A media service that sells live sports, a retailer that lives or dies by checkout speed, and a software company that exposes public APIs all face the same uncomfortable question when traffic starts to matter: is edge control worth paying for? The cheap answer is to push static files through the most bundled content-delivery option already attached to a cloud contract. The expensive answer is to buy an edge platform that lets engineering teams change cache behavior, inspect requests, run code close to users, and bolt security controls onto the same traffic path. Fastly has spent its public-company life arguing for the second answer. Its problem is not that the argument is empty. The problem is that the argument must show up in gross margin, retention, security attach rates, and free cash flow before customers and investors treat programmable edge delivery as more than a premium cache.
The directory row tied to this article is narrow but useful. ARIN's public Whois record for handle FIEE-1 names "FASTLY INC-FASTLY INC. EU ENA", gives a Plano, Texas address, records the handle on December 10, 2019, and identifies the country field as United States (https://whois.arin.net/rest/org/FIEE-1). ARIN RDAP exposes the same entity handle and linked network records, including IPv6 assignments under AT&T parent space (https://rdap.arin.net/registry/entity/FIEE-1). That record is not, by itself, proof of headquarters, revenue, ownership, or the current scope of Fastly's commercial platform. It is better read as a registry artifact around network-resource usage. The commercial question must therefore be answered from filings, investor materials, network records, product pages, and market evidence, not from the directory name alone.
That distinction matters because Fastly is a very different company from the clipped ARIN name. It is a public edge cloud and security vendor whose core economic promise is that customers will pay for a higher-control traffic layer. The latest full-year filing shows the scale of the platform and the limits of the model. Fastly reported 2025 revenue of $624.0 million, up 15 percent from 2024; network services supplied $477.8 million of that total, security supplied $125.1 million, and other products supplied $21.1 million (https://www.sec.gov/Archives/edgar/data/1517413/000151741326000053/fsly-20251231.htm). The same filing shows 2025 gross profit of $356.2 million against $267.8 million of cost of revenue, or a GAAP gross margin of about 57.1 percent. This is the spine of the story. Fastly is large enough to be a real internet infrastructure vendor, but still small enough that traffic mix, customer mix, and security expansion can move the margin line visibly.
By the first quarter of 2026, the picture had improved but not settled. Fastly reported $173.0 million of Q1 revenue, up 21 percent year over year; network services were $126.2 million, security was $38.8 million, and other products were $8.0 million (https://investors.fastly.com/news-releases/news-release-details/fastly-announces-record-first-quarter-2026-financial-results). GAAP gross margin rose to 62.5 percent in that quarter, and remaining performance obligations reached $369.0 million. The same release said the company had 634 enterprise customers, a trailing-12-month net retention rate of 113 percent, and 34 percent of revenue from its top ten customers. Those numbers give Fastly's thesis its best current support: the business is not merely selling commodity delivery by the gigabyte; it is retaining and expanding larger customers. They also show the pressure point: a small number of customers can still shape the economics of a network that must be built ahead of peak demand.
The cache hit is the easy part; the margin is in everything around it
Content delivery looks simple from the outside. A user asks for a file, the edge answers from cache, latency drops, and the origin is protected from load. The economic reality is harsher. A cache hit is only profitable if the vendor buys capacity well, places servers where demand justifies them, keeps utilization high, reduces transit leakage through peering, and sells enough control plane value to avoid a pure bandwidth auction. Fastly's own public pricing illustrates why. Its pricing page lists bandwidth, request, compute, observability, and security dimensions rather than a single flat subscription (https://www.fastly.com/pricing). That structure is rational because the product consumes real network and compute resources, but it also tells customers exactly where to shop if their use case is simple.
The hard-number version of that margin test is visible in the 2025 filing. Network services were still roughly three quarters of Fastly's revenue in 2025, while security was one fifth and other products were only a low-single-digit share (https://www.sec.gov/Archives/edgar/data/1517413/000151741326000053/fsly-20251231.htm). That means the company remains exposed to delivery economics even as it sells itself as a broader edge cloud. It cannot simply declare a transition to security or compute and be done. It must keep the delivery network relevant, protect price realization, and use the traffic relationship to pull more software into the account. The economic prize is not a faster CDN alone. It is a traffic position from which Fastly can charge for application security, bot defense, observability, edge logic, and operational confidence.
The network record gives the operating side of that argument. Fastly's public network map advertises 578 Tbps of connected global edge capacity (https://www.fastly.com/network-map). PeeringDB lists Fastly, Inc. as AS54113, describes its traffic as "100+Tbps", marks the ratio as heavy outbound, and shows a global scope, 152 internet exchange points, and 109 facilities in the API response viewed on July 4, 2026 (https://www.peeringdb.com/api/net?asn=54113 and https://www.peeringdb.com/asn/54113). Fastly's own peering page asks potential peers to exchange traffic with AS54113 and frames peering as a way to improve performance and cost efficiency (https://www.fastly.com/peering). The result is a business where technical architecture and finance are fused. Better peering is not a network engineer's private win; it is a gross-margin lever.
This is why the ARIN/RDAP evidence is relevant but not decisive. The FIEE-1 record and associated RDAP network references show that the directory entry sits in the real world of internet-number administration (https://rdap.arin.net/registry/entity/FIEE-1). They do not tell a reader whether Fastly can turn edge traffic into durable economics. For that, the useful evidence is the combination of registry traces, PeeringDB footprint, Fastly's published capacity, and financial results. When those sources are read together, the company looks like a scaled, technically credible edge network that is still proving that scale can compound into stronger margin rather than merely higher traffic volume.
Developer preference is not the same thing as pricing power
Fastly's appeal has long been strongest among engineering teams that want more direct control over how traffic behaves. Its product pages emphasize programmable delivery, instant purge, observability, and compute at the edge (https://www.fastly.com/products/edge-compute). The Compute documentation describes applications that run on Fastly's edge using WebAssembly, with developers deploying logic close to users rather than sending every decision back to origin infrastructure (https://docs.fastly.com/products/compute). That is a serious architectural proposition. It lets a customer move authentication decisions, personalization, request shaping, redirects, image handling, and API behavior closer to the request path.
The question is whether that architectural proposition creates enough willingness to pay. A developer who loves precise cache control can still be overruled by a procurement team comparing bandwidth bills. A platform team may value Fastly's configuration model and real-time logging, while a finance team asks why a bundled cloud CDN is not good enough. Public community discussions are useful here only as signals, not as measurements. Hacker News threads around Fastly product moves tend to surface respect for the engineering culture alongside questions about price and market positioning (https://news.ycombinator.com/item?id=30611460). Reddit and practitioner forums periodically frame the choice between Fastly, Cloudflare, and hyperscaler CDN services as a trade-off among control, support, security, and cost rather than a simple feature comparison (https://www.reddit.com/r/devops/comments/18ctfth/cloudflare_vs_fastly_for_cdn_and_waf/). Those are not audited facts. They are market texture consistent with Fastly's financial task.
The pricing page makes the same point in harder form. For a large customer, a low cache-miss ratio, favorable peering, and predictable traffic shape can make a premium CDN justifiable; for a simpler site, the customer can benchmark bandwidth and request charges against AWS CloudFront (https://aws.amazon.com/cloudfront/pricing/), Google Cloud CDN (https://cloud.google.com/cdn/pricing), Cloudflare, or Akamai. Fastly's control plane must therefore defend price by reducing operational pain, enabling software behavior that alternatives cannot easily match, or making outages and attacks less likely. A pure speed claim is too fragile. The internet is full of fast-enough delivery options.
This also explains why Fastly's smaller "other" revenue line matters more than its current size suggests. Other products produced $21.1 million in 2025 and $8.0 million in Q1 2026, a small base relative to network services (https://investors.fastly.com/news-releases/news-release-details/fastly-announces-record-first-quarter-2026-financial-results). But compute at the edge is strategically important because it gives Fastly a way to sell something other than delivery volume. If customers run decision logic, API mediation, security response, or application-specific code at Fastly's edge, the vendor's value becomes more embedded than a cache contract. The danger is that compute also consumes resources. If sold poorly, it can add complexity faster than it adds margin.
Fastly's economics are therefore less like a software-only SaaS company and more like a high-control infrastructure business trying to earn software-like attachment. The network has to be physically and commercially present where demand appears; the product has to be programmable enough to win developers; the sales motion has to reach security and platform owners; and the pricing model has to preserve margin when traffic spikes. That is a difficult combination. It is also the reason Fastly remains strategically interesting despite being much smaller than the largest internet infrastructure platforms.
Security is the attach-rate story, not a side business
Fastly's security business is the clearest path to better economics because it sits directly on the traffic layer the company already handles. The acquisition that changed the shape of the company was Signal Sciences. Fastly announced in 2020 that it would acquire Signal Sciences, a web application and API security company, in a deal valued at about $775 million (https://www.fastly.com/press/press-releases/fastly-announces-agreement-acquire-signal-sciences). The SEC-filed transaction release described the consideration as approximately $200 million in cash and $575 million in Fastly stock, plus retention grants (https://www.sec.gov/Archives/edgar/data/1517413/000151741320000194/projectgriffon-pressre.htm). Fastly later announced completion of the acquisition (https://www.fastly.com/press/press-releases/fastly-fastly-completes-acquisition-signal-sciences-acquisition-of-signal-sciences).
The acquisition price set a high bar. Paying that much for security only works if the product becomes part of the core edge sale, improves retention, or raises average account value. The current financials show progress but not a finished transformation. Security revenue was $125.1 million in 2025, up from $103.5 million in 2024, and Q1 2026 security revenue was $38.8 million, up 47 percent year over year (https://investors.fastly.com/news-releases/news-release-details/fastly-announces-record-first-quarter-2026-financial-results). That growth matters because security software usually carries a different customer conversation from delivery. Buyers are not simply asking how cheaply a file can move. They are asking whether a vendor can reduce application abuse, protect APIs, stop automated traffic, and give teams usable evidence during incidents.
Fastly's product pages reflect that move. The company markets a next-generation web application firewall, API protection, bot management, DDoS mitigation, and related application security capabilities alongside delivery and compute (https://www.fastly.com/products/web-application-firewall and https://www.fastly.com/products/bot-management). The more those products are sold into existing delivery accounts, the more Fastly can convert a traffic relationship into a security relationship. That is the cleanest margin story in the company: a customer that already trusts the edge path may be easier to upsell than a new security buyer reached cold.
There is a catch. Security markets are crowded, and buyers expect integration across identity, cloud workload, endpoint, SIEM, and application stacks. Cloudflare sells security as part of a much broader connectivity and zero-trust platform. Akamai has a long security franchise attached to a larger delivery network. Hyperscalers offer native security controls in the same procurement envelope as compute and storage. Fastly's advantage must be operational closeness to high-volume application traffic, not generic security language. If security becomes a feature-matrix purchase, larger platforms can bundle aggressively. If it is about fine-grained protection at the traffic path with developer control, Fastly has a more specific case.
This is where Q1 2026 is encouraging but not conclusive. A 47 percent year-over-year security revenue increase gives Fastly a credible expansion story; a 62.5 percent GAAP gross margin in the same quarter suggests the mix and efficiency can improve (https://investors.fastly.com/news-releases/news-release-details/fastly-announces-record-first-quarter-2026-financial-results). But one quarter cannot settle whether the Signal Sciences bet has paid off at the scale implied by its acquisition price. The important watchpoint is not simply security revenue growth. It is whether security growth lifts net retention, reduces delivery price pressure, and lets Fastly sell higher-value contracts without adding proportionate network cost.
Outage memory keeps the premium honest
Fastly's premium positioning depends on trust. That trust has a scar: the June 8, 2021 outage. Fastly's own summary said a valid customer configuration triggered a software bug, causing widespread disruption until the company disabled the configuration and restored most service; the post says 95 percent of the network was operating normally within 49 minutes after the impact began (https://www.fastly.com/blog/summary-of-june-8-outage). The outage affected high-profile sites and became one of the public examples of how much internet experience can depend on a small number of edge platforms. It did not destroy Fastly, but it changed the way buyers remember concentration risk.
The economic lesson is not that Fastly is uniquely unreliable. Any large CDN or cloud platform can fail. The lesson is that edge platforms sell risk reduction while being themselves a source of concentrated operational risk. That paradox is unavoidable. If a media company chooses Fastly for programmable control, the customer also accepts that a provider-side failure can sit in front of many application paths at once. Fastly's ability to earn premium pricing therefore depends on incident transparency, rapid recovery, configuration safety, and architectural options that let customers reduce blast radius.
Fastly's 2025 filing makes the dependence explicit in financial language. The company describes risks around service disruptions, cyberattacks, infrastructure failure, third-party data center providers, and customer concentration (https://www.sec.gov/Archives/edgar/data/1517413/000151741326000053/fsly-20251231.htm). Those are not boilerplate for this business. They are the conditions under which margin can evaporate. If a large customer loses confidence, the revenue hit is immediate; if Fastly overbuilds capacity to avoid future failure, capital efficiency can suffer; if it underbuilds, performance and reliability become sales problems.
The PeeringDB and network-map numbers sharpen the point. A 578 Tbps connected-capacity claim and a PeeringDB profile with 152 exchange points and 109 facilities are signs of scale (https://www.fastly.com/network-map and https://www.peeringdb.com/api/net?asn=54113). They are also obligations. Every metro, exchange, server cluster, and software release has to be operated well. Fastly's margin is partly an engineering achievement: high cache hit ratios, efficient routing, strong peering, resilient configuration systems, and security controls that can inspect traffic without turning performance into a casualty.
That is why an outage remains commercially relevant years later. The customer deciding whether to pay Fastly is not only comparing unit prices. It is deciding whether the provider can safely carry business-critical traffic. A premium edge vendor can survive an outage if it learns faster than customers lose trust. It cannot build an economic moat if buyers see the control plane itself as a liability.
Competitors attack from every side of the contract
Fastly competes in a field where rivals do not all look alike. Akamai is the older, larger delivery and security platform with a huge enterprise base. Cloudflare is the high-growth network and security platform that packages performance, security, developer services, and zero trust into a broad global network. AWS, Google Cloud, and Microsoft can sell delivery and edge functions as features around larger cloud estates. Smaller specialist vendors can compete on price, geography, bare-metal control, or managed service. Fastly's 2025 filing lists competition across CDN, security, cloud, and other adjacent markets, which is accurate because the buyer can define the problem many ways (https://www.sec.gov/Archives/edgar/data/1517413/000151741326000053/fsly-20251231.htm).
The financial comparison shows the challenge. Cloudflare reported 2025 revenue of $2.168 billion, up 29.8 percent, with a GAAP gross margin of 74.5 percent (https://www.cloudflare.com/press/press-releases/2026/cloudflare-announces-fourth-quarter-and-fiscal-year-2025-financial-results/). In Q1 2026, Cloudflare reported $639.8 million of revenue and a GAAP gross margin of 71.2 percent (https://www.cloudflare.com/press/press-releases/2026/cloudflare-announces-first-quarter-2026-financial-results/). Akamai reported 2025 revenue of $4.208 billion, with security revenue of $2.243 billion, delivery revenue of $1.508 billion, and cloud computing revenue of $457 million (https://www.akamai.com/newsroom/press-release/akamai-reports-fourth-quarter-2025-financial-results). Fastly's $624.0 million 2025 revenue is meaningful, but it sits between much larger platforms with more bundling power.
Those rivals pressure Fastly differently. Akamai can defend delivery accounts with scale, security depth, and enterprise relationships. Cloudflare can turn a CDN conversation into a broader security and connectivity platform sale. Hyperscalers can make CDN and edge functions feel like a native extension of existing workloads, with AWS CloudFront pricing sitting beside EC2, S3, and broader AWS commitments (https://aws.amazon.com/cloudfront/pricing/). Google can do the same around Google Cloud CDN and related cloud services (https://cloud.google.com/cdn/pricing). A buyer that wants one invoice, one cloud architecture, or the simplest procurement route has reasons not to choose a specialist vendor even when the specialist has technical advantages.
Fastly's counterargument has to be specificity. It can win where a customer cares deeply about cache control, instant changes, observability, application security at the edge, or performance on complex traffic patterns. It can also win where a customer wants to avoid putting every traffic layer inside a hyperscaler. But those are selective wins. They require a sales motion that finds buyers with real edge complexity and a product motion that makes the additional control worth the friction.
The risk is that the market splits. Very simple workloads may flow to cheap or bundled CDN services. Very broad platform decisions may flow to Cloudflare, Akamai, or hyperscalers. Fastly's sweet spot would then be high-performance, high-control, technically sophisticated customers. That can be an attractive market, but it is not automatically enormous. The 634 enterprise customers reported in Q1 2026 are therefore important. Fastly does not need every website. It needs enough large, demanding customers that value control and security more than the cheapest traffic path (https://investors.fastly.com/news-releases/news-release-details/fastly-announces-record-first-quarter-2026-financial-results).
Customer concentration is both a strength and a bargaining problem
Fastly's customer base shows the two-sided nature of edge economics. On one side, the company has enough enterprise traction to make the platform relevant. On the other, the largest customers have real negotiating power because their traffic volume can materially affect revenue, capacity planning, and gross margin. Fastly's Q1 2026 release said its top ten customers accounted for 34 percent of revenue and that revenue from those customers grew 25 percent year over year, faster than revenue from customers outside the top ten (https://investors.fastly.com/news-releases/news-release-details/fastly-announces-record-first-quarter-2026-financial-results). That is good news if the largest customers are expanding because they are buying more security, compute, and higher-value delivery. It is less comfortable if expansion comes mainly from traffic volume that must be priced aggressively.
The distinction is important because usage-based models can flatter and punish a vendor at the same time. A customer with surging media traffic may generate immediate revenue, but that traffic also consumes bandwidth, server capacity, cache storage, support attention, and peering resources. If the customer has enough scale to negotiate lower unit prices, the vendor can grow revenue while diluting margin. Conversely, if the same customer buys application security, bot protection, real-time observability, and edge compute, the account can become more profitable and more defensible. The same customer can therefore be either a proof point or a margin risk depending on the product mix inside the contract.
Remaining performance obligations help but do not eliminate the uncertainty. Fastly reported $369.0 million of RPO in Q1 2026, up 63 percent year over year (https://investors.fastly.com/news-releases/news-release-details/fastly-announces-record-first-quarter-2026-financial-results). RPO is valuable because it shows contracted future revenue, but it does not disclose the full margin profile of that revenue. A dollar of committed security revenue is not economically identical to a dollar of low-priced delivery traffic. A multi-year commitment from a sophisticated platform customer can also include price concessions, service commitments, or capacity assumptions that are invisible in the headline RPO number. The headline is positive; the analytical question is how much of that contracted base carries software-like economics.
Net retention points in the same direction. A 113 percent trailing-12-month net retention rate means existing customers, as a group, are spending more with Fastly than they did in the prior period (https://investors.fastly.com/news-releases/news-release-details/fastly-announces-record-first-quarter-2026-financial-results). That is a healthy sign for a company selling into large digital businesses. But for a usage-based infrastructure company, retention has to be read with mix. Expansion can reflect deeper product adoption, traffic growth, price increases, or some combination. The most valuable expansion is deeper adoption because it indicates that Fastly is solving more problems for the customer. The least valuable expansion is low-margin traffic that increases operational load without strengthening the relationship.
This is where the company's public segment reporting becomes useful. In 2025, network services grew, security grew faster, and other products remained small (https://www.sec.gov/Archives/edgar/data/1517413/000151741326000053/fsly-20251231.htm). In Q1 2026, security growth was much faster than network services growth, while other revenue also rose from a small base (https://investors.fastly.com/news-releases/news-release-details/fastly-announces-record-first-quarter-2026-financial-results). That pattern is the one Fastly needs: keep the traffic relationship, but attach higher-value products around it. If the mix continues shifting in that direction, customer concentration becomes less frightening because the large accounts are not merely bulk traffic buyers. If the mix stalls, concentration gives large customers more leverage over a vendor that still needs their volume.
The strategic implication is that Fastly should be judged less by the number of enterprise customers than by the depth of enterprise accounts. A customer that uses Fastly only for content delivery can leave or dual-source when pricing changes. A customer that has built request logic, security policy, API protection, logging, and operational runbooks around Fastly has a higher switching cost. The latter is the company Fastly wants to become: not a CDN line item, but a traffic-control layer embedded in how demanding digital businesses operate. The latest public metrics suggest movement toward that goal, but they do not yet prove that the model has fully escaped bandwidth bargaining.
The balance sheet has room, but the model still has to earn leverage
Fastly is not a distressed infrastructure story. It ended 2025 with $180.6 million in cash and cash equivalents and $181.2 million in marketable securities, while reporting $45.8 million of free cash flow for the year (https://investors.fastly.com/news-releases/news-release-details/fastly-announces-both-record-fourth-quarter-and-full-year-2025). The 2025 annual filing also shows convertible note obligations across maturities, including 2026, 2028, and 2030 notes (https://www.sec.gov/Archives/edgar/data/1517413/000151741326000053/fsly-20251231.htm). In Q1 2026, the company reported $146.7 million in cash and equivalents and $183.8 million in marketable securities (https://investors.fastly.com/news-releases/news-release-details/fastly-announces-record-first-quarter-2026-financial-results). The company has liquidity, but it is not free from the discipline of capital markets.
The operating leverage question is simple: can Fastly grow revenue faster than the combined cost of bandwidth, colocation, depreciation, support, sales, research, and stock compensation? The answer has been mixed. The company has posted strong revenue growth periods and product credibility, but public investors have also seen years of GAAP losses. In 2025, Fastly reported a GAAP net loss of $121.7 million while producing positive free cash flow (https://investors.fastly.com/news-releases/news-release-details/fastly-announces-both-record-fourth-quarter-and-full-year-2025). That gap is not unusual for a technology company with depreciation, amortization, and stock-based compensation, but it forces a more careful reading than revenue growth alone.
Gross margin is the key bridge between the technical story and the investor story. A 57.1 percent GAAP gross margin for 2025 is respectable for an infrastructure-heavy edge business, but it is far below software-only margins and below Cloudflare's reported 2025 GAAP gross margin (https://www.cloudflare.com/press/press-releases/2026/cloudflare-announces-fourth-quarter-and-fiscal-year-2025-financial-results/). The Q1 2026 jump to 62.5 percent is strategically important because it suggests Fastly can improve mix and efficiency. But the durability of that improvement depends on traffic cost, customer pricing, security mix, compute utilization, and capacity planning. One good quarter is not a new economic law.
Capital allocation also has to be judged against the network footprint. The company cannot starve the network and still sell premium performance. It must invest in capacity, server refreshes, software reliability, and security capability. But every dollar spent on underutilized capacity weakens the economics that the platform is supposed to prove. This is why Fastly's public capacity claims should not be read only as marketing. A 578 Tbps connected network is a competitive asset if it is efficiently filled with profitable traffic; it is a burden if it chases low-margin delivery volume (https://www.fastly.com/network-map).
The best version of Fastly's financial story is an operating flywheel: large customers bring traffic; traffic justifies peering and capacity; better edge performance attracts more complex customers; complex customers buy security and compute; security and compute lift gross margin and retention; better margin funds more product and network investment. The weaker version is a traffic treadmill: revenue grows, but competition compresses delivery pricing, security remains a partial attach, compute stays niche, and the network must keep expanding without enough incremental margin. The public numbers do not yet settle which version wins. They show a company moving in the right direction while still exposed to the wrong one.
The ARIN evidence is a starting point, not the identity of the business
Because the directory entry is named "FASTLY INC-FASTLY INC. EU ENA", it would be easy to overread the registry record. That would be a mistake. ARIN's Whois service is designed to publish registry information about internet-number resources and related contacts, not to provide a complete corporate profile (https://www.arin.net/resources/registry/whois/). The FIEE-1 record is valuable because it anchors the directory entry to a public registry artifact. It also carries caveats: the address field, handle name, and associated network references do not explain Fastly's current product mix, customer base, or control structure.
The RDAP view adds detail but not full certainty. It shows linked entities and network objects, with assigned IPv6 ranges associated with AT&T parent space rather than a simple standalone Fastly-owned prefix story (https://rdap.arin.net/registry/entity/FIEE-1). That is consistent with a real operational footprint, but it should not be turned into a claim about Fastly's whole network. The PeeringDB profile for AS54113 and Fastly's own peering page are stronger evidence for Fastly's public interconnection posture (https://www.peeringdb.com/asn/54113 and https://www.fastly.com/peering). The SEC filing and investor releases are stronger evidence for revenue, margin, customers, and liquidity.
This hierarchy of evidence matters for readers. A registry row can show that an internet-infrastructure company appears in the administrative fabric of the network. It cannot answer whether the company has pricing power. The useful analysis comes from connecting layers: ARIN/RDAP for registry existence, PeeringDB for interconnection footprint, Fastly sources for capacity and product claims, filings for audited financial statements, and competitor disclosures for market pressure. The result is a more disciplined picture than either a company brochure or a raw registry record would provide.
There is also a naming caution. The directory name includes "EU ENA", while the ARIN record's country field is United States. The safe reading is not that the name proves a particular European subsidiary structure. It is that this is a Fastly-related ARIN organization record with a United States country field and public network-resource links. The broader article therefore treats Fastly's public company and platform economics as the relevant business context while keeping the registry record in its proper role.
Why Fastly still matters
Fastly matters because the internet's performance layer remains strategically unsettled. The first generation of CDN economics was about moving bits quickly and cheaply. The current market is about where application control sits. If the edge can run code, inspect requests, enforce security policy, observe traffic in real time, and absorb attacks, then the CDN becomes a programmable operating surface for digital business. If those features are too expensive, too complex, or too easy for bigger platforms to copy, then the edge collapses back into a price-sensitive delivery layer.
Fastly is one of the clearest tests of that question because it has not hidden behind an unrelated enterprise software portfolio. Its revenue mix is still visibly tied to network services, yet its strategic upside depends on security and compute becoming more important. The company has real proof points: hundreds of enterprise customers, 113 percent net retention in Q1 2026, a global peering footprint, 578 Tbps of connected capacity, a growing security line, and improving quarterly gross margin (https://investors.fastly.com/news-releases/news-release-details/fastly-announces-record-first-quarter-2026-financial-results). It also has real constraints: customer concentration, aggressive competitors, outage memory, traffic cost, and a business model that cannot escape physical infrastructure.
The incentive structure is unusually clear. Customers want lower latency, lower origin load, fewer security incidents, more control, and predictable bills. Fastly wants those same customers to use enough platform capability that they stop viewing delivery as a commodity. Competitors want to flatten that differentiation by bundling security, CDN, compute, and developer tools into larger suites. Investors want Fastly to show that the premium edge story produces durable gross margin and operating leverage. Every product decision sits inside that four-way tension.
The most important unknown is not whether Fastly can operate a serious network. Public evidence says it can. The unknown is whether enough customers will pay Fastly for the parts of the edge that are not just transport. Security revenue growth suggests the answer may be yes. Compute-at-edge adoption and other product revenue remain too small to declare victory. Network services still pay most of the bills. Fastly's future therefore depends less on winning a generic CDN race and more on expanding the share of customer problems that are solved at its edge.
Bottom line and watchpoints
Fastly's economic story should be judged through margin, not romance. Developer preference, elegant edge logic, and high-capacity network maps are valuable only if they turn into retained enterprise accounts, stronger security attachment, better gross margin, and disciplined free cash flow. The 2025 and Q1 2026 numbers give the company a credible case: $624.0 million in 2025 revenue, a 57.1 percent full-year GAAP gross margin, $45.8 million in free cash flow, $173.0 million in Q1 2026 revenue, a 62.5 percent Q1 GAAP gross margin, and rapidly growing security revenue (https://investors.fastly.com/news-releases/news-release-details/fastly-announces-both-record-fourth-quarter-and-full-year-2025 and https://investors.fastly.com/news-releases/news-release-details/fastly-announces-record-first-quarter-2026-financial-results). The evidence also keeps the case honest: network services remain the dominant revenue line, large customers still matter heavily, and bigger platforms can bundle around Fastly's strongest features.
The first watchpoint is security mix. If security keeps growing faster than network services and lifts net retention, the Signal Sciences acquisition becomes easier to justify and Fastly's margin story improves. If security growth slows or becomes price-competed by broader platforms, Fastly loses its most direct route out of commodity delivery.
The second watchpoint is gross margin durability. Q1 2026 was a better margin quarter, but edge economics can change with traffic mix, customer concentration, capacity investment, peering outcomes, and competitive pricing. Fastly needs several quarters of proof that efficiency and mix are structurally improving.
The third watchpoint is compute relevance. Edge compute is strategically important but still small in revenue terms. The question is whether Compute and related products become a real application platform for large customers or remain a sophisticated feature used by a subset of developers.
The fourth watchpoint is operational trust. The 2021 outage remains a reminder that an edge provider's value proposition includes its own failure modes. Fastly must keep proving that programmable control does not create unacceptable concentration risk.
The fifth watchpoint is the gap between registry evidence and business reality. The ARIN FIEE-1 record supports the existence of a Fastly-related network-resource artifact, but it does not explain the business. The business is the conversion of edge traffic into software-like value. As of July 2026, Fastly has a credible but unfinished answer: a scaled global edge network, improving financial signals, and a security business with momentum, all facing a market that will quickly punish any edge provider that cannot turn control into durable margin.