North v3 puts AI bills and purchasing authority in the same engine

North shipped v3 on August 20th, bringing Microsoft Azure into general availability after a customer beta and putting OpenAI, Anthropic and Snowflake charges next to the underlying infrastructure bill ^[1][2][8]. The part with consequences is not the unified view: North's Autobot can purchase, renew, increase or reduce customer-owned cloud commitments automatically, which turns a reporting tool into a procurement system ^[3].

Autobot models usage each day and manages a mix of one- and three-year commitments across AWS, Google Cloud and Azure ^[4]. North says it evaluates usage patterns, expected demand and renewal windows, then adjusts commitments as demand changes ^[5]. The underlying trade is old: providers discount sustained usage, typically on one- or three-year terms, so on-demand capacity costs more while an oversized commitment converts the discount into waste ^[6]. Automating the decision does not remove that risk, it moves it. The failure mode is no longer a finance team that forgot to buy coverage, it is a model that was wrong, and a customer left paying for capacity it no longer needs ^[7].

The pricing makes the incentive legible. The Startup plan is $199 per month plus an Autobot fee of 3.5% of automated commitments; the $1,399 Premier plan cuts that fee to 1.5% ^[11]. Flexbot is billed separately as a share of savings achieved, 25% on Startup and 20% on Premier ^[12]. The announcement does not state the period the Autobot percentage is assessed over; if it matches the subscription period, the two plans break even at $60,000 of automated commitments, with Premier cheaper above that ^[1]. Either way, North's revenue scales with the volume of commitments it manages and the savings it claims to produce ^[13]. An engine paid on commitment volume is being handed the decision about commitment volume, which is a contract term, not a product detail.

The AI and data integrations are thinner than the framing suggests. They initially provide visibility and governance, with North describing them as a foundation for future optimization, and Autobot's autonomous purchasing still covers cloud commitments only ^[9]. TokenFlow, which tracks token usage, budgets and model health, is an early-beta feature ^[8]. "Cloud spend no longer stops at compute and storage," Biringer said in the announcement, listing AI models, data platforms, GPUs and multiple providers as one financial problem ^[10]. That claim is not disputed by the market: CloudZero, Vantage and Finout all advertise OpenAI cost ingestion alongside cloud spending, Finout supports Anthropic, and Vantage markets allocation, forecasting and anomaly detection for model costs ^[14]. On the automation side, nOps also sells automated commitment management across AWS, Azure and GCP ^[15].

North is a Brooklyn company founded in 2023 by Matt Biringer, who spent 12 years in datacenter technology and growth roles at Pure Storage, CDI and SHI, and Yassine Acoine, who designed cloud systems at Siemens and worked on AWS optimization at 47Lining ^[17][18].

Watch whether the AI and data integrations acquire purchasing rights or stay read-only ^[9], and whether TokenFlow leaves early beta ^[8]. Watch the new interactive simulations and planning tools, which North added so teams can compare strategies before purchases ^[16]: for a buyer, that is the approval and audit surface, and its usefulness depends on whether it records what Autobot decided and why. Ask what happens when the forecast is wrong and who eats the stranded commitment.