Morgan Stanley upgrades Cloudflare: what agent traffic reveals about who controls the infrastructure of the next internet
On June 9, 2026, Cloudflare held its annual Investor Day. In ceremonial terms, it was one more event of the kind that technology companies organize to update their projections and reassert confidence among investors. In structural terms, it was something else entirely. Morgan Stanley understood it that way: it raised its price target on Cloudflare's stock (NYSE: NET) from 245 to 305 dollars, maintaining its overweight rating. It is not the move itself that matters, but the logic that underpins it and what that logic says about how the architecture of the internet is changing.
The data point that anchors the entire analysis is this: artificial intelligence agent requests over Cloudflare's network grew 1,700% between June 2025 and May 2026. Eleven months. This is not a projection, but traffic measured across real infrastructure. And the company itself maintains that the moment when traffic generated by automated agents surpasses human traffic is no longer on the horizon, but "fast-approaching," according to the Investor Day presentation. The company estimates that the total volume of internet traffic could multiply between 10 and 100 times over the next five years as agentic systems proliferate.
When a company that handles more than 20% of all global network traffic makes a claim of that magnitude, it is not speculating about trends: it is describing what is already flowing through its own cables.
---
The change in condition that Morgan Stanley decided to price in
The analytical question is not whether Cloudflare is going to grow. It is whether the nature of that growth implies a structural reordering of the industry or simply more volume on the same business model.
Morgan Stanley bet on the former. And the mechanics of its valuation make that clear: the price target of 305 dollars is based on a multiple of 48 times the projected free cash flow for 2034, discounted at a cost of capital rate of 11%. This is not a short-term valuation based on quarterly results. It is a bet on the position that Cloudflare will occupy in network infrastructure within a decade, when AI agents are the dominant actors of internet traffic.
That distinction matters because it changes the type of risk the investor is absorbing. It is not operational risk in the conventional sense. It is thesis risk: if the transition toward an internet dominated by automated agents occurs as Cloudflare projects, the company holds a network position, a catalog of 75 revenue-generating products, and a global routing capability that no competitor can replicate in the short term. If the transition is slower or more fragmented, the valuation has a severe time-horizon problem.
What Cloudflare presented at its Investor Day was, in that sense, an argument about material conditions. Not about product vision, but about network physics: more than 20% of global traffic already passes through its infrastructure. That is not built in two years. And in an environment where AI agents need to execute workflows with low latency, high computational efficiency, and integrated security, holding that network position is the most difficult barrier to entry to overcome.
The financial commitments the company established for 2027 reinforce the thesis: a "Rule of 50+", which implies growing revenue by more than 30% while sustaining an operating margin of approximately 20%. In infrastructure software, that crossover point between growth and profitability is what separates companies that scale from those that simply spend more to grow more. Cloudflare also raised its long-term targets: operating margins above 30% and free cash flow margins in the range of 30 to 35%, both above prior targets. And it set a concrete milestone: reaching a revenue run rate of 5 billion dollars before the end of 2028, starting from a guidance of approximately 2.809 billion for all of 2026.
---
What the developer market is already confirming
While the agentic thesis is, by definition, a bet on the future, there are present-day metrics that give it operational substance. In 2025, the annual recurring revenue of Cloudflare's developer platform grew 137%. The developer base on the platform grew from approximately 3 million to more than 5.5 million in a single year.
This is not trivial. Development platforms have an adoption logic that operates differently from that of enterprise products. Developers do not purchase through sales cycles: they adopt based on utility, they experiment, and when a platform becomes embedded in their workflows, the cost of switching becomes real. The jump from 3 to 5.5 million developers in 12 months indicates that Cloudflare is gaining ground as construction infrastructure, not just as a distribution or security layer.
On the other hand, Cloudflare One, the company's unified access security and zero-trust platform, recorded growth of 43% in annual recurring revenue during 2025. This growth occurs in a context where enterprises are integrating artificial intelligence workloads into their operations and need security frameworks that understand agent traffic, not just human traffic. The convergence of those two trends — agentic development and enterprise security — is precisely where Cloudflare wants to position itself as a single management layer.
The results of the first quarter of 2026 support the direction: 639.8 million dollars in revenue, with 34% year-over-year growth, and free cash flow of 84.1 million dollars, equivalent to 13% of revenue. Remaining performance obligations also grew 34% year-over-year, indicating that the portfolio of future contracts is maintaining the same pace as recognized revenue. Profitability is not yet at the levels that the long-term targets require, but the trajectory is consistent with the commitments established at the Investor Day.
The total addressable market that Cloudflare estimates for itself reaches 238 billion dollars in 2026 and is projected at 384 billion dollars for 2029, with a compound annual growth rate of 17%. These are large numbers, but the composition matters as much as the size: application services, network security, and developer platforms are segments that have historically been in the hands of different providers. Cloudflare's bet is that the transition toward an internet dominated by agents makes that fragmentation costly for customers, and that a unified platform with network scale is more efficient than the sum of specialized solutions.
---
The concentration point and the friction the model has not yet resolved
No analysis of this thesis can ignore the structural risk that Morgan Stanley itself flagged: the scheduled departure of the company's chief revenue officer at the end of 2026. At a moment when Cloudflare is trying to accelerate enterprise penetration, scale relatively new products such as Cloudflare One, and convert the growth of its developer platform into higher-value revenue, continuity of leadership in the commercial area is not an administrative detail.
Large enterprise accounts are built on relationships, accumulated trust, and long sales cycles. A leadership change at that level, at that specific moment in the growth cycle, introduces a variable that discounted cash flow models do not capture well. Morgan Stanley acknowledged this explicitly. The risk is not that the technical thesis fails, but that commercial execution loses momentum precisely when the targets demand that it accelerate.
There is a second tension worth naming. The valuation of 305 dollars is anchored in projected cash flows for 2034. That implies that whoever buys into this thesis is absorbing eight years of uncertainty in an industry that is being reconfigured. The 1,700% growth in agent traffic is real, but moving from being the infrastructure through which that traffic flows to being the provider that efficiently monetizes that traffic at scale requires that the value proposition remain differentiated as AWS, Google Cloud, and Microsoft Azure also build their own agentic capabilities. Cloudflare's network advantage is real today. Whether it remains sufficiently differentiated by 2030 depends on product, pricing, and architecture decisions that have not yet been made.
The stock has accumulated a return of 15.55% so far this year and 27.02% over the past twelve months, outperforming the S&P 500 index across both time horizons. That already discounts part of the thesis. What remains to be discounted is whether the transition toward the agentic internet occurs with the speed and scale the company projects, and whether Cloudflare has the execution discipline to monetize it before the competition establishes its own control points over the same infrastructure.
---
Traffic concentration as a political asset
There is a dimension to this case that conventional financial analyses tend to undervalue. When a company handles more than 20% of global network traffic, and that traffic is migrating toward automated agents executing workflows with real economic consequences, the position of that company ceases to be purely technical or commercial. It becomes a concentration point with implications over who observes, who can interrupt, and who sets the conditions of access to the infrastructure of the internet that is coming.
That is not an argument against Cloudflare. It is an argument about the nature of the asset that Morgan Stanley is valuing. Network advantages at this scale are not only economic: they are positional. And positions of that kind, historically, generate two things simultaneously: superior returns over long periods and regulatory attention that can change the rules of the game before the model has finished maturing.
The transition toward an internet where agents generate more traffic than humans is not just an engineering change. It is a change in who has visibility over information flows, computational capacity, and automated decisions that move value. Cloudflare is well positioned in that change. Morgan Stanley decided to price it in. The next chapter is not written by the 2026 Investor Day, but by the company's ability to maintain that position when the largest players in the market decide that renting it out no longer serves their interests.
