The past two weeks in AI have been a rollercoaster. While Project Stargate and DeepSeek might sound like titles from a sci-fi paperback, these announcements have serious implications for the industry. We’re here to provide the facts, some interesting takes, and boil it down to how these developments will impact the insurance industry in the near and long term.

What is Stargate?

A $500 billion AI infrastructure initiative led by OpenAI, SoftBank, Oracle, NVIDIA, and Microsoft to build American data centers. Construction has already began in Texas to stand up the initial data centers, which will be used for AI model training and deployment at unprecedented scales.

Ironically, a week after this announcement, a Chinese startup announced a new AI Model, DeepSeek R1, that sent shockwaves throughout the AI community- to the tune of $1T in reduced market value across the AI sector. NVIDIA, the leading manufacturer of AI chips, experienced the largest single-day market value loss in U.S. stock market history, dropping $589 billion in market capitalization.

Why is DeepSeek such a big deal?

DeepSeek developed an open-source model for millions of dollars that matched the performance of OpenAI's latest models, which cost billions to develop. Equal performance for 1/10th the cost.

What does this mean for AI?

Open source models are now highly competitive:

DeepSeek released its models as open-source, allowing developers and researchers to access, modify, and improve them. This has driven widespread adoption and innovation without the need for massive investments to improve the model. OpenAI, ironically, is closed-source- meaning every major improvement takes longer and costs more.

Energy is the ultimate bottleneck:

AI is ultimately constrained by energy—the fundamental input cost is GPU compute, which is powered by electricity. The U.S. energy grid lacks the capacity to meet the demand required for AI expansion at this scale. If we look at energy dynamics in the US compared to China, its no surprise a Chinese company was able to deliver competitive models for a fraction of the cost.

Current US vs China Energy Disparities:

• US stopped major grid investments around 2000 during globalization/deindustrialization

• US electricity production is around 1,600 terawatt-hours compared to China's 9,000 terawatt-hours

• The US pays 1.5-3x more per kilowatt hour than China

• Cost to add new capacity in US is 2-10x higher than China's cost due to regulatory hurdles

Nuclear Power Situation:

• US nuclear capacity has remained flat for 25+ years

• China has been rapidly expanding nuclear capacity

The AI industry is shifting at breakneck speed, with massive investments in infrastructure (Project Stargate) and cost-disrupting breakthroughs (DeepSeek AI) reshaping the landscape. While these macro trends may seem distant from the world of insurance, they have direct and immediate implications for how agencies operate, compete, and grow.

What This Means for Insurance

• Massive Business Opportunity in Data Center Construction Insurance

  • The AI infrastructure boom will drive an explosion in data center construction, creating huge demand for specialized insurance policies.

  • Agencies positioned in construction, energy, and tech sectors stand to gain from this rapidly growing market.

• AI Costs Will Keep Dropping—Efficiency Gains for Agencies

  • AI-powered operations will become even more cost-effective as models get cheaper.

  • Agencies should consider open-source AI if they plan to build their own tools—or at least be strategic about vendor selection to avoid overpaying for AI solutions.

• “Model-Agnostic” AI Vendors Will Win—Be Wary of Proprietary Claims

  • The best AI vendors will integrate the most advanced models, not lock into their own outdated proprietary systems.

  • Be cautious of vendors who claim to have built their own AI—if they don’t keep pace with the rapid evolution of AI, their tools will fall behind in performance and cost efficiency.