There’s no fixed dollar amount. Google Cloud Always Free is a bundle of monthly quotas (like 1 e2-micro VM, 2 million Cloud Run requests, and 1 TB of BigQuery queries) that reset each billing cycle, which can be worth anywhere from “a small bill” to “meaningful savings” depending on what you run. For AI, Gemini API access via AI Studio is rate-limited (requests per minute, tokens per minute, and requests per day) rather than budget-limited, so the value is really “how many calls you can make” for free. If you’re building a lightweight backend plus a small demo UI, Always Free can cover hosting and basic usage for a long time. If you’re pushing heavy traffic or large data scans, you’ll hit the caps fast.

Yes for the GCP Always Free tier (billing account creation requires a credit/debit card), but Gemini API via Google AI Studio does not require a card.

The Always Free tier never expires; the quotas reset every billing cycle.

Yes. If you have Google credits you won’t use before they expire, you can list them on AI Credit Mart and sell them at up to 70% of face value. Companies regularly list surplus credits from startup programs and enterprise agreements.

AI Credit Mart has discounted Google credits available from companies with surplus allocations. Prices are typically 30-70% below retail.

Always Free quotas don’t expire, but if you exceed them on a paid billing account you are billed automatically at standard rates.

Yes. Use Google AI Studio to get an API key without a billing account or credit card, and you’ll be limited by free-tier rate limits (overages return HTTP 429 errors, not charges).

The Gemini API free tier includes daily request caps (RPD) that reset at midnight Pacific Time, plus per-minute limits like RPM and TPM. Limits are per GCP project, not per API key, which means generating extra keys won’t increase your quota. If you suddenly start hitting rate limits more than you used to, you’re not imagining it: Google reduced Gemini free-tier quotas by about 50–80% in December 2025. Also note that Gemini 2.0 Flash is scheduled to retire on March 3, 2026, so plan to migrate to Gemini 2.5 Flash or newer if you’re still on the legacy model.

They’re worth $300 in usage across most Google Cloud services for 91 days, including core GCP products like Compute Engine, BigQuery, Cloud Storage, and Vertex AI (for example, Vertex AI Studio with Gemini models). The practical value depends on what you run: you can build a real prototype, deploy an API, and test data workflows. The biggest caveat is that some high-demand resources are restricted on the trial (notably GPUs). If you need GPUs immediately, you’ll need a different path.

Yes, but it’s for identity verification only, with a temporary $0–$1 authorization hold.

The $300 Google Cloud Free Trial credits last 91 days from the day you sign up.

Yes. If you have Google credits you won’t use before they expire, you can list them on AI Credit Mart and sell them at up to 70% of face value. Companies regularly list surplus credits from startup programs and enterprise agreements.

AI Credit Mart has discounted Google credits available from companies with surplus allocations. Prices are typically 30-70% below retail.

If you don’t upgrade, your trial billing account auto-closes, workloads are stopped, and you get a 30-day grace period before resources are permanently deleted.

No. GPUs for Compute Engine VMs and GPUs on Vertex AI aren’t available during the free trial; they unlock only after you upgrade to a paid account.

Gemini API via Google AI Studio has its own permanent free tier that’s independent of the $300 trial. Vertex AI usage (like Vertex AI Studio with Gemini models) uses your GCP billing setup and can consume the $300 credits. Rate limits for the AI Studio free tier are roughly 5–15 requests per minute and up to about 1,000 requests per day depending on the model, and limits apply per Google Cloud project. Commercial use is allowed on that free tier.

Up to $5,000 (faculty, postdocs, nonprofit researchers) or up to $1,000 (PhD students, renewable annually) in Google Cloud spend on most GCP services. Because GPUs are allowed, the credits can translate into a lot of training time; Google’s example suggests roughly 1,350 hours on a T4 GPU for the $5,000 level. You can also use them for Vertex AI training/deployment, BigQuery analysis, Cloud Storage, and data pipelines. The “real” value depends on whether you use on-demand GPUs, Spot VMs, or mostly CPU and storage.

No.

Once activated, the Research Credits expire after 365 days (or earlier if you spend them all). You also have to activate them within 60 days of the project start date you provided.

Yes. If you have Google credits you won’t use before they expire, you can list them on AI Credit Mart and sell them at up to 70% of face value. Companies regularly list surplus credits from startup programs and enterprise agreements.

AI Credit Mart has discounted Google credits available from companies with surplus allocations. Prices are typically 30-70% below retail.

Running workloads will be stopped unless you have another funding source linked to the billing account.

They can’t be used for Google Maps Platform (Maps/Routes/Places APIs), paid Google Cloud Support packages, or Google Colab products (Colab Pro, Pro+, Colab Enterprise). Cryptocurrency mining is explicitly prohibited. And the credits are for academic research only, so commercial use or personal benefit is out.

Treat the 250-word proposal as the core of the application. Be concrete about the research problem, name the exact GCP services you will use, and tie your Pricing Calculator estimate to those workloads. Include a timeline with milestones and expected outcomes like publications, datasets, or open-source tools. Frankly, vague proposals (“explore cloud computing”) tend to lose to proposals that read like a real plan.

Students get up to $50 in GCP credits per course, and faculty/teaching staff get up to $100 per course, delivered as coupon codes that create a billing account with the credit. The value is straightforward: it offsets paid usage across most Google Cloud services (Compute Engine, BigQuery, Vertex AI, Cloud Storage, and many APIs). For a typical class project, that can cover a small deployment and some experimentation, especially if you lean on Free Tier where possible. If you plan to use GPUs or leave VMs running, assume the credits will vanish quickly and budget accordingly.

No.

You typically have a 16-week window from the course start date to redeem the coupon, and then the credits are valid for 12 months from redemption.

Yes. If you have Google credits you won’t use before they expire, you can list them on AI Credit Mart and sell them at up to 70% of face value. Companies regularly list surplus credits from startup programs and enterprise agreements.

AI Credit Mart has discounted Google credits available from companies with surplus allocations. Prices are typically 30-70% below retail.

When credits run out, running VMs are stopped, storage persists for about 30 days, and then it is deleted.

No. This is not a self-serve student program; a faculty member has to apply, get approved, and send you the distribution URL. If your professor has not applied, your fastest path is to ask them to do it and share the program link. If that is not possible, you will need an alternative like a different education program or a separate platform for compute.

GPU quota is not enabled by default, but you can request GPU access in the Cloud Console (IAM & Admin > Quotas). Approval can take minutes to about a week, and it is not guaranteed because it depends on availability and region. Some zones will not have the GPU type you want, so check pricing and availability before you request quota. This is a key difference from the $300 Free Trial, which blocks GPU access entirely.

They’re worth roughly 15–30 T4 GPU hours per week (dynamic) plus up to 12-hour sessions with about 12–13 GB RAM and 16 GB VRAM when a GPU is available. In practice, that’s enough for quantized 7B LLM inference, LoRA/QLoRA fine-tuning on smaller models, Stable Diffusion image generation, and Whisper transcription. The exact “value” changes with availability because GPU access is not guaranteed at peak times. If you treat it like a disposable GPU sandbox for experiments and demos, it’s one of the better free deals out there.

No. Colab’s free tier does not require a credit card or any payment method.

A Colab free-tier session can run up to about 12 hours, and GPU availability is governed by a dynamic weekly quota rather than a fixed published limit.

Yes. If you have Google credits you won’t use before they expire, you can list them on AI Credit Mart and sell them at up to 70% of face value. Companies regularly list surplus credits from startup programs and enterprise agreements.

AI Credit Mart has discounted Google credits available from companies with surplus allocations. Prices are typically 30-70% below retail.

When your free GPU quota is exhausted, Colab typically restricts you to CPU-only runtimes, and if a session ends or disconnects, the VM is reclaimed and local files are wiped.

Because the idle timeout is based on interaction with the browser tab (clicking, typing, scrolling), not on whether code is running. If you’re inactive for about 90 minutes, Colab disconnects and reclaims the VM. Plan for it. Save checkpoints to Google Drive, and don’t assume a long training job will finish unattended on free tier.

No. The VM disk is ephemeral, so you need to mount Google Drive or download files to persist them.

Up to $350,000 in Google Cloud credits over 2 years, structured as up to $250,000 of 100% coverage in Year 1 and 20% coverage up to an additional $100,000 in Year 2, plus $12,000 in Enhanced Support credits and $10,000 in partner model credits. The “worth” depends on how much of your spend is eligible and how much lands in Year 1 versus Year 2. If you’re building on Vertex AI (Gemini API calls included) and also running typical GCP infrastructure (Compute Engine, GKE, Cloud Run, BigQuery), you can burn through meaningful budget without paying cash early on. Year 2 won’t fully cover bills, so plan for real out-of-pocket spend. Treat it like runway extension, not free infinite cloud.

Yes. You must set up a Google Cloud Billing Account before you apply.

The AI Track credits run for 2 years, and the Enhanced Support credits last 1 year.

Yes. If you have Google credits you won’t use before they expire, you can list them on AI Credit Mart and sell them at up to 70% of face value. Companies regularly list surplus credits from startup programs and enterprise agreements.

AI Credit Mart has discounted Google credits available from companies with surplus allocations. Prices are typically 30-70% below retail.

After the 2-year program period, all program credits expire and you transition to standard pay-as-you-go billing. Always-free tier services continue within their monthly limits, but anything beyond that bills normally, so set billing alerts well before the end date.

Accepted funding includes institutional equity (VC), SAFE agreements, convertible notes, and verifiable token raises or blockchain foundation grants. Private equity, government innovation grants, prizes, crowdfunding, angels only (without institutional backing), and friends-and-family funding aren’t accepted for the Scale/AI tier and may push you toward the Start tier evaluation.

Expect a decision in about 3–5 business days after you submit, but new billing accounts can add another week or so and stealth funding verification can take up to about 2 weeks. If you’re approved, credits are deposited immediately into your billing account.

Up to $200,000 over two years (and up to $350,000 for AI-first startups), applied to eligible Google Cloud and Firebase usage on your billing account.

Yes. You must create a Google Cloud billing account and provide the 18-character billing account ID as part of the application.

The Scale tier runs for two years: Year 1 coverage, then a 12-month Year 2 reimbursement period, and then you switch to pay-as-you-go.

Yes. If you have Google credits you won’t use before they expire, you can list them on AI Credit Mart and sell them at up to 70% of face value. Companies regularly list surplus credits from startup programs and enterprise agreements.

AI Credit Mart has discounted Google credits available from companies with surplus allocations. Prices are typically 30-70% below retail.

Your resources keep running and you’re billed at standard pay-as-you-go rates unless you shut things down or set budgets to prevent it.

Year 2 is not a lump-sum grant. You pay for a month of usage, then Google issues credits equal to 20% of that previous month’s spend, applied to your billing account. This continues until you hit the $100,000 Year 2 cap or the 12 months end. Because it’s retroactive, cash flow planning matters, and low-spend teams will not come close to maxing the Year 2 amount.

Institutional VC equity and SAFEs qualify, and Google also lists verifiable token raises and blockchain foundation grants. Angel investment, friends-and-family rounds, private equity, government innovation grants, and prize/crowdfunding do not qualify for Scale.

Up to $2,000 in billable usage on Google Cloud and Firebase, applied to your Cloud Billing Account for 1 year. Practically, that can cover an MVP stack (Cloud Run + Cloud SQL + Storage + Firebase Auth/Hosting) or a meaningful batch of Vertex AI experimentation, depending on how heavy your workloads are. The credits can be spent across standard GCP services like Compute Engine, BigQuery, GKE, and Cloud Storage, plus Firebase services such as Firestore and Cloud Functions.

Yes. You need a Google Cloud Billing Account with a payment method (credit card or bank account) on file.

The Start tier credits expire 1 year after they’re deposited into your billing account.

Yes. If you have Google credits you won’t use before they expire, you can list them on AI Credit Mart and sell them at up to 70% of face value. Companies regularly list surplus credits from startup programs and enterprise agreements.

AI Credit Mart has discounted Google credits available from companies with surplus allocations. Prices are typically 30-70% below retail.

Your services keep running, and charges shift to the payment method on your billing account. If there is no valid payment method (or it can’t be charged), services may be suspended. Google recommends setting budget alerts in the Billing console to avoid surprise charges. The expiration happens 1 year after deposit regardless of remaining balance, so it’s on you to plan the burn.

Most rejections come down to domain mismatch: the billing admin email domain doesn’t match the company website domain, or the application is submitted from a personal Gmail address. Google also rejects companies that fall into excluded categories (agencies, consultancies, nonprofits, government entities, and similar). And sometimes you’re simply not a fit for Start because your funding counts as institutional equity.

Yes, but you need verifiable equity funding and you reapply at the same application page.

They’re worth 30 hours/week of NVIDIA GPU time plus about 20–30 hours/week of TPU v3-8 time, which is often enough for a full week of experiments or a few serious training runs. In terms of capability, you’re getting access to a P100 (16 GB) or dual T4s (32 GB total) for PyTorch/TensorFlow training, and a TPU v3-8 (128 GB HBM) for TensorFlow/JAX workloads. The real “value” comes from the environment too: pre-installed ML libraries, easy dataset attachment, and background execution for long runs. If you checkpoint well, you can push surprisingly far without paying anything.

No.

For Kaggle notebooks, GPU and TPU access runs on a rolling weekly window that resets over time rather than expiring on a single date. CPU notebooks don’t have a weekly cap, only a per-session limit.

Yes. If you have Google credits you won’t use before they expire, you can list them on AI Credit Mart and sell them at up to 70% of face value. Companies regularly list surplus credits from startup programs and enterprise agreements.

AI Credit Mart has discounted Google credits available from companies with surplus allocations. Prices are typically 30-70% below retail.

On Kaggle, you don’t get charged when your GPU/TPU quota runs out; you simply lose accelerator access until the rolling weekly quota window refreshes.

Kaggle offers NVIDIA Tesla P100 (16 GB), NVIDIA T4 x2 in beta (2 × 16 GB), and TPU v3-8 (128 GB HBM across 8 cores) as notebook accelerators once your account is phone-verified.

Use “Save Version” and choose “Save & Run All (Commit)” so the notebook runs in the background. Interactive sessions can prompt “Are you still there?” after inactivity and may terminate if you don’t confirm, which is brutal if you walked away for lunch. Commit runs still obey the session caps (9 hours for GPU/TPU and 12 hours for CPU), so save checkpoints to /kaggle/working/ and plan to resume in a new session if needed.