CoreStack Expands FinOps+ to Govern AI Spend

By: Parthu Kishen, Senior Vice President of Product

We’re excited to announce a major expansion of CoreStack’s FinOps+ product capabilities to discover, manage, and govern AI workload spend and users, including agents as a central cost alongside cloud providers. Our September release includes the launch of TokenOps, a new module within the FinOps+ product to give Finance, FinOps, and IT teams a single place to onboard AI services (including Amazon Bedrock and Anthropic), see real-time spend across major providers and models, and deploy out-of-the-box agents to gain insights and provide decision intelligence for supported AI services through conversational chats.

Over the past year, AI spend has moved from experimental budgets to board-level scrutiny as more powerful models and compute-heavy agents become substantial operating expenses. The inaugural State of Tokenomics Report from the Tokenomics Foundation shows that proving the value and ROI of AI spend is the single biggest challenge enterprises are facing.^[1]

Governing this new spend category is largely falling to FinOps teams, with 98% of FinOps teams now managing AI spend, up from just 31% two years earlier, according to the FinOps Foundation’s latest State of FinOps 2026 Report.^[2] Without visibility into AI workload spend and clear cost attribution, teams can't prove ROI or make the case for continued investment to the CEO, CFO, and board.

“We're doubling down on our investment in FinOps+ to cover AI because our customers' governance mandate has expanded,” said Parthu Kishen, Senior Vice President of Product at CoreStack. “FinOps teams are being asked to govern a spend category that didn't exist two years ago, using tools never designed to see inside a model bill. We think that's the single biggest governance risk in the enterprise today.”

With TokenOps, FinOps teams can auto-discover Amazon Bedrock and Claude API usage within a few hours. From there, they can see a granular view of costs, as TokenOps supports all models available on and supported by Bedrock, as well as agents and users that invoke those models, including agents from Agentcore. With TokenOps, FinOps teams can see which Anthropic models were purchased through the AWS Marketplace, which typically accounts for the majority of a Bedrock bill, across four separately priced token dimensions (cache write, cache read, uncached input, and output), map every dollar spent to the agent, user or API key that generated it, and understand full AI workload costs. For teams using Anthropic's platform directly, TokenOps unifies Bedrock purchased and direct Claude usage into one view, connecting agent, user, and workload costs back to the business context that matters for real showback and chargeback. Data is ingested multiple times a day to constantly report on agent runtimes, users, and components of both Amazon Bedrock and Anthropic. These new capabilities build on CoreStack’s previous releases for FinOps+, including AI token visibility as part of multicloud reporting, as well as support for cost, usage, and rate metrics for multicloud providers, including AWS, Azure, GCP, and OCI.

“Customers have told us for months that they can explain their cloud bill but not their AI bill,” said Sabapathy Arumugam, Chief Technology Officer at CoreStack. “TokenOps closes that gap by giving teams the same level of attribution for a model call that they've had for a compute instance, so they can finally tie AI spend to the outcome it produced and make the case for continued investment with real evidence.”

The new Cloud Commitment Agent gives FinOps teams one view of portfolio health, including commitments, utilization, coverage, waste, and maturity stage across AWS and Azure today, with GCP and AI commitments coming in future releases. Every recommendation is checked against what an organization already owns before it's actioned, with reasoning built on longer lookback data, cost of capital, expiry timing, and commitment laddering attached. Easily collaborate with team members on commitment analysis through shared chats and custom reports to get ahead of renewal cliffs and prevent wasted spend.

As AI adoption accelerates, enterprises need a unified cross-functional solution to see, govern, and prove the value of ever-increasing spend. CoreStack gives teams one place to see spend across every major provider, optimize and recover waste automatically, and prove ROI to justify continued investment, closing the AI governance gap before it becomes a board-level issue. Request a demo to see how CoreStack works to get started.