The short answer
It depends on your agreement with Google Cloud. If you reach the end of your commitment period having spent less than you committed, you may face a shortfall, even though you negotiated good pricing to get there. One option is to ask Google Cloud about moving some of the unused commitment into a future period.
Another option is often overlooked. Google Cloud Marketplace lets you buy third party software on your Google Cloud bill, and eligible Marketplace purchases can count toward your commitment. If you already pay outside vendors for work like data preparation and labeling, you may be able to move some of that spending onto your Google Cloud bill instead of buying compute you don’t need.
The goal isn’t to spend more. It’s to get more out of money you’re already spending on your data. This article covers how Marketplace purchases count toward your commitment, a worked example, what private offers allow, what the fees look like, and a checklist to take to your Google Cloud account team.
Why Google Cloud commitments go unused
Many companies sign spending commitments with cloud providers like Google Cloud, AWS, and Microsoft Azure. In exchange for committing to a minimum level of spending over a set period, you get better pricing on compute, storage, and other infrastructure.
Those agreements can save a lot, especially if your workloads are heavy on compute. Lower infrastructure costs go straight to your gross margin.
But plans change. Say you commit to $1 million a year on Google Cloud, and as the year draws to a close you can see you’ll spend $800,000. That leaves $200,000 of your commitment unused, and depending on your agreement, a bill for some or all of it.
Do Google Cloud Marketplace purchases count toward your commitment?
Yes, if the purchase is eligible. Google’s commitment drawdown page says up to 100% of your spend on Google Cloud Marketplace services can count toward your minimum commitment, apart from a short list of excluded categories. Your own agreement with Google Cloud decides how that applies to you.
Google has made this a selling point. Its October 2024 announcement on Marketplace private offers describes buying from independent software vendors (ISVs) as a way to draw down your Google Cloud commitment.
That changes the question. Instead of asking what else you could run on Google Cloud, you can ask which vendors you already pay directly that sell through Marketplace.
A worked example: a $1 million commitment, $200,000 short
Here is how that could look with illustrative numbers.
| Annual spending | Amount |
|---|---|
| Google Cloud commitment | $1,000,000 |
| Expected infrastructure spending | $800,000 |
| Remaining commitment | $200,000 |
| Current data labeling spending, paid directly to vendors | $500,000 |
Illustrative figures, not a customer example. What counts toward your commitment depends on your agreement with Google Cloud.
You also spend $500,000 a year on outsourced data labeling. Your labeling vendors have their own tools, workforces, and processes, and you pay them directly, outside your Google Cloud bill.
Now say you route $200,000 of that existing labeling work through an MLtwist private offer on Google Cloud Marketplace. If that purchase counts in full, it closes the $200,000 gap. Your total labeling budget doesn’t go up. You’ve moved part of it onto the bill that counts toward your commitment.
Put everything in one private offer. What’s billed through the private offer draws down your commitment. Anything a vendor invoices you directly, such as a separate statement of work, an onboarding fee, or extra labeling hours, doesn’t. Bills you’ve already paid outside Marketplace can’t be turned into Marketplace spending after the fact.
Check the timing. The commitment period your purchase counts toward depends on your agreement and on how the private offer is structured. If you pay several years up front, Google’s announcement says the drawdown is amortized annually, so a three year prepayment won’t cover a single year’s gap all at once.
Why route your data labeling through MLtwist
The commitment is only part of the case. If you work with several labeling vendors, each with its own tools and processes, it’s hard to standardize how your data is prepared, bring in AI prelabeling, check quality the same way, or combine the results into one dataset for training.
MLtwist sits across those workflows and prepares, labels, checks, and delivers your data in the format your training pipeline expects. MLtwist’s listing on Google Cloud Marketplace covers converting images, prelabeling through Google Cloud, sending work to the labeling platform of your choice, human quality control if your team wants it, and returning JSON to your storage.
How MLtwist prepares and prelabels your data
Preparation before labeling. MLtwist reads directly from your Google Cloud Storage, Amazon S3, or Azure Blob bucket. Corrupt and duplicate files are dropped, long captures are split, images are resized, large imagery is cut into tiles, and video is transcoded to formats labeling tools can play. Each step runs as a Twist, a versioned pipeline step, so your next batch is prepared exactly the same way.
Prelabels people correct. Foundation models draft labels before anyone opens a file, so your labelers correct drafts instead of starting from a blank frame. The prelabels are kept alongside each source file, so you can see what the model proposed and what a person changed. Prelabeling is optional: you can skip it and have your data labeled from scratch, or bring prelabels from your own models instead. Learn more about data preparation at MLtwist.
How MLtwist works with the labeling tools and teams you have
Keep your tools and your vendors. Moving your labeling through MLtwist doesn’t mean replacing the vendors you already use. MLtwist prepares and prelabels your data, sends it to the labeling tool your team knows, and brings the labels back for review, versioning, and delivery. The labeling itself can be done by our team, yours, or both.
One quality check and one delivery. Whichever vendor or tool did the labeling, the results go through the same built in quality control before handover, and your finished data is written to your storage or pulled through the REST API into your workflows. Every file carries a Data Card recording where it came from, which model drafted its labels, and who labeled and reviewed it. Learn more about labeling on the MLtwist platform.
What Google Cloud Marketplace private offers allow
Private offers let you and a vendor negotiate terms instead of buying a standard listing. MLtwist’s listing can only be bought through a private offer, so the price, term, and payment schedule are set around your data volumes and workflows. Google’s announcement lists these options:
| Private offer option | What Google Cloud Marketplace supports |
|---|---|
| Payment schedule | Monthly, quarterly, annual, or custom installments, including prorated payments |
| Contract length | Up to 7 years |
| Upfront payment | Up to 5 years in advance, drawn down against your commitment on an annual amortized basis |
From Google’s October 16, 2024 announcement on Marketplace private offers. Check current terms with your Google Cloud account team.
Private offers can also make buying simpler. You buy on your existing Google Cloud billing account, which can save you from setting up a new vendor from scratch. You should still review the vendor’s agreement, your security requirements, and any additional terms.
What about Google Cloud Marketplace fees?
Google takes a share of each Marketplace transaction from the vendor, not from you. Its Vendor Net Revenue Schedule, effective April 21, 2025, sets that share by deal type and total contract value:
| Deal type | Total contract value | Google’s share |
|---|---|---|
| New private offer | Under $1M | 3% |
| New private offer | $1M to under $10M | 2% |
| New private offer | $10M or more | 1.5% |
| Channel shift, migration, or native renewal | Any | 1.5% |
| Standard offer (not a private offer) | Any | 3% |
Google’s schedule lists the percentage the vendor keeps, from 97% to 98.5%; this table shows the remainder. It applies to eligible vendors and transactions, as defined on the schedule.
A channel shift, in Google’s definition, is a new private offer that replaces a contract you already have directly with the same vendor for the same product, where the workload already runs on Google Cloud. If you already buy MLtwist directly, moving that contract to Marketplace falls in the lowest fee tier.
Vendors account for these fees in their pricing. The case for routing your labeling through Marketplace is strongest when the platform also saves you money on the work itself: preparation that no longer runs on one off scripts, labelers correcting prelabels instead of drawing from scratch, and one quality check across every vendor. If those gains are bigger than the fee, you come out ahead on more than your commitment.
Managed, or in your own cloud
MLtwist can run as a managed platform, with our team doing the work for you. If you need more control over your infrastructure and data, you can run the MLtwist platform in your own cloud environment, with your own database, while it still sends work to the labeling tools you use. Either way, your files stay in your own cloud storage, encrypted at rest.
A checklist before you use Marketplace for your remaining commitment
- How much is left. Know your remaining commitment and when the period ends.
- What counts. Confirm with your Google Cloud account team how Marketplace purchases count under your agreement.
- What you already pay for. List the data preparation and labeling work you pay vendors for directly today.
- One private offer. Put everything you want to count in the same private offer, so nothing is invoiced separately.
- Timing. Check which commitment period the purchase counts toward, and how upfront payments are amortized.
- The agreement. Review the vendor’s terms and your security requirements before you sign.
Not every purchase qualifies, and invoices you’ve already paid to other vendors can’t be converted into Marketplace spending.
Explore an MLtwist private offer on Google Cloud Marketplace
If you have a Google Cloud commitment, it’s worth looking at what you already pay for outside it. Data preparation, labeling, and quality control are recurring costs, and they’re also where automation and one consistent process pay off.
MLtwist can review how your data is labeled today, find what can be automated, and set up a private offer around it, with our team, yours, or both doing the labeling. View MLtwist on Google Cloud Marketplace, or get in touch about a private offer for your own data.
This article is general information, not financial or legal advice. Marketplace eligibility, commitment drawdown, and private offer terms vary by agreement and product; confirm how they apply to you with your Google Cloud account team. The example above is illustrative. Google Cloud’s drawdown rules, private offer options, and fees are as published in the documents linked above as of May 2025.