It was a study in extremes this week for the edge computing industry. In a simultaneous collision of news, Akamai Technologies and Cloudflare—two long-time rivals in the race to power the internet’s periphery—found themselves on opposite sides of a fortunes-reversed narrative.
While Akamai announced the largest deal in its corporate history, a massive seven-year commitment from a leading AI developer, Cloudflare spent the same evening notifying 1,100 employees that their roles were being eliminated. The contrast was reflected almost instantly in the markets: Akamai’s stock surged 26 percent on Friday, while Cloudflare’s plummeted 23 percent.
For those of us who spent years in software engineering before moving into reporting, this isn’t just a story about stock tickers or headcount. It is a glimpse into the brutal transition from the “cloud era” to the “AI era.” The battle is no longer just about who can store the most data or block the most DDoS attacks; it is about who can handle the massive, latency-sensitive compute requirements of Large Language Models (LLMs) at a global scale.
The pivot point is “inference”—the process of a trained AI model actually generating a response to a user’s prompt. While training a model happens in massive, centralized GPU clusters, running that model for millions of users requires a distributed footprint to avoid the lag that kills user experience. This represents where Akamai’s legacy infrastructure is suddenly becoming its greatest asset.
The $1.8 Billion Bet on Distributed Inference
During Thursday’s first-quarter earnings call, Akamai CEO Tom Leighton revealed a seven-year, $1.8 billion deal with a frontier LLM provider. While Akamai did not officially name the partner, Bloomberg has identified the client as Anthropic, the AI safety and research company behind the Claude models.
This deal follows another $200 million agreement signed last quarter with a different, unidentified AI developer. For Akamai, this represents a successful execution of a long-term strategy to move beyond its roots as a Content Delivery Network (CDN) and become a distributed compute powerhouse.
The technical appeal for AI labs is simple: scale and proximity. Akamai operates 4,300 locations across 700 cities in 130 countries. When a user in Tokyo asks an AI a question, the response is faster if it’s processed in Tokyo rather than routed to a data center in Virginia. Leighton noted that this ability to manage complex distributed systems and maintain low latency is exactly what tipped the scales in Akamai’s favor over the “hyperscalers” (like AWS or Azure) and “neoclouds.”
From a financial perspective, the deal is structured as a consumption-based contract. In other words Akamai will realize revenue as the AI provider ramps up its use of the capacity. CFO Ed McGowan addressed concerns regarding the global chip and memory shortage, stating that the company has already secured the necessary supply chain requirements for the next seven years within the next 12 months.
Cloudflare’s Painful Pivot to ‘Agentic AI’
While Akamai celebrated, Cloudflare was in the midst of a painful realignment. The company is cutting roughly 20 percent of its workforce—about 1,100 people—in a move co-founders Matthew Prince and Michelle Zatlyn describe not as a cost-cutting measure, but as a structural necessity.
In a blog post, the founders argued that the company must be “intentional” in architecting itself for the “agentic AI era.” In the industry, agentic AI refers to systems that don’t just chat, but can independently execute tasks—like booking a flight or managing a calendar. This shift requires a fundamental change in how the network is managed and how products are built.
The financial paradox for Cloudflare is striking. The company is not failing in the traditional sense; its revenues grew 34 percent year-over-year, reaching $639.8 million in the first quarter. However, it posted a net loss of $22.9 million and expects to spend up to $150 million on severance and benefits for the departing staff.
The market’s reaction suggests a lack of confidence in Cloudflare’s current trajectory compared to Akamai’s tangible wins. Despite the stock dip, Cloudflare still maintains a market capitalization over $69 billion—more than three times that of Akamai—meaning it has a significant cushion, but the pressure to prove its AI utility is mounting.
Comparing the AI Edge Strategies
The divergence between these two companies highlights two different approaches to the AI boom. Akamai is leveraging its massive, existing physical footprint to provide the “plumbing” for inference. Cloudflare is attempting to reinvent its organizational structure to build the “intelligence” layer of the agentic web.
| Metric/Event | Akamai Technologies | Cloudflare |
|---|---|---|
| Recent Major AI Win | $1.8B deal (7-year contract) | Strategic realignment for “Agentic AI” |
| Workforce Impact | Growth/Stability | 1,100 layoffs (~20% of staff) |
| Stock Reaction (Fri) | +26% Surge | -23% Drop |
| Core AI Value Prop | Distributed Inference/Low Latency | Network Architecture for AI Agents |
The Bottom Line: Why This Matters
The stakes here extend beyond two companies. This is a battle for the “Edge.” For a decade, the trend was centralization—moving everything into a few massive data centers. AI is forcing the pendulum back toward the edge. If the goal is to have an AI agent that responds in milliseconds, the compute must happen as close to the user as possible.
Akamai’s success suggests that the “old” infrastructure—the thousands of small points of presence (PoPs) built for streaming video and website caching—is actually the perfect foundation for the next generation of AI.
Disclaimer: This article is for informational purposes only and does not constitute financial, investment, or legal advice.
The next critical checkpoint for both companies will be their upcoming quarterly filings, where investors will look for evidence of Akamai’s revenue ramp-up from the Anthropic deal and whether Cloudflare’s leaner structure begins to translate into profitability.
What do you think about the shift toward distributed AI? Does the “old guard” have the advantage in the AI race? Let us know in the comments or share this story on social media.
Related reading
