Neal's Deals (Vol. 78) - Power Surge ?: The Rising Energy Costs of the AI Revolution ????

Neal's Deals (Vol. 78) - Power Surge ?: The Rising Energy Costs of the AI Revolution ????

Hey everyone – Yesterday, some of the world’s biggest tech leaders, including Nvidia CEO Jensen Huang, OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, and Google President Ruth Porat, gathered at the White House to discuss energy and infrastructure. Why was energy consumption on the agenda for both tech leaders and politicians? The answer lies in the immense power demands of AI. On average, a ChatGPT query consumes nearly 10 times more electricity than a Google search. This signals a major shift in how the world will consume and pay for power in the future.

For years, data centers have managed to keep their energy use relatively stable, even as workloads increased. But now, with the AI revolution accelerating and efficiency gains in electricity use slowing, Goldman Sachs Research predicts that data center power demand could soar by 160% by 2030.

As AI development ramps up, so do concerns about its strain on America’s aging power grid, especially as the Biden administration pushes to transition away from fossil fuels. While AI holds the potential to tackle big issues like climate change and cancer, it also brings complex challenges—chief among them, how to meet its massive energy requirements.

In this edition of Neal’s Deals, let’s dive deeper into AI’s energy demands, how the White House is responding, and what this means for the tech and venture world moving forward.


Energy demands:

Over the past decade, U.S. power demand has remained flat despite population growth and increased economic activity, thanks to efficiencies like LED lighting reducing energy use. However, this is about to change. Projections suggest a 2.4% increase in power demand between 2022 and 2030, with 0.9 percentage points of that driven by data centers. This surge, reminiscent of power demand spikes from the early 2000s, will be fueled by factors like electrification, industrial reshoring, and AI. By 2030, data centers are expected to consume 8% of U.S. power, up from 3% in 2022.

To meet this rising demand, U.S. utilities will need to invest around $50 billion in new generation capacity just to support data centers. Additionally, analysts anticipate that increased data center energy consumption will drive around 3.3 billion cubic feet per day of new natural gas demand by 2030, necessitating the construction of additional pipeline capacity.

The response:

Following yesterday’s meeting, the White House announced the formation of an interagency task force to promote data center development in the U.S. and expedite permitting for those facilities.

Additionally, the Energy Department will now guide data center owners and operators toward loans, grants, and tax credits to help them access clean and reliable energy sources. OpenAI, plans to invest tens of billions of dollars in domestic AI infrastructure, including data centers, energy capacity, transmission, and semiconductor manufacturing, with funding from global investors. The company has been in talks with government officials for months regarding the national security implications of foreign capital in these projects, while also highlighting the potential for creating 40,000 jobs across several states. OpenAI pointed to China’s AI investments, as it aims to lead the global AI race by the end of the decade, further underscoring the importance of strong U.S. energy infrastructure for maintaining competitiveness in the field.


Investments in data center solutions:

It's well known that the private sector typically moves faster and spots trends before the public sector, and this has been evident in the growing focus on AI energy demand. Companies have been building and investing in this space for some time.

Notably, there’s been a surge in software and hardware innovations aimed at developing new materials for computer chips and optimizing energy consumption in existing hardware. Investors of all sizes are also actively seeking out companies deploying cooling solutions for data centers.

In early-stage venture, investing in the infrastructure behind AI has been popular for some time, but the solutions addressing the growing energy demands of AI are some of the most compelling opportunities out there. If you know anyone building in this space, feel free to reach out!

Let’s get to it:

Brico, a San Francisco startup that automates the financial licensing process for businesses, particularly in areas like money transmitter and credit licenses, raised a $8 million seed round led by Pear Ventures.

Why this is interesting: Operating a financial services business in the U.S. requires obtaining and maintaining various licenses, a process that is often manual and expensive. Brico simplifies this with an automated license management tool that handles the full lifecycle—from applying for licenses to maintaining compliance and meeting regulatory requirements. It provides a centralized repository for licensing information, streamlines reporting across all 50 states, and offers AI-powered automation to handle forms, documents, and regulatory submissions. The company is certainly addressing a clear need, but it will be interesting to see how they evolve the product. I'm curious about their future product roadmap, as it currently feels more like a feature than a fully developed core offering.

Portex, a San Francisco startup that helps SMBs manage shipping tasks, communicate with freight partners, and automate tasks like price quotes and freight tracking, raised a $6.25 million seed round led by Footwork Ventures.

Why this is interesting: Portex founder Brittany Ennix's interest in supply chains started in childhood, inspired by her grandfather's work at a Ford glass manufacturing plant. Her passion deepened during her time at Uber and Flexport, which ultimately led her to create Portex. The company aims to make supply chains more efficient for small and mid-sized businesses (SMBs) by offering a more accessible alternative to traditional, costly Transport Management Systems (TMS). Portex connects SMBs with freight partners, streamlines shipments, and reduces manual processes and freight costs. The founder-market fit is strong, and it would be interesting to explore the market composition of freight forwarders and any trends in consolidation to assess the venture scalability within this focused customer segment.

LineLeap, a New York startup whose app?allows users to purchase passes for instant entry, VIP access, or other perks at partnered bars, clubs, and popular events,?raised a $10 million round?led by Y?Combinator.

Why this is interesting: Although this raise is larger than what I typically cover in the newsletter, I wanted to highlight LineLeap because I've followed the company since my college days. LineLeap allows users to skip lines at bars via its app, focusing on nightlife and entertainment. Initially targeting college bars, the platform has grown to over 1 million users and 400+ venue partnerships, generating revenue through convenience fees and data-driven marketing. While growth was slowed by COVID and the challenge of signing up venue owners, it's also what sets them apart. I do wonder about the company's long-term future—could it be acquired, and by whom? Or might it go public? There's little precedent for outcomes like this in the space, so it's hard to say. Curious how investors underwrote this outcome.


Deals in the Works:?If you want to learn more - feel free to reach out

  1. Marketplace for industrial components
  2. Marketplace for commercial office space
  3. Compliance automation for the credit industry
  4. Trust and estate settlement sold as a corporate wellness benefit
  5. Modern inventory and manufacturing software

_____

Quote of the week:

“Life is not about waiting for the storms to pass. It’s about learning how to dance in the rain.”

Vivian Greene

_____

Have a great weekend everyone!

要查看或添加评论,请登录