Cloudflare: A Great Network at an Impossible Price

The Setup

This week offered a clean natural experiment. The momentum trade that carried the market through the first half of 2026 finally cracked, with the semiconductor names falling more than 10% in the first days of July. And in the middle of that, Cloudflare went the other way, rising about 9% in a single session to near a fresh record high around $277, lifted by an analyst upgrade and a $300 price target. When a stock rallies hardest in the week its peers break, it is worth asking what the buyer at the top is actually paying for.

Let us be clear about what this is not. It is not an argument that Cloudflare is a weak company. Cloudflare operates one of the most impressive networks in the world, sitting in front of a large share of the internet, and its positioning in front of AI traffic is real and early. The technology is not in question. This is a narrower argument, the same one this column keeps returning to in different sectors: that price and business are two different things, and Cloudflare's price has traveled a long way past what the financial statements currently support.

Cloudflare belongs to the broader debate about whether the market has paid too much for AI infrastructure, a theme we examined in Who Is Winning the AI Race and, on the software side, in The SaaS Reckoning. The question here is specific and answerable from the filings: what does Cloudflare earn, what is it worth on those earnings, and how far has the price run beyond them.

The Agentic Internet Story

The bull case is genuinely compelling and belongs up front. Cloudflare sits in front of more than 20% of all internet traffic, which gives it a vantage point no new entrant can easily replicate. As the internet shifts from human browsing to AI agents acting on our behalf, Cloudflare has positioned its network as the layer that routes, secures, and increasingly monetizes that agent traffic. Management has framed AI as the largest growth tailwind in the company's history, and the traffic data supports the framing: AI agent activity on the network has been growing far faster than ordinary traffic.

The commercial evidence is accumulating. Cloudflare reported first-quarter revenue growth of 33% and record net new annual recurring revenue, beating estimates and prompting a double-digit rally. Its Workers developer platform is being positioned as a default runtime for AI-native applications. It has launched Pay Per Use tools that let publishers charge AI crawlers for content, aiming to become the toll collector of the agentic internet. And it has attached itself to the biggest names in AI: an alignment with Nvidia's infrastructure products, a new research pilot with OpenAI to improve AI search, and a partnership with Anthropic to run Claude-powered autonomous agents on Cloudflare Workers.

In the interest of full disclosure, Anthropic makes the Claude AI that powers portions of the Wealth Engine Pro platform, so we note that relationship plainly and it is stated again in the disclaimer below. It does not change the analysis, which rests entirely on Cloudflare's public financials. Take the bull case at face value, because it is real: Cloudflare has a rare network position, a genuine AI tailwind, and marquee partners. The rest of this article is about what the market is charging for that story.

A Great Network That Loses Money

Here is the fact the narrative tends to skip: Cloudflare does not make money. Despite roughly $2 billion in annual revenue growing in the low thirties percent, the company runs a negative operating margin, around -9%, and reported operating earnings that were negative on a GAAP basis. This is not a company modestly below breakeven by accident. It has chosen to spend aggressively on growth, and by its own guidance it does not expect to reach sustained GAAP profitability until 2028.

That matters more than usual because of the balance sheet. Cloudflare carries roughly $3.5 billion of debt, and its operating earnings are not currently sufficient to cover the interest on it. A high-growth company that is not yet profitable and carries debt is entirely viable while capital is cheap and enthusiasm is high. It is far more fragile if either of those conditions changes, and the entire premise of this week's market wobble is that both can change quickly.

None of this makes Cloudflare a bad business. Plenty of great companies invested through years of losses on the way to dominance. The point is that when you buy Cloudflare today, you are not buying current earnings, because there are none. You are buying a claim on earnings that the company itself does not expect to materialize for another two years, and you are paying a price that, as the next two sections show, already assumes those future earnings arrive and then some.

27 Times Sales

Because there are no earnings to anchor to, the market values Cloudflare on revenue, and the multiple is extraordinary. At roughly $90 billion of market value against about $2 billion of revenue, Cloudflare trades near 27 times sales. For perspective, a mature, highly profitable software business is often considered richly valued in the high single digits to low teens on sales. Cloudflare is being valued at roughly double the level that would already be considered expensive for a company that actually earns money.

The other valuation lenses tell the same story. Cloudflare trades at about 56 times book value, against an industry average closer to 8 times, and independent estimates put the shares at roughly a 40% premium to their calculated intrinsic value. Every angle points the same direction: this is one of the most expensive large-cap software stocks in the market on essentially every metric that does not depend on imagining 2028 and beyond.

A multiple like this is not a description of the present. It is a forecast. It says the market expects Cloudflare to grow rapidly for many years, expand margins dramatically as it finally turns profitable, and defend its position against every hyperscaler and security vendor that would like the same agentic internet revenue. That can happen. But at 27 times sales, a great deal of it is required simply to justify today's price, which leaves very little room for the reward if it all goes right, and a long way to fall if it does not.

What the Insiders Are Doing

One data point deserves attention precisely because it cuts against the enthusiasm. Over the past three months, while analysts raised price targets and the shares climbed to records, Cloudflare insiders sold roughly $160 million of stock, with additional sales reported as recently as this week. Insider selling is not, on its own, proof of anything. Executives sell for many ordinary reasons: diversification, taxes, planned trading programs, and simple liquidity. A single sale tells you very little.

What is worth noting is the pattern and the scale. The people with the most complete view of Cloudflare's business have been consistent net sellers of a size that stands out, during exactly the period when the outside narrative was most bullish and the price was highest. It does not prove the stock is overvalued. It simply means the group with the best information was reducing exposure while the market was adding it, and that contrast belongs in any honest assessment. A governance note reinforces the theme: an advisory group publicly urged shareholders to withhold votes from two directors at the June annual meeting, a reminder that not everyone with a close view is uniformly comfortable.

What the Wealth Engine Scores Say

Before the valuation verdict, here is what the Wealth Engine Pro platform's systematic scoring shows for Cloudflare right now.

Two things need saying plainly. First, the fair value of $26.06 is aggressive, and the reason is structural: the platform's valuation models lean on earnings and cash flow, and Cloudflare has neither in positive territory yet, so the models assign it a harsh number. Do not anchor on the precise figure. Anchor on the direction and the size of the gap, which every model agrees on: Expensive. Second, the Moat score of just 3 out of 15 is striking for a company whose entire bull case is a network moat. The systematic read is that the durability of that advantage, in a field crowded with hyperscalers and security vendors, is less established than the narrative assumes.

These scores are systematic. They evaluate companies based on reported financials, balance sheet quality, moat characteristics, and valuation models. They measure what a company is today, not what it might become, and they give no credit for an agentic-internet future that has not yet reached the income statement.

In this case, the editorial thesis and the platform scores point in the same direction. The article argues the price has detached from the financials, and the systematic data, reading Weak strength against an Expensive, Bearish valuation, lands in the same place from a different angle. When both the quantitative model and the qualitative analysis flag the same gap between price and substance, that convergence is worth paying attention to. Research Cloudflare yourself on the platform and decide which signal matters more for your situation.

The Valuation Verdict

Put the pieces together and the asymmetry is stark. Cloudflare is a roughly $90 billion company that does not earn a profit, trades at 27 times sales and 56 times book, is guided to reach GAAP profitability only in 2028, and whose insiders have been heavy sellers into the rally. The price is not a claim on what Cloudflare is. It is a claim on what Cloudflare must become, and it assumes that becoming is close to certain.

Consider the reward if the story works. The most bullish target on the Street, the fresh $300 upgrade that helped drive this week's rally, implies roughly 15% upside from the current price. That is the optimistic case, and it is a modest return for owning an unprofitable company at a record valuation in a market that just demonstrated how quickly enthusiasm for exactly these names can reverse. The downside, by contrast, is not modest. A company priced at 27 times sales re-rates violently when growth decelerates even slightly, and history is full of great businesses whose stocks fell by half while the business kept growing, simply because the multiple normalized.

That is the core of the avoid case. The buyer near $265 is paying a price that already contains the good outcome, and being offered perhaps 15% if everything goes right against a much larger loss if the multiple simply returns to earth. A great network does not make that a good trade at this price.

What Could Go Wrong

This is a demanding steelman, because Cloudflare is a real leader and the AI tailwind is not imaginary. A buyer at today's price is making several credible bets.

The network position may be a durable moat after all. Sitting in front of 20% of internet traffic is genuinely hard to replicate, and if the agentic internet routes through that vantage point, Cloudflare could hold pricing power the platform's Moat score underestimates. Network effects, once established, tend to be underpriced by backward-looking models.

The AI tailwind is real and early. One analyst notes that Cloudflare traffic historically leads revenue by about three quarters, and that agentic-AI demand is inflecting now. If that relationship holds, revenue could beat and management could raise guidance repeatedly, and a growth surprise is exactly what justifies a high multiple after the fact.

Monetizing agent traffic could open a new market. The Pay Per Use tools and the partnerships with OpenAI, Nvidia, and Anthropic point toward revenue streams that barely exist today. If Cloudflare becomes the toll collector of AI traffic, its addressable market expands in a way current revenue does not capture.

Unprofitability is a choice, not a constraint. Cloudflare spends heavily because it is chasing a land grab. It could pivot toward margin whenever it chose, and the path to 2028 profitability may prove conservative if AI revenue scales faster than planned.

The thesis breaks if growth accelerates on agentic-AI demand, the moat proves durable, and Cloudflare converts its network position into high-margin AI revenue ahead of schedule. In that world the multiple is justified by growth and the stock keeps working. What the avoid thesis argues is not that this is impossible. It is that the current price already assumes it, and pays a record valuation for an unprofitable company to leave almost no margin of safety. It is the same asymmetry, a real opportunity priced as though the hard part is finished, examined in The Quantum Computing Bubble.

The Bottom Line

Cloudflare is a great company. Its network is a genuine asset, its AI positioning is real, and its growth is impressive. If you own it and believe the agentic internet runs through Cloudflare's pipes, nothing in this thesis says the business is failing, because it is not. The caution is entirely about price, and about the distance between what Cloudflare earns and what the market is charging for it.

That distance is large. A roughly $90 billion valuation on a company with no profit, at 27 times sales and 56 times book, guided to profitability only in 2028, with insiders selling $160 million into the rally and a platform read of Weak, Expensive, and Bearish. The most bullish case on the Street offers about 15% upside. The market just reminded everyone how fast the downside on these names can arrive.

That is the entire philosophy here. Wealth Engine Pro evaluates companies on what they are and what they earn, not on the most exciting version of the story during the most enthusiastic part of the cycle. The narrative says Cloudflare is the toll booth of the agentic internet. The data says it is an excellent, unprofitable network trading at one of the richest multiples in software, whose own insiders are reducing their stakes into the rally. When the gap between those two pictures is this wide, the data is the part worth trusting until the earnings arrive to prove otherwise.