California: A Technology Overview

AI Moves From Model and Software Narrative to New Reality of Digital Infrastructure

California remains the center of gravity for technology transactions in the United States, but the market’s defining shift is that AI is no longer just a model and software story. It is increasingly an infrastructure business. Capital, talent and model quality still matter, but so do chips, cloud capacity, data-center buildouts, network infrastructure, power availability, and long-term commercialization channels. That shift is changing both the structure and the substance of technology deals. A defining characteristic of this new normal is that transactions now require deep integration across multiple disciplines – M&A, credit and leveraged finance, real estate, energy and project finance, and technology transactions must work together seamlessly to structure and execute these complex, multifaceted deals.

Firms with established strength across each of these practices, and a culture of cross-practice collaboration, such as Sullivan & Cromwell, are particularly well positioned to deliver results. The leading transactions are no longer merely acquisitions, divestitures or ordinary software partnerships; they are also strategic financings, compute arrangements, custom silicon collaborations, hyperscale supply agreements, real estate and power development, and commercialization alliances designed to secure scarce capacity and accelerate deployment. In California, where many of the market’s most important AI companies, investors, chip providers and strategic counterparties are based, that shift has been especially pronounced.

Landmark AI Transactions Show How California Deals Are Being Built Around Compute, Commercialization and Scale

That shift has kept activity high. Over the past year, some of the most consequential California technology transactions have reflected the same core market priorities: securing access to compute, supporting commercialization, and positioning businesses to scale. Recent examples on which Sullivan & Cromwell advised include:

• OpenAI’s October 2025 agreement with Microsoft, which preserves and builds on key elements of the parties’ landmark partnership, including OpenAI’s purchase of USD250 billion of Azure services;
• OpenAI’s February 2026 strategic partnership with Amazon, including OpenAI’s commitment to consume 2 gigawatts of Trainium capacity; and
• Broadcom’s April 2026 expanded strategic collaboration with Google and Anthropic, under which Anthropic, beginning in 2027, will access through Broadcom approximately 3.5 gigawatts of next-generation TPU-based AI compute capacity.

These transactions underscore that the most strategically significant AI transactions now extend well beyond traditional software licensing or ordinary procurement models.

California Remains the Control Room, Even as Infrastructure Spreads Outward

California remains where many AI strategies are conceived, financed and negotiated, even as the infrastructure required to support them is increasingly built out elsewhere. Silicon Valley continues to serve as a commercial and innovation hub, but the largest deployments often depend on access to power, land and permitting conditions that are easier to secure outside California, though the parties and deals originating in California remain critical.

Infrastructure Demands Are Reshaping Key Technology Deal Terms

The infrastructure-focused shift is changing the nature of technology transactions. Terms that might once have been peripheral in a conventional software deal – long-term capacity reservations, phased deployment rights, ramp schedules, energy strategy, resiliency planning and upgrade paths – are now often central. So too are provisions addressing supply constraints, changing technology roadmaps, and reliance on a small number of critical providers. In practical terms, AI companies and their counterparties are negotiating for capacity, time, certainty and optionality as much as price. In a market where delay can be as costly as spending, those provisions can determine whether a business is able to scale.

Commercialization Partnerships Are Becoming as Important as M&A

In many ways, these deals have become at least as important as traditional M&A in shaping competitive landscapes. The most valuable AI partnerships now solve multiple problems at once: securing the hardware supply chain at the chip and rack-level, locking in power and data center capacity, allocating supply chain risks, establishing compute services, enabling distribution of AI solutions, and – through it all – sourcing massive amounts of financing. Strategic collaborations have therefore become the driving force of the growth of an entire ecosystem of technology and infrastructure market participants. They help determine how products reach the market, how priority and exclusivity rights operate, and how parties preserve room to expand into adjacent offerings as technology develops.  

This is one reason the current market cannot be understood through a conventional lens that separates financing, supply and commercialization into silos. A cloud relationship may also be a distribution arrangement. A chip or infrastructure commitment may also support a long-term product roadmap. A financing may accompany strategic rights and market positioning that matter well beyond the financing economics. Similar dynamics are emerging in semiconductor and custom compute arrangements, where the traditional lines between supply agreement, platform partnership, and ecosystem strategy continue to blur. At Sullivan & Cromwell, we have embraced these challenges by bringing together technology, financing and infrastructure practices to pioneer the leading transactions in this market. 

The Next Phase Of California AI Dealmaking Will Be Bigger, More Integrated and More Operational

AI-driven deal activity is expected to remain robust, but the center of gravity has shifted toward large-scale financings, infrastructure commitments and commercialization partnerships. The companies best positioned to lead will be those that can secure compute, power, supply continuity and market access while maintaining flexibility to evolve their products and business models. That is why the leading California transactions increasingly look less like traditional technology deals and more like multidisciplinary, integrated strategies for building, supplying and commercializing AI at scale. The past year’s headline transactions suggest that this model is not a passing feature of the market. We look forward to continuing to partner with leading California companies to meet the challenges of this dynamic market.