Blue Bear Capital closes its third AI-energy fund

Back in 2016, when the popular embrace of artificial intelligence was still years away, Blue Bear Capital was already pitching itself as an investor in the space.

Today, the venture capital firm announced the close of its third venture and growth equity fund, Blue Bear Capital Partners III, at $160 million, with an additional $40 million in co-investment vehicles. The capital comes from institutional investors such as Rockefeller Brothers Fund, the McKnight Foundation, Zoma Capital, UBS, and WovenEarth Ventures. 

The firm invests in what partner Ernst Sack calls “machine intelligence” and applied AI across energy infrastructure, transportation and logistics, and energy-intensive industries, selecting software-heavy companies that plug the gap between tech venture capital and the energy industry. Before Blue Bear, Sack worked in energy private equity, including for nearly ten years at Riverstone Holdings — a background that he considers instrumental to understanding the pain points in energy infrastructure that can be solved with AI. 

Raising a first-time fund at the nexus between AI and energy in 2016 was challenging, Sack told Latitude Media. But having been in the market ever since means the firm now has a good track record to show investors, which is especially valuable in a challenging fundraising environment.

“[AI] has gone from the challenge that nobody thinks is real, to a moment when it was a fresh and exciting idea for the market, to now feeling like there are too many buzzwords and noise,” Sack said. “The fact that we can point to about three dozen portfolio companies that are building and deploying AI-enabled solutions in the real world with big enterprise customers who operate grids and power plants… helps differentiate buzzwords from real-world applications.” 

Blue Bear’s first fund — with a pitch-deck that mentioned AI and machine learning at least eight times in the first four pages — was in the low tens of millions. In 2021, its successor was oversubscribed at $150 million.

Among what Sack considers Blue Bear’s successful portfolio companies is Raptor Maps, a solar management platform for large-scale utility-scale solar plants. The startup raised a $5 million series A co-led by Blue Bear Capital back in 2020, and it’s now on the verge of closing a successful series C, according to Sack. Meanwhile, TruckLabs, a technology platform improving the fuel efficiency or electric range performance of trucking by around 5%, was acquired by ConMet in November 2023 after closing a successful series B raise the same year. 

Both investments are emblematic of Blue Bear’s strategy of investing in the “business model acceleration” of low-cost digital technologies that have just started being commercially viable. The types of software at that inflection point, though, have changed over the years.

“In the early days, categories like solar and wind were the best candidates because they had the lowest cost of generation, they were deployed on a large scale across the country, and needed workflow and supply chain technologies and performance technologies that software can provide,” Sack said. “Then across our funds two and three, the push for greater energy efficiency across the industrial economy became more central.” 

Pani Energy, an AI business that improves operational efficiency at wastewater treatment and desalination plants, is an example of the kinds of companies Blue Bear is now focusing on. Other recent investments include ChargerHelp!, which optimizes electric vehicle charging, and Delos, which streamlines wildfire insurance. 

Chain-of-thought reasoning 

When it comes to the rapid evolution of AI in recent years, Sack says “chain-of-thought” reasoning, or prompting, is one of its most encouraging emerging properties of the technology. The method encourages large language models to process responses in distinct steps.

“That provides the user better visibility and audibility into why the model is providing the answer, which leads to more reliable answers, but also more trust,” he said. “The more understanding there is of the process that's led to the answer, the more likely a manufacturing plant operator is to trust and be comfortable letting that model recommend and even execute operational decisions.” 

Such trust is essential for AI applications in industries like the utility sector, which is generally risk-averse and has a slow enterprise sales cycle speed. The latter, in fact, is one of the most consistent hurdles that Blue Bear’s portfolio companies face. 

“For new entrants that don't yet have any brand or professional reputation as a management team, building trust with customers can take time,” Sack said. “That will continue to be a challenge.”