Anthropic now spends more on United States federal lobbying than Nvidia, a symbolic crossing point that shows how frontier AI labs have become some of the most aggressive political actors in technology. The timing ahead of the 2026 midterm elections and widely expected stock market debuts is not accidental and it tells us a lot about where AI policy is heading next. In the first quarter of 2026, Anthropic’s $1.6 million federal lobbying bill marked a sharp increase from the same period in 2025.
A new hierarchy in tech lobbying
Regulatory disclosures compiled by multiple outlets show that Anthropic spent about 1.97 million dollars on federal lobbying in the second quarter of 2026, a 26 percent jump from the first quarter. That pushed its quarterly total past Nvidia and close to Oracle, placing a still private AI lab alongside long established hardware and enterprise software giants in Washington spending tables.
Anthropic’s near-$2 million Q2 lobbying surge vaulted it past Nvidia and into Washington’s top tier of tech powerbrokers
Across the first half of 2026 Anthropic has already crossed 3.5 million dollars in lobbying outlay, more than the 3.1 million dollars it reported for all of 2025. Its first quarter bill, roughly 1.6 million dollars, was more than four times what it spent in the same period a year earlier.
In the second quarter Anthropic and OpenAI together put 3.17 million dollars into federal lobbying, a record for both companies and an increase of about 23 percent over the prior quarter.
Those numbers matter for two reasons. First, they move Anthropic out of the niche category of research lab and into the same lobbying league as major chipmakers and defense contractors. Second, they show that AI policy is no longer a side project for these companies. It is becoming a core strategic function on par with compute procurement and model development.
How we got here
When OpenAI and Anthropic emerged as leading generative AI labs earlier in the decade they presented themselves as research driven organizations that were cautious about deployment and wary of the worst case scenarios that powerful models could introduce.
For the first few years their policy footprint was relatively modest. Reports from 2023 and 2024 show six and seven figure federal lobbying totals that were small compared with the multibillion dollar platforms they were often critiquing.
By 2025 the picture had changed. Anthropic and OpenAI each spent several million dollars on direct federal lobbying that year, with additional hundreds of thousands of dollars going to state level efforts in California.
Politico reported that both companies were adding Washington staff, seeking government contracts from a federal information technology budget that was steering tens of billions of dollars into AI systems, and taking positions on high profile regulatory fights like proposed moratoria on certain kinds of advanced model development.
In other words, the AI labs stopped being outside critics of big tech and started behaving like big tech themselves. The 2026 data simply makes that transition impossible to ignore.
Why the money is flowing now
Several forces are converging.
First, the policy agenda has moved from broad principles to concrete rulemaking. Anthropic has concentrated its lobbying on export controls, cybersecurity requirements and formal AI safety standards, areas that reach directly into how and where frontier models can be trained and deployed. These efforts align with current global standards for AI governance, which emphasize safe governance and interoperability.
Second, there is a clear electoral calendar. The 2026 midterms are drawing large sums from across the technology sector, and AI labs face the prospect that future Congresses will revisit key issues such as liability for model outputs, data access and chip export regimes.
Getting in front of those debates now can shape the baseline from which future legislation is negotiated.
Third, expectations of initial public offerings for leading AI labs create new incentives. Once companies are public, regulatory risk is immediately priced into valuations and can affect everything from capital costs to employee compensation.
Building relationships with lawmakers and regulators before listing gives them more visibility and potentially more leverage when inevitable policy shocks arrive.
The Commerce Department scare and operational risk
Anthropic’s sharp increase in second quarter spending has been linked by industry observers to a recent Commerce Department action that temporarily took its flagship models offline, forcing customers to pause or reroute critical workloads.
That episode crystallized a simple but important reality. Oversight of frontier models is no longer only about long term safety. It can suddenly affect day to day operations, revenue streams and service reliability.
In response, Anthropic appears to be treating Washington engagement as an operational risk management function. The company is not only increasing direct lobbying but also investing in broader political advocacy and public education that can reduce the chance of abrupt, poorly understood interventions.
That is a rational move for any firm whose core product might be regulated as a strategic technology rather than as ordinary software.
Beyond lobbying: the Public First Action donations
The most striking example of this broader strategy is Anthropic’s support for Public First Action, an AI safety focused political group. In early 2026 reporting revealed a 20 million dollar contribution from Anthropic to the organization, which backs congressional candidates who favor stricter AI safety rules and more robust oversight.
Anthropic later disclosed an additional 20 million dollar donation to Public First Action, bringing total support to 40 million dollars. In its own statement the company emphasized that these funds are restricted to public education and policy work and cannot be used to finance the election of specific candidates for federal, state or local office.
That framing positions the spending as an extension of Anthropic’s mission oriented branding rather than as classic corporate campaign finance.
Even so, channeling tens of millions of dollars into a group that is active in congressional races inevitably blurs the line between issue advocacy and electoral influence. It also illustrates how frontier AI labs are willing to spend far beyond traditional lobbying to shape the ecosystem in which safety debates and regulatory design occur.
What this means for technology and business
For technology builders the most immediate implication is that AI regulation is likely to remain deeply intertwined with national security, export control and infrastructure agendas.
Anthropic’s focus areas mirror a trend in which lawmakers treat frontier models like sensitive dual use technologies rather than neutral tools. That may lead to stricter controls on access to advanced chips and model weights, more demanding security audits and tighter rules around enterprise deployment.
Large incumbents such as Nvidia now find themselves in a more complicated policy landscape. On one side they are suppliers to AI labs and beneficiaries of AI infrastructure spending. On the other side they are being outspent in Washington by customers whose primary concern is model governance and safety rather than hardware policy.
This can shift the balance of influence in debates over things like export restrictions, domestic manufacturing incentives and shared safety baselines.
Smaller AI startups and open source communities face a different challenge. As frontier labs lock in extensive contact with regulators and pour millions into shaping standards, there is a risk that compliance regimes and safety frameworks evolve in ways that are tailored to very large, capital intensive organizations.
Meeting those requirements could be far harder for lean teams or non-profit projects that lack dedicated policy staff or legal budgets.
For enterprises adopting AI systems, the upside is that more structured regulation and intense engagement from major labs can eventually produce clearer rules, stronger safety documentation and more predictable liability frameworks.
The downside is that rapid, politically driven changes in oversight such as the Commerce Department action mentioned earlier can still create abrupt service disruptions and contractual uncertainty.
Societal stakes and governance questions
From a societal standpoint Anthropic’s lobbying surge and surpassing of Nvidia underscores how quickly AI governance has become a central political issue.
What used to be a debate among researchers and ethicists is now a multimillion dollar contest among corporations, defense officials and lawmakers over who defines acceptable model behavior, access controls and risk thresholds.
There are real benefits to having technically sophisticated labs at the table. They can explain how model evaluations work, what current safety techniques can and cannot do, and where additional research is required.
They can also push for guardrails that some less safety minded actors might resist.
At the same time concentrated corporate influence raises predictable concerns. When the labs that stand to profit most from favorable rules also spend the most to shape those rules, there is ongoing tension between expertise and self-interest.
Civil society groups, academic researchers and public interest technologists will need enough resources and access to provide countervailing views, especially on issues such as transparency, competition and long term societal impacts that may not align with short term commercial incentives.
Anthropic’s large donations to Public First Action add another layer. Funding a safety oriented political group may help ensure that AI risks are taken seriously in Congress, but it also gives one company disproportionate influence over which versions of safety and responsibility gain political traction.
That is a subtle but important form of agenda setting.
Key takeaways and what to watch
- Anthropic has moved into the top tier of United States tech lobbyists, outspending Nvidia and nearly matching Oracle on federal lobbying during the second quarter of 2026.
- Its lobbying spend has grown from hundreds of thousands of dollars per quarter in 2025 to millions per quarter in 2026, with more than 3.5 million dollars already booked in the first half of the year.
- Beyond formal lobbying, Anthropic is channeling tens of millions of dollars into Public First Action and similar initiatives, signaling a long term strategy to shape both policy and public understanding of AI safety.
- The combination of heavy spending, recent regulatory scares and looming midterm elections suggests that AI governance will be set in a political environment where frontier labs, traditional tech giants, defense agencies and civil society all compete to define the rules that will govern the next phase of AI growth.
Over the coming year the most important signals to watch will be how concrete safety standards and export controls evolve, whether midterm campaigns make AI regulation a front line issue, and how much transparency companies provide about the specific policies they are pushing for.
Those details will reveal whether this new wave of lobbying leads to robust, broadly legitimate guardrails or to a framework that reflects the priorities of a handful of well funded firms.
Conclusion
Anthropic has quietly become one of the most aggressive lobbyists in artificial intelligence policy, spending more in Washington in the second quarter of 2026 than Nvidia and nearly matching Oracle as the regulation fight enters a decisive phase. That shift marks a turning point for frontier AI labs, which are no longer just technical outfits but full political actors trying to shape how their own products are governed.
A new phase in the AI policy fight
For most of the past decade, the biggest spenders on technology policy in the United States were the familiar cloud and social media giants along with hardware companies such as Nvidia. Frontier AI labs like Anthropic and OpenAI operated more like research institutions, emphasizing safety, alignment and cautious deployment and staying relatively small in Washington.
That posture has changed quickly. Anthropic began building a serious lobbying operation around 2023 and has increased its spending almost every quarter since. At the same time, lawmakers are struggling to write rules for powerful general purpose AI models, from export controls and model evaluations to liability, copyright and national security use cases.
The result is a convergence. Traditional big tech firms that care about cloud infrastructure and chip policy now share the field with younger AI labs that are directly exposed to frontier model regulation, and both sides are spending heavily to influence the outcome.
How Anthropic became a major player in Washington
Anthropic did not even report federal lobbying until 2023, when its annual outlay was just over two hundred thousand dollars. By the end of President Donald Trump second term, its monthly lobbying expenditures had risen by more than five hundred percent, reaching about 1.1 million dollars in the final quarter of 2025.
That trajectory has steepened in 2026. In the first quarter Anthropic spent roughly 1.56 million dollars on federal lobbying, more than tripling its spending from the same quarter a year earlier. In the second quarter it reported 1.97 million dollars in lobbying costs, a twenty six percent increase over the first three months of the year and its highest quarterly spend to date.
Over the first half of 2026 the company has already spent more than 3.5 million dollars on lobbying, surpassing the 3.1 million it reported for all of 2025. That places Anthropic among the more serious policy players in Washington, especially for a privately held firm that is still preparing for a public listing.
What the numbers tell us
The most striking data point is comparative. In the second quarter of 2026, Anthropic spent 1.97 million dollars on lobbying, more than Nvidia and only slightly below Oracle, which reported about 2 million dollars.
Nvidia remains a large and influential actor, with about 2.1 million dollars in federal AI lobbying recorded for the first quarter of 2026 and roughly 8 million dollars in total tracked spending. Yet the fact that a single frontier AI lab can surpass Nvidia over a quarter illustrates how quickly the lobbying landscape is tilting toward companies whose main product is AI itself, rather than chips or platforms.
Anthropic is not alone. OpenAI also boosted its spending, reporting 1.2 million dollars in the second quarter of 2026, an increase of nearly eighteen percent from the first quarter. Together, the two labs spent 3.17 million dollars during that period, up twenty three percent from the previous quarter and signaling a shift from niche advocacy to sustained political engagement.
What Anthropic is trying to influence
Anthropic reports lobbying on a wide set of issues that map directly onto frontier model deployment. Disclosures mention export controls on advanced AI systems, cybersecurity, standards for AI safety, defense procurement, and funding for research projects such as the Energy Department Genesis Mission, as well as applications in health care.
Earlier filings and reporting also show the company pushing back against proposals such as moratoria on certain AI uses, while simultaneously advocating for more structured oversight in areas it considers high risk. Anthropic has hired both Republican and Democratic aligned lobbying firms and has spent more than 5 million dollars in total on lobbyists since launching its Washington operation.
Beyond direct lobbying, a super PAC network with ties to Anthropic has become active in the 2026 election cycle, spending roughly 12 million dollars on ads supporting candidates who favor stronger AI rules and committing additional hundreds of thousands of dollars to specific congressional races. This combination of inside lobbying and outside election spending gives the company influence over both policy design and the political environment in which those policies are debated.
Why frontier labs see regulation as a strategic battlefield
For frontier AI labs, regulation is no longer just a compliance concern. It is a central strategic variable that can determine which business models are viable, which countries can access advanced systems and how quickly competitors can release their own models.
If regulators decide that only companies with certain security practices or model evaluation workflows can deploy frontier systems, that will favor labs that have invested heavily in safety infrastructure. Conversely, if policymakers opt for lighter oversight with strong protections for open source development, that could tilt the playing field toward communities and firms that are less capital intensive but more distributed.
Anthropic has consistently presented itself as safety focused, investing in interpretability research and structured deployment policies for systems like its Claude models. Strong safety standards that reflect its own practices would in effect codify many of its design choices and favor firms that adopt similar approaches. On the other hand, strict licensing regimes or export controls that are miscalibrated could slow innovation, push development offshore or consolidate power among a handful of players. Frontier labs are lobbying hard because they know the rules will lock in advantages and constraints that might last for years.
Implications for technology companies and competition
For hardware and cloud providers such as Nvidia and large platforms, Anthropic rising spend sends a clear signal. AI labs are no longer junior partners in the policy conversation. They are peers bringing their own priorities, and sometimes those priorities diverge from traditional big tech positions on issues like open source, liability and national security access.
In practical terms, more intense lobbying by Anthropic and OpenAI means that lawmakers will hear more arguments framed around frontier model safety, capability thresholds and systemic risk, rather than only around innovation, competition and intellectual property. Hardware and cloud firms will still press for favorable export rules, procurement decisions and industrial policy, but they now have to share the stage with companies whose main concern is how their own models are assessed, constrained or mandated.
This competition for influence can cut both ways. It may produce more nuanced regulation that integrates safety perspectives and operational realities from labs that understand the technology intimately. It can also increase the risk that rules are crafted in ways that entrench incumbent labs and make it harder for smaller teams, open source projects or academic groups to participate in frontier research.
Risks for democracy and the public interest
The scale of spending now involved in AI policy is substantial. Big tech companies collectively poured millions of dollars into Washington during the first half of 2026, even before counting the new wave of AI specific lobbying and aligned super PAC spending.
Public interest organizations and civil society groups often operate with a fraction of these resources, which means their ability to match technical detail and sustained engagement in complex rulemakings is limited. There is a real risk that the regulatory framework for powerful AI models is shaped primarily by those who stand to profit from them, even if many are sincerely concerned about safety and misuse.
At the same time, the presence of safety focused labs in the debate is not purely negative. When companies like Anthropic push for more robust evaluation standards, clearer incident reporting and stronger guardrails around high risk deployment, they can help lawmakers move beyond vague concern to concrete policy. The challenge is ensuring that these proposals are scrutinized not only for their safety value but also for their competitive impact and implications for individual rights.
What to watch in the next regulatory chapter
Several trends merit close attention over the next year.
First, spending levels are likely to climb further as the United States approaches the 2026 elections and as OpenAI and Anthropic move closer to public offerings, increasing scrutiny of their business models.
Second, the issues on which Anthropic and its peers lobby will reveal where they see genuine risk to society and to their own interests. If disclosure reports focus heavily on export controls and national security, that suggests concern about geopolitical escalation and model proliferation. If copyright, liability and cloud procurement dominate, the emphasis is more on monetization and market share.
Third, the relationship between direct lobbying and election spending will matter. The growing role of super PACs tied to AI firms shows that influence is not confined to regulatory agencies and committees but extends to who is in office and what their baseline attitudes toward technology are.
For businesses and citizens trying to make sense of this landscape, several takeaways stand out.
- Frontier AI labs have become central political actors, and their lobbying choices will shape the rules that govern future systems.
- Outspending hardware and legacy tech firms in key quarters signals a shift in who drives the AI policy agenda.
- Safety and regulation are now strategic levers, not just compliance tasks, and will increasingly be used to gain or defend competitive advantage.
- Without strong public interest engagement, there is a real possibility that AI rules will tilt toward the preferences of a small number of well funded companies.
The battle over AI regulation is ultimately a battle over who sets the terms of the next era of computation and decision making. Anthropic willingness to surpass Nvidia in lobbying spend is not just a line in a disclosure report. It is a signal that the frontier labs that build these systems intend to be just as present in the corridors of power as they are in research labs and data centers.
