Interview
Sponsored
AI
VC

The energy transition’s digital layer still has holes

Allison Myers, co-founder of Buoyant Ventures, said there’s a missing link when it comes to contextual, decision-enabling data for utilities.

|
Published
March 15, 2024
Listen to the episode on:
Data center infrastructure

Photo credit: Google

Data center infrastructure

Photo credit: Google

Despite the major buzz around artificial intelligence in recent years, digital solutions for the energy transition have fallen behind in funding. In the last five years, nearly 75% of climate tech deals that exceeded $70 million funded physical solutions, according to data from Net Zero Insights.

On the one hand, digital products don’t have the same capital intensity as their physical counterparts. On the other, there’s also the reality that utilities, which are key market segments for many energy-focused products, remain reticent when it comes to implementing new digital technologies, including AI.

Ultimately though, reliable AI models require good data, and according to Allison Myers, cofounder and general partner at digital climate tech-focused firm Buoyant Ventures, that remains one of the biggest holes when it comes to the energy transition’s digital layer.

“One of the really big gaps in the space is the data to make the right decisions,” Myers told Latitude Media, pointing to optimization opportunities in data collection, aggregation, and interpretation.

The size of that gap, however, remains an open question: “I think it depends on how quickly we want to move, which somewhat depends on political and individual forces at play, and is different by region,” she added.

Gaps in foundational data span many sectors across the energy transition, Myers added. That’s one of the reasons for Buoyant’s recent investment in HData, a software startup using AI to browse energy regulatory filings.

“HData is really filling [the data gap] from the energy transition intelligence perspective,” Myers said. 

But other gaps remain; it’s not just the raw data that’s needed, she said, but rather “it’s the data in the right context that’s really the gap.”

Myers defines that “contextual data” as an aggregation of the wide variety of data points a utility, for example, would need to consider to make the right decisions about a pilot or an investment, such as where to build new distribution versus utilize behind-the-meter resources. 

And it’s not just contextualizing or aggregating that data that’s a challenge — passive data collection is another area where Myers sees a gap.

“I think we need to find new ways to find points at which you can capture data without having to redeploy all the time,” she said. Proliferation of tools like computer vision and remote sensing will be key, but costs will have to come down before they’re widely deployed. Myers pointed to Noteworthy AI — which isn’t a part of Buoyant’s portfolio — as an excellent example of “ground-truthing,” or using software to understand what and where physical assets exist.

That data, once collected, can power machine learning algorithms and other AI solutions. And having enough of the right data to train those solutions is needed, she said, because most of the modeling and planning software that currently exists and is used by utilities can’t handle the scale of data needed to make extremely complex decisions.

This is a point that was echoed by Titiaan Palazzi, the head of power and utilities at Snowflake, in a recent episode of The Carbon Copy. 

“There is a lot of time-series data coming from the grid,” he told host and Latitude Media executive editor Stephen Lacey. “Using AI for forecasting is quite challenging because every time you need to create a new prediction, you need to have the latest data. And so from an engineering perspective, it was quite complicated to do.”

For the solutions that do exist — and Myers pointed to Pearl Street Technologies, whose software helps automate grid interconnection studies, as an example — the relative newness of artificial intelligence has made some legacy grid actors hesitant to implement them.

For more on the data gaps that exist for deploying AI in the power system, listen to The Carbon Copy:

Listen to the episode on:

Next up for the digital layer

One barrier to overcome that doesn’t necessarily depend on software, Myers said, is the disconnect between the tech world and the utility world when it comes to readiness and timelines.

“The startup world is always going to say the tech is ready, but you have to remember the position of the utility here,” she said. “They are far more risk-averse than the CEO of a startup, and the tech has to work. It can’t just be ready, it has to be ready for a different type of scale.”

And the energy sector has generally been more comfortable deploying physical solutions than digital ones.

“For a lot of these physical technologies, we have the reps down, we know the steps to take, and there are a lot of stage gates that are decision points,” Myers said, adding that with digital deployment, there are fewer opportunities to double- and triple-check that all is working. “I think it tends to feel as if you’re not sure what’s going to happen the second you deploy because it goes out the door so quickly.”

In that sense, part of filling in holes in the digital layer is about developing comfort with new digital solutions, Myers said. That means that more deployments, and more software champions will be essential to push some of the existing solutions out more broadly.

“Digital pilots need to be bigger bets,” she said. “We’re starting to see that more and more, but in a way this is really about skilling — there’s a talent gap inside companies that can take these things to scale.”

No items found.
No items found.
No items found.
No items found.
No items found.