California’s technology policy choices rarely stay in California. That is especially true when lawmakers target the platforms, tools, and increasingly the AI-enabled services that many startups and small businesses rely on every day. That is why AB 1776, the COMPETE Act, has taken on such importance. Even after the BASED Act failed to advance on its latest vote, California’s broader antitrust agenda remains active, and the debate over these proposals still matters for startups, small businesses, and the digital ecosystems they rely on.
ACT’s April 10 webinar brought together legal, economic, and business perspectives on the bill and California’s wider antitrust agenda. The discussion surfaced a fundamental question: should policymakers address specific competitive harms through evidence-based enforcement, or should they move toward sweeping rules that presume ordinary product design and platform integration are suspect? That framing ran through the conversation with Satya Marar, Dr. Ted Bolema, Bilal Sayyed, and Ben Golombek, each of whom approached the issue from a different angle but reached a similar concern about where this policy model leads.
Innovation Under Attack: A panel of legal and economic experts discusses the potential impacts of California’s AB 1776 (COMPETE Act), SB 1074 (BASED Act), and proposed merger reforms on the tech ecosystem.
The BASED Act pointed decisively toward the latter approach. As discussed during the webinar, the bill would import key elements of the European Union’s Digital Markets Act (DMA) into California law, targeting the largest digital platforms and restricting what it characterizes as self-preferencing across ranking, bundling, defaults, data use, and certain AI-driven features. Its stated goal—fairness for third-party businesses that depend on those platforms—should be taken seriously. But the structure of the bill undermines that aim. Instead of asking first whether a feature harms competition or benefits users, it starts from the assumption that integration itself is suspect.
That matters because integration is not a side issue in digital markets. It is often how platforms reduce friction, improve trust, and create a more usable experience. Satya Marar put the point directly when he noted that self-preferencing is often good for consumers because it creates incentives for firms to vertically integrate in ways that can produce efficiencies and lower prices. That is not a defense of every platform practice. It is a reminder that product integration often reflects competition on the merits rather than a departure from it. Rankings help users find relevant results. Bundled tools lower costs. Integrated payment, fulfillment, and security systems help smaller firms reach customers without recreating costly infrastructure on their own. When policy starts from the assumption that these features are inherently suspect, it risks dismantling the practical value the features create.
The bill’s affirmative defense deepened that concern. Covered companies would bear the burden of showing that challenged conduct is reasonably necessary, narrowly tailored, and non-pretextual. Bilal Sayyed emphasized during the panel that shifting the burden this way is not a minor procedural detail. It creates substantial legal uncertainty and expense before a court ever reaches the merits. For a product team considering a new integrated feature, the question is no longer just whether the feature helps users. It becomes whether the company can defend that decision through costly litigation under an ambiguous standard. That kind of uncertainty gives new product development a wide berth, and it chills experimentation long before any final judgment.
The consequences become sharper when the bill reaches AI. The BASED Act extends to generative summaries, recommendation engines, conversational tools, and shopping assistants, all systems whose core function is to rank, predict, and prioritize outputs. Ted Bolema captured the problem when he explained that AI is not “neutral” in the way lawyers use that term. These systems are built on modeling assumptions that make them useful in the first place. A broad anti-preferencing framework maps poorly onto tools built around probabilistic relevance. Consider a generative shopping assistant that surfaces a platform’s integrated payment option because it has the strongest fraud-protection record for the user’s transaction type. Under a rule like this, that recommendation could be treated as suspect self-preferencing, even if the alternative would expose a small merchant to higher chargeback risk. As Bolema warned, this is how judges and lawyers can end up imposing design standards on technologies they are not well-positioned to specify.
The small-business lens makes this especially urgent. The BASED Act purports to target a narrow set of large companies, but its effects would radiate outward. Smaller businesses rely on platform ecosystems for customer acquisition, payments, security, logistics, discoverability, and increasingly, AI functionality. Ben Golombek described the business community’s reaction to the proposal as “hair on fire,” which is striking not because it is colorful, but because it reflects how broad the concern became across industries. He also described the downstream effect in concrete terms, warning that the bill would “break the e-commerce ecosystem as we know it” and recounting the story of a small business owner whose company depends overwhelmingly on customers finding her through Google or Amazon. That is the core policy point. When those systems become less functional, more fragmented, or harder to improve, the burden falls first on the firms least able to build alternatives.
Europe’s experience reinforces the concern. Policymakers no longer have to speculate about how DMA-style rules work in practice. As Marar noted during the webinar, the DMA has already delayed some AI-enabled features in Europe while rollouts proceeded in the United States. More broadly, ACT has pointed to growing evidence that ex-ante digital regulation can degrade functionality, create friction for users, and reshape product design in ways that do not necessarily improve outcomes for the businesses these rules are supposed to help. None of that means every concern about platform conduct is unfounded. It means lawmakers should be cautious about importing rigid frameworks without fully accounting for their downstream effects.
California’s antitrust agenda extends beyond the BASED Act. AB 1776 (COMPETE Act) would broaden single-firm liability and weaken the limiting principles that help businesses distinguish unlawful exclusion from ordinary competition, raising the risk that aggressive but lawful competitive behavior becomes actionable. ACT led a coalition of small business and innovation-focused organizations in a letter opposing the bill, warning that the expanded liability standard would fall hardest on startups and smaller firms that lack the resources to navigate open-ended litigation risk. Bilal Sayyed warned during the webinar that stripping away familiar screens and standards changes how firms think about ordinary competitive conduct. Instead of asking whether a decision makes economic sense, they begin asking whether it might trigger a competitor claim. That economic structure is anti-innovation.
The California Law Revision Commission (CLRC) merger reform proposals raise parallel concerns by making acquisitions less predictable in an ecosystem where mergers and acquisitions remain a critical pathway for startup growth and exit. Ben Golombek described that risk bluntly during the webinar, warning that shutting down acquisition pathways could become a death knell for California’s venture-backed innovation model. ACT provided comments before the Commission and submitted a letter outlining how rigid structural presumptions and untested legal standards could discourage the very investments and partnerships that drive innovation in California. Both proposals reflect the same underlying shift: less weight on predictable standards, less attention to how startups actually grow and deliver products in integrated digital ecosystems.
March 2026 CLRC Meeting: Providing perspective on California’s antitrust expansion and the importance of the acquisition ‘exit’ for small app developers. (ACT comments start at 2:26:00)
These remain live debates, even if the immediate status has changed. The BASED Act failed to advance out of the Senate Privacy, Digital Technologies, and Consumer Protection Committee after falling short of the votes needed. The COMPETE Act has passed out of the California State Assembly and is headed to the Senate side. The CLRC proposals remain active ahead of further discussion.
California does not need to choose between competition enforcement and innovation. But it does need to distinguish between conduct that harms competition and design choices that create value for users and business customers. The BASED Act’s failure in one committee does not end the debate it represents. If policymakers get that distinction wrong, the seller who loses access to integrated fraud protection, the developer who can no longer rely on a streamlined API, and the small business that depends on AI-powered discoverability will feel the consequences long before the largest platforms do.