California Nears Passage of 24 AI Bills, Targeting Surveillance Pricing and Data Center Power Grids

Imagine logging onto a retail website and seeing a pair of shoes priced at $120, while your neighbor, browsing the exact same site from their phone, sees them for $95. Or imagine opening your monthly electricity bill to find a sudden spike, driven not by your own air conditioning use, but by a massive artificial intelligence server farm built three towns over.

For millions of consumers, the invisible mechanics of the AI boom are beginning to hit their wallets directly. These dual pressures—algorithmic price discrimination and the staggering energy costs of AI infrastructure—have moved from theoretical debates to immediate economic realities.[1][2]

That dynamic is why California lawmakers are currently finalizing a sweeping package of roughly 24 artificial intelligence bills ahead of the August 31 legislative deadline. While much of the national conversation around AI regulation has focused on existential risks, copyright disputes, or deepfakes, California's legislative slate is distinctly focused on consumer economics and physical infrastructure. The state legislature is attempting to establish firm guardrails before the technology becomes so deeply entrenched in daily life that its economic side effects are impossible to reverse.[1][6]

Two of the most consequential measures in the package target the immediate, tangible impacts of the technology. The first aims to ban "surveillance pricing," a practice where algorithms use personal data to customize prices for individual shoppers.

The second seeks to rein in the massive energy consumption of AI data centers, ensuring that regular utility ratepayers do not subsidize the grid upgrades required by tech giants. Together, these bills represent a shift in how lawmakers approach AI, treating it not just as a software innovation, but as a heavy industry with physical and financial footprints.[1][2]

To understand the surveillance pricing ban, it helps to look at how modern retail algorithms function. Retailers and third-party data brokers increasingly harvest vast amounts of consumer data, including web browsing history, physical location tracking, app usage, and past purchase behavior.

AI models process this data to generate a real-time profile of a shopper's price sensitivity. The algorithm essentially calculates the absolute maximum a specific individual is willing to pay for a good or service at a given moment, allowing the retailer to extract the highest possible margin from every single transaction.[3]

Assembly Bill 2564, authored by Assemblymember Chris Ward, explicitly prohibits businesses from using a consumer's personally identifiable information to determine these customized prices. If passed, a retailer would be legally barred from charging a user more simply because their data profile suggests they have a higher income, live in a specific zip code, or have a desperate, immediate need for the product. The legislation aims to restore a baseline level of price transparency, ensuring that the price tag on a digital storefront is the same for everyone who visits it.[1][7]

The legislation does carve out specific exemptions to avoid breaking standard retail practices. Companies can still offer uniform discounts to broadly defined groups, such as seniors, veterans, or students. They can also maintain loyalty programs where discounts are tied to membership rather than hidden algorithmic profiling. The core target of the bill is not the concept of a sale or a discount, but rather the opaque, individualized price discrimination that happens entirely without the consumer's knowledge or consent.[7]

On the physical infrastructure side, the AI boom is creating an unprecedented strain on the electrical grid. Training and operating frontier AI models requires massive clusters of specialized processors, housed in data centers that draw immense amounts of power. While traditional data centers might require 5 to 15 megawatts of electricity, modern AI facilities routinely demand 100 to 500 megawatts—enough to power hundreds of thousands of homes. This scale of energy consumption fundamentally alters the calculus for local utilities and grid operators.[4]

When a new gigawatt-scale data center connects to the grid, utility companies often have to build new transmission lines, substations, and power generation facilities to accommodate the load. Historically, the cost of these grid upgrades has been distributed across all utility customers. This means that residential ratepayers end up subsidizing the infrastructure required by highly profitable technology companies, leading to higher monthly electricity bills for everyday citizens who see none of the direct financial benefits of the AI facility.[2][5]

Senate Bill 886, introduced by State Senator Steve Padilla, addresses this economic fallout directly. The bill requires the California Public Utilities Commission to establish a new rate structure specifically for data centers with a peak demand of at least 25 megawatts. Under this tariff, the data center operators would be legally required to fully cover the costs of the transmission upgrades and new power generation they necessitate. The goal is to isolate the financial impact of the AI boom, protecting everyday Californians from footing the bill.[2][4]

A companion measure, Senate Bill 887, uses a carrot-and-stick approach to address the environmental impact of these facilities. The bill offers accelerated environmental review and permitting for data center developers who commit to using clean energy and meeting strict water-use standards. In a state where local opposition and environmental concerns have already led to municipal data center moratoriums—such as a recent permanent ban passed by the city of Monterey Park—SB 887 attempts to create a sustainable pathway for the industry to grow without overwhelming local resources.[2][8]

These economic measures are moving alongside a broader suite of AI regulations in Sacramento. The 24-bill package also includes provisions requiring digital health education in public schools, mandating clear disclosures when AI is used to digitally alter real estate promotional materials, and establishing strict worker protections against automated decision systems. The sheer volume of legislation reflects a growing consensus among state lawmakers that AI requires comprehensive, sector-specific regulation rather than a single, overarching framework that attempts to cover everything at once.[6]

As the August 31 legislative deadline approaches, the final shape of these bills remains fluid. Tech industry groups and business chambers are actively lobbying to soften the language, warning that overly aggressive regulation could drive AI investment out of the state. They argue that strict tariffs and algorithmic bans will push valuable engineering jobs and infrastructure development to states with more permissive environments, threatening California's status as the global epicenter of the technology industry.[5]

Ultimately, the outcome of California's legislative sprint will likely set a national precedent. With the federal government largely stalled on comprehensive AI regulation, California is stepping into the void, attempting to write the rules of the road for how artificial intelligence intersects with consumer privacy, utility infrastructure, and everyday commerce. Whether these bills become law or falter under industry pressure, they have already redefined the debate around what it means to regulate artificial intelligence.[5]