Fastly Insider Bergman Files Form 144 and Form 4 Simultaneously

Artur Bergman sold 32,181 shares of Fastly on May 18, 2026, at an average price of $16.85, for a total of $542,249.85 [1]. Two days later, on May 20, he filed both a Form 144 and a Form 4 within 90 seconds of each other [2]. The CFO sold shares the same night [3]. That is two senior insiders, same company, same evening. The supply pressure is directional and documented. The only variable still missing is the exact share count from the May 20 EDGAR filing.

This essay argues one thing: a dual Form 144 and Form 4 filing from a control person is a concrete, actionable signal for anyone holding FSLY as an AI infrastructure or edge-compute proxy. The magnitude is unconfirmed. The direction is not. And the pattern across multiple tech names on May 20 makes this more than a single-company story.

The Signal: What Happened on May 20

A Form 144 is a notice of intent to sell shares under SEC Rule 144. Rule 144 governs how insiders and control persons, meaning people who hold enough shares to influence the company, can sell restricted or control securities into the public market. A Form 4 is the disclosure that follows when the transaction actually happens. Both forms are public filings on EDGAR.

When both land on the same night, within 90 seconds of each other, the sequence is unambiguous [2]. The Form 144 precedes the Form 4. That is the standard Rule 144 resale protocol. The insider files intent, then files the resulting ownership change. You do not need to guess at motivation. The regulatory structure tells you what happened.

Bergman is Fastly's founder and Executive Chairperson [4]. He is a control person by definition. That classification matters because Rule 144 imposes specific conditions on how control persons can sell, including volume limits and manner-of-sale requirements. The dual filing is not unusual for someone in his position. What is worth noting is the timing: May 20 was a busy night for insider filings across the tech sector. I flagged simultaneous Form 4s at Astera Labs from three insiders within a 35-minute window the same evening, and a Form 144 from Fortinet's CFO on the same date [5]. One filing is a data point. A cluster across names is a pattern worth tracking.

The specific share count and proceeds from the May 20 Bergman filing are not confirmed in the summary data available at time of writing. The EDGAR filing contains the full numbers. Do not size a position around this signal until those are pulled and verified.

Pattern and Prior Evidence

This is not a one-off. The documented record shows a consistent pattern of dispositions by Bergman across early 2026.

On February 23, 2026, Bergman sold 40,000 shares of Fastly Class A Common Stock for approximately $683,199 [6]. On March 2, 2026, a Form 144 filing recorded a disposition of 274,174 shares for $5,638,448.50 [7]. On May 18, 2026, he sold 32,181 shares at $16.85 per share, totaling $542,249.85 [1]. The May 20 filing adds another data point to that sequence.

Prior filings reference 10b5-1 plans [7]. A 10b5-1 plan is a pre-scheduled selling program. An insider sets it up weeks or months in advance, before they have access to any material non-public information. The plan specifies the price, volume, and timing of future sales. Once the plan is in place, the insider does not make active decisions about when to sell. The sales execute automatically according to the schedule.

That context matters for interpretation. A sale under a 10b5-1 plan is not a reaction to something happening at the company right now. It was decided earlier. The SEC introduced stricter rules on 10b5-1 plans in 2023, including mandatory cooling-off periods between plan adoption and first sale, to reduce the potential for abuse. Those rules make the plans more credible as genuine pre-commitment devices.

So what does a pattern of 10b5-1 sales tell you? It tells you that Bergman has been systematically reducing his FSLY position across multiple months. As of early May 2026, his estimated beneficial ownership was approximately 3.7 million shares [8]. The sales documented so far represent a meaningful but not catastrophic reduction relative to that total. The May 20 filing will clarify whether the pace is accelerating.

The CFO selling on the same night adds a second data point [3]. Two senior insiders, same evening, both with documented prior sales. That is supply pressure from multiple directions. Even if both sales are plan-driven, the market has to absorb the shares.

Why Fastly Specifically Matters for Capital Markets Operators

Fastly operates at the edge of content delivery and AI workload infrastructure. Its compute platform handles traffic routing, security filtering, and increasingly, inference workloads at the network edge. Some funds hold FSLY as a proxy for AI infrastructure exposure, specifically the edge-compute layer that sits between centralized data centers and end users.

That framing matters for how you read insider dispositions. A founder selling shares at a company held purely for its balance sheet is one thing. A founder selling shares at a company held as a structural AI infrastructure bet is a different signal. It does not mean the thesis is broken. Founders sell for many reasons, including estate planning, trust distributions, and diversification. But it is information about how the person with the most context on the company is valuing their own equity right now.

Fastly's compute revenues have been cited in recent coverage as showing growth tied to AI demand [9]. If that narrative is driving the stock, insider dispositions at this price level are worth factoring into your position sizing, even if the sales are pre-scheduled.

The broader context from May 20 reinforces this. The Astera Labs cluster, the Fortinet CFO sale, and now the Fastly dual filing all landed on the same evening [5]. Astera Labs is a data center interconnect company. Fortinet is network security infrastructure. Fastly is edge compute. These are not random names. They are all companies that trade, at least partially, on AI infrastructure narratives. A single night of concentrated insider selling across that cluster is worth building into your signal stack as a systematic input, not just a one-off alert.

The Bear Case and Why It Does Not Change the Directional Read

Skeptics will argue that 10b5-1 plan sales are noise. The plan was set up months ago. The insider had no view on current conditions when the schedule was established. The sale tells you nothing about what the insider thinks today. This argument has real merit. The SEC's enhanced 10b5-1 rules exist precisely because pre-scheduled plans, when properly constructed, are supposed to be information-free. A sale under a compliant plan is closer to a passive index rebalance than an active sell decision.

The rebuttal is simple: even plan-driven sales create real supply. Bergman's documented dispositions from February through May 2026 total well over $6 million in proceeds across confirmed transactions [1][6][7]. That is shares entering the market regardless of the motivation behind the schedule. Supply pressure is supply pressure. The question for a portfolio manager is not whether Bergman is bearish today. The question is whether the stock can absorb the volume without price impact. That is a mechanical question, not a psychological one.

Who Should Care

If you are a portfolio manager with FSLY exposure: confirm the share count and proceeds from the May 20 EDGAR filing before making a sizing decision. The supply pressure is real. The magnitude is the variable you are missing. Pull the filing, calculate the shares sold as a percentage of Bergman's total beneficial ownership of approximately 3.7 million shares [8], and decide whether the pace of reduction changes your conviction on the position.

If you are an AI infrastructure allocator: insider dispositions at edge-compute names are a relevant input to your signal stack. A founder selling under a scheduled plan does not mean the thesis is broken. But a pattern of sales across multiple months, combined with a CFO selling on the same night, is worth weighting alongside your fundamental view on Fastly's compute revenue trajectory [9]. Build the insider filing cadence into your monitoring process as a systematic variable, not an ad hoc check.

If you are a treasury manager or family office with tech equity exposure: the May 20 cluster of insider filings across Fastly, Fortinet, and Astera Labs is worth a systematic review of your positions [5]. One filing is noise. Three companies, same night, all in the AI infrastructure adjacency, is a prompt to look harder at your edge-compute and network infrastructure holdings. The question is not whether to sell. The question is whether your position sizing still reflects your actual conviction given the new information.

What to Watch Next

First, pull the EDGAR filing for the confirmed share count and proceeds from the May 20 Bergman transaction. That number determines whether this is a routine trim or a material disposition relative to his total beneficial ownership. A sale of under 50,000 shares is consistent with the cadence of prior months. A sale materially above that range changes the read.

Second, watch FSLY price action over the next two to three sessions. If the stock absorbs the supply without notable weakness, the market is already pricing the 10b5-1 plan cadence and the signal has limited incremental weight. If the stock shows sustained selling pressure following the filing disclosure, the supply impact is larger than the plan narrative suggests.

Third, watch for any amendment or termination of Bergman's 10b5-1 plan in the next Form 144 or Form 4 filing. Execution under an existing plan is a low-order signal. A change to the plan itself, whether an acceleration, a pause, or a termination, is a higher-order signal. Under the SEC's current 10b5-1 rules, plan modifications require a new cooling-off period, which means any change would be visible in the filing record before new sales could execute. That is the filing to watch.