AI infrastructure cost management startup PointFive wins $60 million to help companies reduce costs

FinOps startup PointFive Inc. today announced that it has raised $60 million from investors betting that companies can curb the soaring costs of operating artificial intelligence infrastructure.

Accel led the round, with participation from venture capital firms including Index Ventures, Entre Capital, Perpetual Growth, Vesey Ventures, Sheva Ventures, and Salesforce Ventures.

New York-based PointFive has created what it calls an “AI and cloud efficiency platform” that allows businesses to visualize, control and optimize costs associated with cloud-based computing infrastructure. This round brings the total raised to date to $96 million and validation to $500 million, just halfway to unicorn status.

The name PointFive reflects the company’s claim that it can cut customers’ cloud spending in half. This is done by scanning the cloud environment being used to uncover wasteful spending on things like idle servers and unused storage resources.

It focuses specifically on AI infrastructure, where it says many organizations waste significant amounts of money on memory resources. Many companies try to ensure that their AI models have enough context available to them so that they can respond accurately. However, too much context in an AI model can lead to problems such as slow response times and inflated costs.

PointFive CEO Aaron Arvatz (pictured, center) told the Wall Street Journal that an “always-on” AI agent is also a huge financial waste.

PointFive’s platform integrates with customers’ cloud environments and runs in the background to help guide cloud engineers. Arvatz explained that it scans your complete infrastructure configuration and recommends ways to optimize it, reducing costs without impacting performance. For example, depending on the nature of the task, a customer might suggest using a lower-cost AI model for some applications. “We are their efficiency coach,” he explained, adding that some large companies spend millions of dollars a year on unnecessary resources.

Accel partner Philippe Bottelli told the Journal that the cost of running AI applications has become one of the biggest expenses for many organizations. As they encourage their employees to adopt AI tools to get more done, they are facing increasing bills for computing, networking, and storage resource usage. This practice, which Botteri calls “token maxing,” causes costs to quickly rise as companies run out of tokens, the units that track the consumption of AI resources.

Meta Platforms appears to be one of the biggest companies to learn this the hard way. In a widely publicized memo to Meta employees in April, the company’s chief technology officer, Andrew Bosworth, mandated that “we shouldn’t use AI tools just for the sake of using AI tools. Every move is not progress, and token usage alone does not measure any kind of impact.”

Arvatz founded PointFive in 2023 with cybersecurity veterans Gal Ben David (left) and Amir Hozez (right). They previously co-founded and sold a company called IntSights Cyber ​​Intelligence Ltd., which was acquired by Rapid7 Inc. in 2021 for approximately $335 million. It was during the time when InSights’ technology was being integrated with Rapid7’s that they realized the scale of wasted cloud spending.

The company says its services are attracting a lot of attention. The company didn’t give numbers, but said its annual recurring revenue increased sixfold last year, much of it from existing customers who doubled their spending on average. Albats said they are spending more money to save more money and believes the company is on track to increase its revenue fivefold this year.

These customers include global companies such as German utility company E.ON GmbH, Brazilian neobank Nubank, and US-based Fanatics Inc., which sells sporting goods and collectibles and operates a gambling platform. According to Arvatz, Nubank was able to recoup the money spent on its services within just 10 days thanks to the cost savings of cloud infrastructure.

Arbats said the company will use the funding from today’s round to expand into Europe and Israel. The company plans to increase its workforce, adding about 40 new employees to its marketing and research and development teams this year. However, the company had originally planned to hire up to 80 new employees, but the adoption of its own AI led to scaling back those plans. “Our unique use of AI allows us to avoid the problem that many companies currently face: having to ‘right-size’ after overly aggressive hiring,” Arvatz explained.

The startup also plans to launch new services later this year, including TokenShift, which debuted today and helps customers track and control usage of internal AI tools.

Featured photo: Eyal Marils