Cloudflare’s AI Bet Is Working, but Insiders Are Selling Into It

Cloudflare (NYSE: NET) posted first-quarter 2026 revenue of $639.8 million, up 34% year-over-year, and raised its full-year outlook on May 7. The same afternoon, the company said it would cut roughly 1,100 jobs, or about 20% of its workforce, framing the reduction as a shift to an agentic AI-first operating model. The same print now carries both reactions.

Shares fell 18% to 19% in extended trading. A few weeks later, the stock was up 8.3% on news of a Claude Managed Agents integration with Anthropic. Wall Street firms lifted price targets into the $200 to $265 range. Insiders, meanwhile, kept selling.

The Q1 Print That Both Bulls and Bears Can Quote

Cloudflare’s Q1 2026 results, detailed in the company’s Q1 2026 results and guidance release, beat the Street on the top line, the bottom line, and free cash flow, per Quartz’s recap. Revenue of $639.8 million topped the $622 million Wall Street had pencilled in, and the $0.25 adjusted EPS beat the 23-cent estimate. Free cash flow reached $84.1 million, equal to 13% of revenue, up from $52.9 million, or 11% of revenue, a year earlier. Current remaining performance obligations grew 34% year-over-year. The strong print set the table for the company’s full-year guidance raise.

Profitability is the more complicated story. The GAAP net loss narrowed to $22.9 million, or $0.07 per share, from a $38.5 million loss a year earlier, while non-GAAP net income was $94.0 million. Gross margin compressed to 71.2% on a GAAP basis from 75.9%. Cash, cash equivalents, and available-for-sale securities stood at $4,163.9 million as of March 31, 2026, giving the company a war chest to fund the AI-first transition.

Guidance went higher. Cloudflare raised full-year fiscal 2026 revenue to $2,805.0 to $2,813.0 million, with non-GAAP EPS of $1.19 to $1.20, and pointed Q2 to $664.0 to $665.0 million in revenue. The Street had been pencilling in roughly $622 million for Q1, so the raise was the meaningful data point.

The 20% Layoff Framed as an Operating Model Shift

Cloudflare had 5,156 full-time employees at the end of 2025. Cutting 20% of the workforce, on a single afternoon, is not a routine reorg. The company estimates $140.0 to $150.0 million in total charges, of which $105.0 to $110.0 million is cash for severance and benefits and $35.0 to $40.0 million is non-cash equity vesting. Most of that will hit the second quarter, and the company expects the action to be substantially complete by the end of the third quarter.

Management did not pitch the move as cost-cutting. In a letter to staff, co-founder and CEO Matthew Prince and co-founder and President Michelle Zatlyn said the workforce reduction was about how Cloudflare intends to operate in a new era, not about individual performance. They pointed to internal AI tool adoption as the operational reason a leaner headcount could carry the same workload. A layoff letter to staff published after the print made the framing explicit: the cut is a one-time event, not a rolling program, and the company is being rebuilt around an agentic model in which AI does work previously done by staff. Internal AI tool adoption rose more than 600% in the quarter, by the company’s own count.

Cloudflare did not disclose a productivity figure, an output-per-headcount number, or a revenue-per-employee projection to back the new operating model. The 20% cut is the largest single workforce move Cloudflare has ever made, and the speed of it leaves no room to phase in a transition. The next two quarters of operating data will show whether the AI rationale holds up.

The Q2 print is the next test of the operating model. Investors have until then to decide which version of the story they believe.

Today’s actions are not a cost-cutting exercise or an assessment of individuals’ performance; they are about Cloudflare defining how a world-class, high-growth company operates and creates value in the agentic AI era.

Matthew Prince and Michelle Zatlyn, in their letter to Cloudflare staff, May 7, 2026.

Why the Stock Dropped 18% in One Night

The print was strong. The reaction was not. Shares fell roughly 18% to 19% in the extended session on May 7, according to Quartz’s recap. The catalyst was the workforce cut and the implication it carried: if 20% of the workforce can be eliminated in one afternoon with no obvious product regression, the market has to ask what that says about the productivity of the rest of the company. A 34% revenue beat and 600% AI tool adoption gain were not enough to offset the workforce cut.

The shape of the week tells the story. Five events defined the move, and the market is still working through the third. Piper Sandler and BTIG both raised targets the day after the print, the rebound to 8.3% in late May added the Claude Managed Agents integration, and analysts are now trying to figure out which signal to weight more. The post-earnings period isn’t over, with fresh analyst updates still landing in the days that follow.

1. May 7, 2026, after market close: Cloudflare reports Q1 FY2026 results, raises full-year guidance, and announces the 20% workforce reduction.
2. May 7, 2026, extended trading: NET falls roughly 18% to 19%.
3. May 8, 2026: Piper Sandler sets a $250 target; BTIG sets a $243 target.
4. May 11, 2026: Susquehanna sets a $200 target.
5. May 28, 2026: NET jumps 8.3% on AI-agent infrastructure momentum, anchored by the Claude Managed Agents integration with Anthropic.

Analyst Targets Say Buy, With Caveats Attached

The post-earnings analyst round was not unanimous, but it skewed constructive. A running tally of insider sales and analyst targets tracks 14 price targets on NET in the last six months with a median of $237.5. The most recent calls were a mix of upward revisions and fresh price targets, with the sell-side cohort still carrying a Buy consensus in the wake of the print.

The caveats are in the spread. The high end of the recent cohort sits in the $260s and the low end below $200, with most calls clustered in the $220 to $250 range. The full roster is below. The dispersion leaves a stock that closed at $228.48 on June 12, 2026 with a median target 3.9% above the last close. The dispersion is wider than usual for a name that just reported a beat-and-raise quarter with layoffs.

BTIG and Piper Sandler issued new targets within a day of the print, while Susquehanna set a $200 target that sits at the low end of the recent cohort. Mizuho, Citigroup, Jefferies, and RBC all carried targets in the $225 to $265 range from earlier in the year.

Insiders Have Been Selling the Whole Way Up

While analysts were setting targets in the $200 to $265 range, the people running the company were trimming. According to the same running tally cited above, co-founder and CEO Matthew Prince has made 152 open-market sales over the past six months, totalling 942,912 shares for an estimated $188,642,236, with zero purchases. Co-founder and President Michelle Zatlyn made 82 sales covering 480,416 shares, an estimated $94,368,574. CFO Thomas Seifert made 49 sales covering 91,898 shares, an estimated $17,671,550.

The pattern holds across the C-suite: 338 insider transactions in six months, all sales. A separate twelve-month insider transaction history notes that the single largest sale in the period was a $13 million block by Prince at $226 a share, close to the recent trading range. Insiders still hold about 10% of the company, valued around $7.2 billion, and they have not bought a single share on the open market in the past six months.

• CEO Matthew Prince: 152 sales, 0 buys, 942,912 shares, ~$188,642,236 over 6 months
• President Michelle Zatlyn: 82 sales, 0 buys, 480,416 shares, ~$94,368,574 over 6 months
• CFO Thomas Seifert: 49 sales, 0 buys, 91,898 shares, ~$17,671,550 over 6 months
• All insider trades, last 6 months: 338 sales, 0 buys
• Insider ownership of NET: ~10% (~$7.2 billion at recent prices)
• Largest single insider sale in 12 months: $13 million by Prince at $226 per share

The AI-Agent Infrastructure Bet Behind the Move

The operating model story is downstream of a product story, and the product story is the part of the thesis the layoffs were made to defend. Cloudflare ran Agents Week 2026 from April 12 to 17, positioning itself as the substrate for AI agents running on the open Internet. A math-and-economics post on scaling agents from Cloudflare’s blog laid out the case. If more than 100 million US knowledge workers each use an agentic assistant at 15% concurrency, the system needs capacity for roughly 24 million simultaneous sessions. At 25 to 50 users per CPU, that is 500,000 to 1,000,000 server CPUs, just for the US, before scaling to the rest of the world.

Cloudflare’s bet is that its Workers platform, which is built on V8 isolates rather than containers, is the right unit of compute for that world. Isolates, the company wrote, are 100x faster, up to 100x more memory-efficient than a container, and the Dynamic Workers open beta lets them be spun up at runtime in milliseconds. Coding agents need full containers, and Cloudflare is keeping that product alive. For the long tail of agents that just need to make API calls and return a result, the company is positioning isolates as the per-request primitive. Per-agent cost falls as a result.

The Anthropic deal is the first visible customer. The Quiverquant piece on the 8.3% rebound flagged the launch of Cloudflare Environments for Claude Managed Agents, which lets Anthropic’s Claude agents run inside sandboxes backed by Cloudflare’s network and Workers. The deal positions Cloudflare as a viable alternative to hyperscalers for production-grade agent infrastructure.

Whether that thesis holds depends on whether the agent economy routes through Cloudflare or through the cloud incumbents. The 600% surge in internal AI tool adoption is suggestive but not proof. The Q1 revenue beat proves the platform is selling today. The two are different statements.

AI is driving a fundamental re-platforming of the Internet and a paradigm shift in how software is created and consumed; it’s shaping up to be the biggest tailwind we’ve ever seen in Cloudflare’s history.

Matthew Prince, co-founder and CEO of Cloudflare, in the Q1 2026 results release.

The New Investment Narrative, in Three Claims

Strip the AI framing away and the new investment narrative has three load-bearing claims. First, 34% revenue growth in Q1, plus raised full-year guidance, plus 13% free-cash-flow margin prove Cloudflare is a real growth company at scale. Second, operating leverage will come from a leaner headcount and from AI being used internally, not from a product mix shift. Third, Cloudflare’s network, Workers, and now Claude Managed Agents make the company a structural beneficiary of the agent economy. The recent stack forming around centralized AI’s gaps is a parallel bet on distributed compute, and Cloudflare’s share of agent compute is the open question for the next year.

The bear case is that the layoffs are a cost cut with AI branding, and the analyst optimism is a holdover from the pre-print narrative. The 18% post-earnings drop and the 338-to-zero insider sales-to-purchases ratio give the bear case real weight. The median analyst target of $237.5 sits 3.9% above the June 12 close of $228.48, which is not the kind of upside that compensates for execution risk on a workforce reset.

Q2 is the first print that will test the thesis. Restructuring charges of $105 to $110 million in cash will land in the quarter, dragging reported earnings even as the underlying business grows. Whether the operating margin moves higher, whether the workforce reduction is absorbed without a product or sales-comp regression, and whether insider selling continues are the three things to watch. The Q2 revenue guide of $664 to $665 million gives analysts a fixed point, and a swing of the agent stack toward non-developer knowledge workers would broaden the compute base in the direction Cloudflare’s narrative requires.

Disclaimer: This article is commentary on publicly reported financial events and is for informational purposes only. It is not financial advice. Investment decisions carry risk, including the loss of principal, and figures cited are accurate as of the publication date. Consult a qualified financial professional before acting on any of the views expressed.