Bezos Earth Fund exec: There's something missing from the AI-climate discussion

The fund's latest challenge aims to close what it perceives as an enterprise-scale gap in activity.

July 1, 2024
Listen to the episode on:
Apple Podcast LogoSpotify Logo

Photo credit: NASA / AFP via Getty Images // Artur Widak / NurPhoto via Getty Images

Photo credit: NASA / AFP via Getty Images // Artur Widak / NurPhoto via Getty Images

In mid-June, the Bezos Earth Fund opened submissions for what it’s calling the “AI for Climate and Nature Grand Challenge.” Up for grabs is $100 million in grants, provided by Jeff Bezos’ $10 billion pot directed at combating climate change.

The first round of grants, currently accepting applications, is targeting AI-backed solutions focused on sustainable proteins, biodiversity conservation, and power grid optimization. Up to 30 proposals will be awarded seed grants, and then will be eligible to apply for additional funding.

It might seem an unusual move for a charity-adjacent organization that has previously focused on topics like conserving nature and the future of food, given that AI already seems to be the subject of almost rabid interest from investors.

A glance at recent climate tech headlines, including those on our own site, don’t paint a picture of a corner of the industry that’s in dire need of charitable funding. In 2024 already, we’ve seen one VC close a fund entirely focused on AI and the energy transition, and the AI-backed vegetation management company AiDash rake in $50 million — plus a slew of funding announcements for companies using Ai to alleviate the critical minerals crunch.

But Amen Ra Mashariki, the Bezos Earth Fund’s head of AI and data strategies, told Latitude Media that the organization sees a gap in the market, one they hope the competition can help fill. 

“A lot of people are attempting, I think clumsily, to connect the space that we’re in now to the dot-com era,” he added. But in his view, there’s a key difference: that era was defined by thousands of small organizations, garage startups, building from the bottom up.

"Hyperscalers are absolutely jumping in with both feet,” Mashariki said. “What’s missing are small organizations who say…'I want to understand how I can take advantage of AI.’ And we’re not seeing that at any scale.”

That’s according to a “landscape assessment” that the fund commissioned from Columbia University earlier this year, which dug into the ways AI can target biodiversity loss and climate change, and the key players operating in the space — and found that hyperscalers are dictating the market’s direction. From Mashariki’s perspective, the enterprise-level activity represents a hole.

“It’s these organizations on the ground that need to get in on the demand for AI on the ground floor,” he added. “And that’s what we’re funding.”

Listen to the episode on:
Apple Podcast LogoSpotify Logo
No items found.
No items found.
No items found.
No items found.

The fund’s aims

The new fund is seeking out climate- and nature-focused organizations who have an idea for using AI to scale or improve their operations, Mashariki said. It will then pair winners with one of a list of “trusted AI partners” (a group that is still in the works) to help them bring their idea to life.

It’s similar in structure to to an accelerator, he added, though for climate orgs rather than tech startups.

Within the challenge’s power grid optimization focus area, a lot of the conversation among applicants so far is on AI power forecasting systems — and particularly on how to scale them.

“We’ve seen point solutions, but what we are looking at are solutions that can be applied on the ground, that have an ability for scale,” Mashariki said. 

As the fund evaluates and compares proposals, he said, ease of scale, along with ease of adoption and implementation, are key

And energy efficiency is also a significant factor, he added. In the long run, it’s the fund’s expectation that technological innovation will catch up with the “greedy” energy needs of today’s compute power — but in the meantime it’s looking for solutions with small footprints.

“Part of what we’re judging is your ability to maximize impact and effectiveness of an AI solution while minimizing compute access and power,” he said. 

The phase one awardees will be announced in October.

No items found.