Anthropic Eyes Mandatory 10b5-1 Plans for All Employees—Why the Real Story Is the S-1 Around It

Reuters reported on July 23, 2026 that Anthropic is debating whether to require every employee—not only executives and finance/legal staff—to sell shares after its IPO through Rule 10b5-1 trading plans. Plans of this kind pre-set the timing, amount, and price of stock sales so that trades look automated rather than discretionary, which is the standard defense against insider-trading allegations. Anthropic's own status, not valuation, makes this unusual: most issuers limit 10b5-1 use to Section 16 officers, and broader adoption is rare even at the largest tech companies.

The rule itself is well-trodden. Under the SEC's 2022 amendments to Rule 10b5-1, plans adopted by directors and Section 16 officers require a cooling-off period of at least 90 days (no more than 120) before any trades can occur; all other persons are subject to a 30-day cooling-off. Single-trade plans are capped at one every 12 months, and insiders must certify they are not in possession of material non-public information at adoption. Anthropic's twist is to push rank-and-file employees—who normally sell during post-earnings trading windows under company policy—into the same preset framework that executives already use.

Anthropic filed a confidential draft S-1 with the SEC on June 1, 2026, just over a month after closing a $65B Series H on May 28, 2026 at a $965B post-money valuation. The Series H was led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, and co-led by Capital Group, Coatue, D1 Capital Partners, GIC, Iconiq Capital, and XN; Amazon also rolled in previously committed capital. Inside five weeks, Anthropic went from raising private capital to beginning the SEC review that ends in a public listing.

The disclosure timing is what makes the policy decision load-bearing. The S-1 will publish first-day selling limits and lockup length alongside the 10b5-1 framework. Together they define the float profile: how many shares trade on day one, when the post-IPO supply wall lands, and whether employee selling is pre-scheduled or discretionary. Anthropic's revenue trajectory—$9B at year-end 2025, $30B in April 2026, $47B by mid-May—gives underwriters unusually high elasticity on deal size, but it does not solve the orderly-distribution question.

Anthropic runs Claude on three chip platforms—AWS Trainium, Google TPUs, and Nvidia GPUs—and the company has spent close to or above 100% of revenue on AWS compute at points through 2025. The April 2026 expansion with Google and Broadcom added multiple gigawatts of next-generation TPU capacity starting in 2027, and the October 2025 announcement committed access to as many as one million TPUs and well over a gigawatt of capacity coming online in 2026. Amazon's $30B deal announced in November 2025 set the multi-year commercial floor; a separate April 20, 2026 update added up to an additional $20B from Amazon on top of the existing $13B. These contracts are the inputs to Anthropic's gross margin, and a clean IPO prints them on the page.

Read through the listed peers, a clean Anthropic print translates into three things: confirmation that the hyperscaler-AI lab revenue model scales (Amazon and Alphabet), validation that custom-silicon demand is real beyond internal Google use (Broadcom), and incremental demand visibility on the memory bottleneck (SK Hynix, Micron). Nvidia is the indirect beneficiary: Anthropic's published commitments to Nvidia GPUs in the S-1 are a forward indicator of cloud-provider capex for the next four quarters.

Three structural mechanics interact. Lockup length sets when employee and insider shares can first be sold; first-day selling caps set how much existing holders can unload when the stock opens; and 10b5-1 plans determine the cadence of any sales after lockup. Mandating 10b5-1 for all employees removes one source of post-IPO volatility: large blocks of rank-and-file sales concentrated in post-earnings windows. It does not eliminate selling pressure, but it does turn that pressure into a slow, pre-committed drip rather than a series of lumpy decisions.

The 10b5-1 question is one part of an established IPO design menu. Recent tech listings have used staggered lockup releases, blackout pull-forwards, and price-triggered early unlocks: Snowflake released 25% of locked-up shares at day 91, DoorDash released 40%, Braze released 20% at day 50, and Datadog-style structures conditioned early release on the stock trading 20–50% above the IPO price for 10 of any 15 days. The Airbnb, Robinhood, and Unity playbook added day-one liquidity for non-executive employees—15% of shares available from listing. Meta's 2012 IPO used Zuckerberg's explicit no-10b5-1 posture as a signal of long-term commitment; the same page lets other executives adopt plans quietly via Form 4.

Anthropic's combination—a $965B valuation, ~$30B+ revenue, and a 10b5-1 mandate extending to all staff—is not in the playbook. Snowflake, the closest data-platform comparable, IPO'd at a $33B valuation on $592M of FY2021 revenue with no such employee mandate. The closest labor-of-labor analog is the post-IPO RSU-to-cash pressure at DoorDash, which proved that without preset schedules, employee selling concentrates in predictable post-earnings windows.

In May 2026, Anthropic widened its ban on gray-market secondary transfers: SPVs, forward contracts, and tokenized products tied to its private shares were declared potentially void, and publicly traded funds with synthetic Anthropic exposure sold off on the news. A mandatory 10b5-1 policy for employees after listing is the post-IPO mirror of that pre-IPO posture. Together they close the gray-market channels on both ends of the listing event—private trading in, and discretionary trading out.

The window from now to the first 180 days post-IPO carries the bulk of the signal. Public S-1 amendments will publish first-day sell caps and lockup length; the IPO prospectus will name bookrunners and the percentage of shares offered; the first earnings release will overlap the lockup window. Investors looking at the listed peers should track two specific milestones: the date the public S-1 becomes effective (which locks the 10b5-1 language into the insider trading policy exhibit) and the date of the first post-IPO 10-Q (which is when the first scheduled employee sales under 10b5-1 plans will show up in Form 4 filings).

Structural risk is concentrated in two places. First, a public listing reveals employee concentration data and equity overhang that private rounds obscured, which can suppress multiples even with strong growth. Second, mandating 10b5-1 for all employees effectively makes the company co-manage every worker's personal financial planning—an unusual employer posture in any industry and one that competitors can either match or use as a recruiting wedge.