Anthropic's $40M Policy Bet, AMD's $5B Deal, IPO Stock Rules

Three separate Anthropic developments landed in the last 48 hours, spanning election-year politics, a multibillion-dollar chip supply chain, and the mechanics of the company’s eventual public offering.

Key points

• Anthropic has spent $40 million backing a nonprofit that has clashed with a Super PAC backed by OpenAI President Greg Brockman, deepening a proxy war over AI regulation ahead of the midterms.
• AMD has committed to supply Anthropic up to 2 gigawatts of Instinct MI450 GPUs and invest as much as $5 billion in the company, reigniting scrutiny over circular AI chip financing arrangements.
• Anthropic is reportedly considering a rule that would require all employees to sell shares only through preset plans after an IPO, a governance step typically associated with companies managing lock-up volatility.

How deep does the political spending go?

The $40 million figure from Business Insider is a notable escalation. This site previously covered Anthropic’s $20 million donation to an AI regulation-focused nonprofit, so the updated total suggests either additional tranches or a broader accounting of total support. The recipient nonprofit has directly clashed with a Super PAC linked to OpenAI’s Greg Brockman, meaning the two leading frontier AI labs are now financing competing political vehicles heading into the 2026 midterms.

The strategic logic is straightforward: whichever regulatory framework takes hold at the federal level will shape compliance costs, liability exposure, and ultimately the competitive order among frontier labs. Anthropic, which has consistently positioned safety and oversight as core to its brand, has reason to back groups that favor structured regulation over a lighter-touch approach preferred by some rivals. Whether $40 million buys meaningful influence in a cycle already saturated with tech money is a separate question the sources do not answer.

What does AMD’s $5B deal actually mean?

The AMD commitment, reported by FinanceFeeds, is framed as both a supply agreement and an investment. AMD would supply Anthropic up to 2 gigawatts of MI450 GPU capacity while investing up to $5 billion into the company. The source notes this revives questions about circular AI chip deals, a pattern where chipmakers invest in AI labs that then spend that capital back on chips from the same vendor.

Whether the deal is structured to avoid that circularity, or embraces it openly, is not clear from available reporting. What is clear is that 2 gigawatts of compute is a substantial commitment, and AMD is signaling it wants a meaningful position in the frontier AI supply chain currently dominated by Nvidia. For Anthropic, locking in guaranteed GPU supply at this scale matters for both training future models and supporting the commercial inference load that drives revenue. Investors watching Anthropic’s path to IPO should note that large, multi-year infrastructure commitments like this one have balance sheet implications that will eventually appear in any S-1 filing.

Why preset stock sale plans matter for an IPO

The Live Mint report on employee stock sale rules is thin on specifics but raises a meaningful governance point. Requiring employees to sell shares only through preset Rule 10b5-1-style plans (a U.S. regulatory mechanism allowing insiders to schedule sales in advance) is a way to reduce the appearance of insider trading and manage the price volatility that typically follows a lock-up expiration.

That Anthropic is reportedly thinking through this now, before filing, suggests IPO planning is concrete enough to be working through operational details rather than broad strategy. The same report notes that both Anthropic and OpenAI are racing to go public as investor appetite for AI fuels valuations. Neither company has filed publicly, and the sources provide no updated timing or valuation figures beyond what has been reported previously.

The bigger picture for investors

Taken together, the three stories reflect a company managing simultaneously on multiple fronts: political positioning, infrastructure scale-up, and IPO governance design. Each carries its own risk profile. Political spending in a contested regulatory environment can backfire if the favored framework fails or triggers backlash. Circular chip financing structures invite regulatory and accounting scrutiny. And IPO governance choices, once set, are difficult to revise after listing.

None of this constitutes investment advice, and none of the sources provide audited financials or confirmed IPO timelines. But the density of activity suggests Anthropic is not treating any of these as peripheral concerns.