code Assist users into Gemini Enterprise at $30 to $50 per seat, then separate the Starter and Base populations that lost the upgrade path entirely and now have no eligible product to buy.
- Diary the non-renewal notice date and serve it as a placeholder, roughly 90 days out, so the notice lands before Google's fiscal pressure builds rather than after; in our experience a served notice is worth more at the Google table than any verbal threat, and it costs nothing to withdraw.
- Table the three limbs with parity and revert as the lead ask, holding no-net-increase on repackaging as the concession you trade last, and pair it with the mechanics in the Workspace price hold and uplift cap language so a fold-in cannot be reclassified as a list adjustment.
- Thread Workspace into the GCP commit as one Private Pricing Agreement if combined spend approaches $10M, where three-year commitments at that scale routinely close at 28 to 35% effective discount; use the commit and Workspace redlines to target 15 to 20% better blended rates than the two agreements priced separately.
Frequently asked questions
What exactly is a repackaging clause in a Google Workspace contract?
It is a protection that says if Google changes how a capability you already pay for is packaged, bundled, unbundled or withdrawn, your effective per-seat cost does not rise for the balance of the term.
It normally has three limbs: no net price increase, functional parity in the replacement SKU, and a right to revert or exit the affected line without penalty.
Google's standard Workspace paper contains none of these, which is why the March 2025 Gemini fold-in and the July 2026 AI Ultra Access withdrawal were both contractually free for Google.
Did the Gemini fold-in actually increase prices, or was it just added value?
Both, but the price movement is the part that matters commercially.
Google discontinued the $20 Gemini Business and $30 Gemini Enterprise add-ons and raised base plan prices by $2 to $4 per user per month, with commonly cited moves of Business Starter $6 to $8, Standard $12 to $18 and Plus $22 to $28, a 16 to 22% increase.
There was no opt-out on Standard and above, so the value argument does not answer the buyer who did not want the capability.
What happened to AI Ultra Access and what were customers actually owed?
Google stopped selling AI Ultra Access and began removing it from Workspace accounts on July 7, 2026. Administrators who purchased before May 5, 2026 kept the features until at least that date, then licenses transitioned automatically to AI Expanded Access unless the buyer acted first.
On cancellation the only remedy is proration to the cancellation date, with credit overages unchanged, and no compensation for functionality lost in the swap.
Is AI Expanded Access a like-for-like replacement for AI Ultra Access?
No, and that gap is the entire case for a parity limb.
From July 7, 2026 AI Expanded Access no longer includes Google Flow, and Google redirects Antigravity users to Gemini Enterprise Agent Platform and Gemini CLI or Code Assist users to Gemini Enterprise Standard or Plus, which are separately priced at roughly $30 and $50 per seat.
Eligibility also narrowed, since AI Expanded Access is limited to Business Standard, Business Plus, Enterprise Standard and Enterprise Plus, dropping the Base and Starter populations entirely.
How much discount can we realistically get on Workspace Enterprise?
Reported bands put 250 to 1,000 seats at roughly 15 to 30% off list and above 1,000 seats at 25 to 35%, with renewal discounts more typically 10 to 20%. Three-year commitments add around 7 to 10 points over one-year and five-year another 3 to 6.
A genuinely evaluated Microsoft 365 alternative lands deals 5 to 10 points above the median band, while pricing-only threats are recognised and discounted by Google deal desks.
When should we raise the repackaging clause, and what if we are mid-term?
Raise it at least 120 days before renewal, because it is close to unwinnable once a repackaging has been announced.
Mid-term you have less leverage, but you still have two things: the 60-day non-renewal notice, which you can serve as a placeholder to force a conversation, and any live or pending GCP commitment negotiation, which is where the clause can be traded rather than begged for.
Organisations with $10M or more in combined GCP and Workspace spend have negotiated 15 to 20% better overall rates by threading both lines into a single Private Pricing Agreement.
What evidence should we capture before a withdrawn AI SKU disappears?
Usage data and log events for removed AI Ultra licenses remain available for review for 180 days after removal, so set your capture deadline inside that window.
Pull per-user activity, the feature set actually consumed, and the date the entitlement changed, because that is the record that quantifies the parity gap if you later argue for a credit. Without it you are negotiating a substitution loss on assertion rather than on data.
