패스트리(FSLY) 2026년 3분기 실적 발표 컨퍼런스 콜: 20% 성장 및 전망치 상향

핵심 요약

• 패스트리는 2026년 3분기 매출이 전년 동기 대비 20% 증가한 1억 9,020만 달러로 역대 최고치를 기록하며, 자체 전망치인 1억 8,400만~1억 9,000만 달러를 상회했다고 발표했습니다. 이는 4분기 연속 20% 이상의 매출 성장세를 이어간 것입니다.
• 비GAAP 영업이익은 3,120만 달러로 역대 최고치를 기록했으며, 영업이익률은 전년 동기 대비 약 900bp 확대된 16.4%를 나타냈습니다. 매출총이익률은 역대 최고 수준인 65.8%를 기록했습니다.
• 최근 12개월 순매출 유지율(NRR)은 2분기 117%, 전년 동기 106%에서 119%로 상승하며 6분기 연속 확대 흐름을 보였습니다.
• 다중 제품 도입도 지속적으로 증가했습니다. 대형 고객사 중 2개 이상의 제품 군을 사용하는 비중은 74%, 4개 이상을 사용하는 비중은 34%로 2분기(30%) 대비 상승했으며 2년 전 수준의 3배 이상으로 늘어났습니다.
• 경영진은 2026년 전체 실적 전망치를 매출 7억 4,400만~7억 4,800만 달러, 비GAAP 영업이익 1억 700만~1억 1,100만 달러, 잉여현금흐름(FCF) 5,000만~6,000만 달러로 상향 조정했습니다.
• 스콧 러빗(Scott Lovett) 시장개척(Go-To-Market) 사장은 2027년에 퇴임할 계획입니다. 킵 컴프턴(Kip Compton) CEO와 러빗 사장은 신임 최고수익책임자(CRO) 선임을 함께 이끌 예정이며, 러빗 사장은 과도기 동안 조직 운영을 지속할 예정입니다.

주요 재무 실적

별도로 명시되지 않는 한, 경영진이 논의한 수익성 수치는 비GAAP 기준입니다.

사업 및 영업 실적

업셀링과 크로스셀링이 주요 매출 성장 동력으로 작용했습니다. 패스트리는 3분기 실적 호조의 원인으로 일부 대형 고객사의 예상보다 강한 트래픽 증가와 보안 제품의 지속적인 성장을 꼽았습니다.

상위 10개 고객사가 매출의 39%를 차지하여 2분기(37%) 대비 상승했습니다. 이들 고객사로부터의 매출은 전년 동기 대비 46% 증가한 반면, 상위 10개 외 고객사의 매출은 8% 증가에 그쳤습니다. 패스트리의 분기 말 기준 대형 고객(연간 매출 10만 달러 초과 계정) 수의 경우 634개로 집계되었습니다.

보안 매출은 2025년 3분기 대형 보안 수주로 인한 역기저 효과에도 불구하고 32% 증가했습니다. 경영진은 봇 관리 및 DDoS 방어 제품이 이번 분기에도 세 자릿수 성장을 기록했다고 밝혔습니다.

패스트리는 AI 런타임 컨트롤(AI Runtime Control), AI 방화벽(AI Firewall) 및 확장된 API 보안 기능을 통해 AI 포트폴리오를 확장했습니다. 주요 AI 기업 한 곳은 에이전트(agentic) 트래픽의 보안 확보와 확장을 위해 7자리 수 규모의 계약을 체결했습니다. 경영진은 텍스트 기반 트래픽이 상대적으로 적은 바이트를 사용하기 때문에, 현재 에이전틱 워크로드는 전송 매출보다 보안 측면에서 더 의미 있는 기회를 제공한다고 설명했습니다.

해외 투자를 통해 일본의 유통, 스트리밍, 가전 분야에서 신규 고객을 확보했습니다. 또한 컴캐스트(Comcast) 네트워크 내에 패스트리의 전송, 보안 및 컴퓨팅 플랫폼을 구축하는 통합 구축 계약을 발표하여 최종 사용자에게 더 가까운 곳에서 더 낮은 지연 시간의 서비스와 추가 용량을 지원하게 되었습니다.

패스트리는 2027년부터 기존 3개 매출 카테고리를 하나의 통합 매출 라인으로 개편할 계획입니다. 회사는 고객의 6개 제품 군 활용 현황을 파악할 수 있도록 다중 제품 도입 현황을 지표로 제공할 예정입니다.

경영진 실적 전망(가이던스)

경영진은 시장개척 인센티브 증가, 계절적 마케팅 활동, IT 지출 등으로 인해 4분기 영업비용이 전분기 대비 수백만 달러(한 자릿수 중간대) 증가할 것으로 예상하고 있습니다. 3분기 비용은 150만 달러 규모의 비단발성 세금 항목 환급 혜택을 받았습니다.

패스트리는 2026년 인프라 설비투자 전망을 기존 매출 대비 10%~12%에서 약 10%로 하향 조정했습니다. 회사는 용량 확장을 조기 진행함에 따라 연간 설비투자의 70% 가까이가 상반기에 집행되었다고 밝혔습니다.

리스크 및 주요 점검 사항

• 패스트리의 사용량 기반 매출 모델은 특히 고객 트래픽 수준과 일회성 이벤트로 인해 분기별 변동성을 유발할 수 있습니다.
• 성장이 여전히 대형 고객사에 집중되어 있습니다. 상위 10개 고객사가 3분기 매출의 39%를 차지했으며, 나머지 고객군에 비해 훨씬 빠른 속도로 성장했습니다.
• 경영진은 패스트리가 상위 고객사에 지나치게 집중했을 가능성을 인정했습니다. 더 넓은 범위의 성장을 뒷받침하기 위한 마케팅 전환, 해외 확장, 전략적 파트너십 구축에는 더 많은 시간이 소요될 것으로 예상됩니다.
• 시장개척 부문 리더십 교체 계획은 실행 리스크를 수반할 수 있으나, 패스트리가 후임자를 찾는 동안 러빗 사장은 계속 업무에 관여할 예정입니다.
• 패스트리는 특히 메모리 부품을 중심으로 공급망 상황을 계속 모니터링하고 있습니다. 경영진은 선제적 구매와 부품 업그레이드를 통해 잠재적 차질을 완화하는 데 도움이 되었다고 밝혔습니다.
• 4분기 트래픽 가정에는 게임 다운로드, 연말 쇼핑, NFL 트래픽, 미국 중간선거 등 일회성 수요가 포함되어 있어 실제 소비량을 정밀하게 예측하기는 어렵습니다.

애널리스트 Q&A 주요 내용

AI 수익화: 경영진은 AI 에이전트가 전송 트래픽에 기여할 수 있지만, 단기적으로 더 큰 기회는 보안 부문에 있다고 밝혔습니다. AI 기업의 에이전틱 워크로드를 보호하고, 봇 관리, AI 런타임 컨트롤, AI 방화벽 등의 제품을 통해 기업이 봇과 에이전트를 관리할 수 있도록 지원함으로써 매출이 발생할 수 있습니다. 대부분의 보안 계약은 구독 기반이며 가격은 예상 트래픽이나 요청 수량의 영향을 받습니다.

AI 워크로드를 위한 자본 요건: 7자리 수 규모의 AI 계약은 GPU나 특정 거래를 위한 전용 설비투자를 필요로 하지 않습니다. 패스트리는 단일 통합 인프라를 통해 전용 네트워크 투자 없이도 새로운 워크로드를 지원한다고 밝혔습니다.

순매출 유지율 확대: 경영진은 119%의 순매출 유지율이 업셀링과 크로스셀링 모두에 힘입은 것이라고 설명했습니다. 더 많은 제품 군을 도입하는 고객일수록 전반적으로 더 높은 유지율을 보이며, 상위 고객군과 비상위 고객군 모두에서 확대가 확인되었다고 덧붙였습니다.

상위 10개 외 고객사의 성장: 경영진은 대형 고객사 외의 성장 둔화가 시장개척 이니셔티브의 시기와 연관되어 있다고 설명했습니다. 브랜드 개발, 신임 마케팅 리더십 구축, 아시아태평양·일본(APJ) 및 라틴아메리카 지역 확장, 파트너십은 기존 주요 고객사와의 심화된 관계 구축보다 더 많은 시간이 소요될 것으로 예상됩니다.

시장개척 승계 계획: 패스트리는 차기 최고수익책임자(CRO) 선임을 위해 내부 및 외부 후보를 모두 검토할 예정입니다. 경영진은 현재 전망치에 리더십 교체의 영향 평가가 반영되어 있다고 밝혔으나, 2027년 전망치 제시에 대해서는 말을 아꼈습니다.

GTA 6 트래픽: 경영진은 다운로드 용량, 향후 패치, 패스트리의 배급 점유율 등의 요소를 바탕으로 4분기 GTA 6 관련 가정이 이전에 포트나이트(Fortnite)에 적용했던 수준보다 크다고 밝혔습니다. 다만 트래픽 기여는 일회성 부문이며 변동성이 매우 크다고 강조했습니다.

실적발표 콘퍼런스 콜 전문

전체 실적 발표 컨퍼런스 콜 녹취록

경영진 발표

Operator

Good afternoon. My name is Michelle, and I will be your conference operator today. At this time, I would like to welcome everyone to the Fastly Third Quarter 2026 Earnings Conference Call. [Operator Instructions]

Please be advised today's conference is being recorded. I would now like to turn the conference over to Vernon Essi, Investor Relations at Fastly. Sir, please go ahead.

Vernon Essi

Thank you, and whom everyone to our third quarter 2026 earnings conference call. We have Fastly's CEO, Kip Compton, and CFO, Rich Wong, with us today. The webcast of this call can be accessed through our website, fastly.com and will be archived for 1 quarter. A copy of today's earnings press release, related financial tables and supplements, all of which are furnished in our 8-K filing today can be found in the Investor Relations portion of Festa's website, along with the investor presentation.

During this call, we will make forward-looking statements, including statements related to the expected performance of our business, future financial results, product and services, sales and growth, strategy, long-term growth and overall future prospects. These statements are subject to known and unknown risks, uncertainties and assumptions that could cause actual results to differ materially from those projected or implied during the call.

For further information regarding risk factors for our business, please refer to our filings with the SEC, including our most annual report filed on Form 10-K and quarterly reports filed on Form 10-Q filed with the SEC and our third quarter 2026 earnings press release and supplement for a discussion of the factors that could cause our results to differ. Please refer in particular to the sections entitled Risk Factors. We encourage you to read these documents.

Also note that the forward-looking statements on this call are based on information available to us as of today's date. We undertake no obligation to update any forward-looking statements, except as required by law. Also during this call, we will discuss certain non-GAAP financial measures and certain key performance indicators. Unless otherwise noted, all numbers we discuss today other than revenue will be on an adjusted non-GAAP basis. We do not provide reconciliations of forward-looking non-GAAP measures because quantitative reconciling of information for these measures is unavailable without unreasonable effort.

Reconciliations to the most directly comparable GAAP financial measures are provided in the earnings release and supplement in our Investor Relations website and filed with the SEC. These non-GAAP measures are intended to be a substitute for our GAAP results. Before we begin our prepared comments, please note that during the fourth quarter, we will be attending the RBC 2026 Global TIMT Conference in New York on November 17. The UBS 30th Annual Global Technology and AI Conference in Scottsdale on December 2 and the Wells Fargo Tenth Annual TMT Summit in Rancho Palos Verdes on December 8.

Now I'll turn the call over to Kip.

Kip Compton

Good afternoon, everyone, and thank you for joining us. Fastly delivered another strong quarter with record revenue, operating income and free cash flow. Q3 revenue was $190.2 million, up 20% year-over-year and above the high end of our guidance. Operating income was $31.2 million for an operating margin of 16.4%. Gross margin was 65.8%, matching record levels. This is our fifth consecutive quarter of operating profit and our seventh consecutive quarter of positive free cash flow.

Trailing 12-month net retention rate rose to 119%, and we continue to see customers expanding across our platform and driving more revenue. Today, I'll cover what's driving these positive results. Our customers are growing with us. We're delivering outcomes that matter most to them, and we're building for AI scale. I will also share details of our planned transition in our go-to-market leadership in the year ahead. Customers come to us with business problems, traffic they can't afford to drop, an attack they need to stop or a wave of AI agents they need to govern.

The breadth of our platform and relationships built on trust allow us to solve those challenges, and that's how we grow with our customers. Customers increasing the value that they derive from the Fastly platform is a central driver of our financial performance and underpins our ability to create long-term value for you, our shareholders. As we continue to refine our operating metrics, multiproduct adoption, along with incremental margin will be key to how we evaluate our performance.

In the third quarter, 74% of our large customers use 2 or more of our product suites and 34% used 4 or more, up from 30% in the second quarter and more than 3x the level we saw just 2 years ago. Customers who use more of our product suite see more value are stickier and spend more with us over time. Our go-to-market transformation is a multifaceted strategy with investments aimed at driving growth across customer cohorts. Parts of it are already paying off. We have seen tremendous success with our largest customers, driving robust growth that, in turn, is powering our overall growth.

Expanding the growth we are seeing with our largest customers to our other customer cohorts is our biggest near-term opportunity and a clear execution priority. The other go-to-market investments that we have made, including our international expansion, strategic partnerships and marketing transformation, are progressing well, and we expect them to have a larger impact over time. Our customer wins this quarter show this strategy at work.

Many span multiple product suites and highlight our platform's high performance. As AI use cases continue to accelerate, one of the world's leading AI companies signed a 7-figure commitment with Fastly to secure and scale its agentic traffic. Agents are central to its strategy. They chose Fastly to protect that mission-critical work.

Our track record of delivering reliability and performance at scale continue to pay dividends with a major U.S. media and entertainment company, home to leading new sports and entertainment brands selected fastly to power live streaming across its digital properties. We are also seeing significant success in unseating incumbent providers. One of the largest restaurant chains in the U.S. replaced a legacy security vendor with Fastly to protect and accelerate its online checkout and has already seen critical performance and reliability metrics improve for its most important digital revenue stream.

Security and Control are winning us business, a global online marketplace with tens of millions of members added Fastly delivery, WAF and bot management, giving it network resilience, improved security posture and the transparency and control at once over its network. In the APJ region, our investments are producing results. A major Japanese apparel and lifestyle retail group moved to Fastly for delivery, security and bot management to optimize infrastructure costs while maintaining performance and security.

A leading Japanese streaming service chose Fastly to scale while reducing infrastructure costs. And one of the world's largest consumer electronics makers is using Fastly to protect user privacy and deliver more reliable software updates as it launches AI services on its devices. As the breadth of our platform expands, we are also building new routes to market through strategic partnerships, a digital experience platform company that runs websites, customer portals and Internet is building Fastly's delivery, WAF, bot management and DDoS protection into its platform through a managed service provider partnership, expanding our reach to critical new customers.

In September, we announced an industry-first deployment in partnership with Comcast. It puts Fastly's entire platform, including security and compute inside Comcast network to deliver content and applications closer to users. Uniquely, we have integrated Fastly software directly into Comcast's networking elements, delivering high capacity and performance. For customers, it means lower latency and faster downloads, even when millions are streaming at once. For Comcast, it means delivering a new level of speed, reliability and quality.

And for Fastly, it adds capacity close to end users and serves as a foundation for latency-sensitive applications from online gaming to emerging AI services. We look forward to adding more network partners and deployment models in the coming quarters. We're also investing for what's next as enterprise AI shifts from experimentation to business-critical production, the focus turns to leveraging AI at scale.

AI use cases on Fastly are showing up across our network in more ways from agents interacting with our customers' applications to code generation, model distribution and content authenticity. In Q3, we expanded the Fastly for AI portfolio beyond capabilities like bot management and NextGen WAF. We announced new innovations, including AI run time control to govern AI access and usage, AI firewall to protect AI applications and expanded API security to control how AI agents access enterprise applications and websites.

We built these solutions with our customer design partners who helped shape what we shipped. We'll keep working closely with customers as their challenges evolve. This quarter also shows progress on 3 priorities: increasing product velocity, transforming our go-to-market motion and driving profitable growth with record revenue, operating income and free cash flow. I'm proud of this team, grateful for the trust our customers place in us and as optimistic as ever about the future of Fastly.

Before I hand it over to Rich, I want to share a leadership update. Scott Lovett, President go-to-market plans to retire next year. Together, Scott and I will lead a broad search for our new Chief Revenue Officer, and Scott will continue to run the go-to-market organization during the transition. Since joining Fastly in 2024, Scott has transformed how we sell. He brought executive selling and true customer partnerships to our largest accounts, aligned our sales motion and conversation around the platform and expanded our presence in APJ.

He also built a seasoned leadership team across sales, marketing, channel and customer success. That strength and our strategy are not changing. Scott's leadership through the transition gives us time to find the right leader for the next phase of our go-to-market transformation with continuity for our customers and our teams. On behalf of the entire Fastly team, Scott, thank you.

And now I'll hand it over to Rich, who will take you through our results and our outlook. Rich?

Richard Wong

Thank you, Kip, and thank you, everyone, for joining us today. Before I review our detailed financial results, I want to emphasize that Q3 demonstrated the strength of our underlying business model that Kip and I discussed at our first earnings call together a year ago and highlighted again at our Investor Day back in September. We delivered durable revenue growth of 20% or more for 4 consecutive quarters while driving meaningful operating leverage across our expense base. Q3 was also our fifth consecutive quarter of non-GAAP operating profit and our seventh consecutive quarter of positive free cash flow.

Now on to the specifics of our Q3 results. I'd like to remind you that unless otherwise stated, all financial results in my discussion are non-GAAP based. Revenue for the third quarter increased 20% year-over-year to $190.2 million, above the high end of our guidance range of $184 million to $190 million. This result was a record high for Fastly and as I mentioned, the fourth consecutive quarter of 20% or more year-on-year revenue growth.

The revenue strength in the quarter was driven by continued success in our go-to-market upsell and cross-sell motions as we see customers adopt more product suites within our platform. The upside was driven by better-than-anticipated traffic at some of our larger customers and continued success with our security products. At Investor Day, we highlighted our shift toward a single unified edge platform with 6 main product suites.

Beginning in 2027, we're aligning our financial metrics to match the strategy by transitioning from 3 separate revenue lines, network services, security and other, down to one consolidated revenue line. To give investors full visibility into platform adoption, we're introducing a new reporting metric, multiproduct adoption. This metric measures the percentage of large customers adopting 2 or more product suites and those adopting 4 or more product suites.

In the third quarter, 74% of our large customers adopted 2 or more product suites, up 7 percentage points year-over-year and up 2 percentage points quarter-over-quarter. More importantly, 34% of large customers have now adopted 4 or more product suites, up from 30% in the second quarter and more than 3x the level we saw just 2 years ago.

For more information on multiproduct adoption, including a description of the product suites and historical trended view of this data, please refer to our investor presentation and supplement on our IR website. I will now review our prior 3 revenue line reporting. In the third quarter, Network Services revenue of $138 million grew 16% year-over-year. Security revenue was $44.7 million, which represented a growth of 32% year-over-year and represented 24% of revenue.

Recall that our security revenue growth faced tougher year-over-year comps as it was impacted by a large security win with a top 10 customer in the third quarter of 2025. However, on a sequential basis, security revenue has demonstrated growth in excess of 7% for every quarter in 2026. Our other products revenue of $7.4 million grew 37% year-over-year. We will report these 3 revenue lines one final time for the fourth quarter of 2026, before retiring them in favor of our streamlined single revenue structure and multiproduct adoption metrics.

In the third quarter, our top 10 customers represented 39% of revenue, a moderate increase from 37% in the second quarter. Revenues from our top 10 customers grew 46% year-over-year. Revenue from our customers outside of our top 10 grew 8% year-over-year. Also, no single entity accounted for 10% of our revenue in the third quarter, a group of entities under common control as a single customer accounted for 12% of the company's revenue for the quarter. Our large customer account, which represents customers with more than $100,000 in annualized revenue in the quarter with 634 customers.

Our trailing 12-month net retention rate was 119%, up from 117% in the prior quarter and up from 106% in the year ago quarter. This marks Fastly a sixth consecutive quarter of NRR expansion. The quarter-over-quarter and year-over-year increases were due to revenue increases in our larger customers as they expand their use of our platform. We exited the third quarter with RPO of $334 million, growing 24% year-over-year.

The current portion of RPO was 81% of total RPO and grew 31% year-over-year. We believe improvements in our go-to-market motion have contributed to year-over-year RPO growth, though RPO will fluctuate quarter-to-quarter with the timing of large renewals. Given this variability, we continue to encourage investors to focus on our current RPO, which gives a more accurate indication of our revenue visibility and year-over-year progress.

I will now turn to the rest of our financial results for the third quarter. Our gross margin was 65.8% in the third quarter, matching our record high for Fastly. Gross margin was 80 basis points above our guidance midpoint of 65% and up 300 basis points from 62.8% in Q3 2025. Our gross margin performance is driven by revenue scale relative to our cost-effective software-defined infrastructure. Combined with our continued financial discipline and our cost of revenue, we believe our gross margins are sustainable at these levels.

This is further substantiated by our incremental gross margin flow-through on a trailing 12-month basis increasing to 93% in the third quarter, up from 58% a year ago. Operating expenses were $94 million in the third quarter coming in better than anticipated due to disciplined expense management and a nonrecurring tax benefit of $1.5 million that favorably impacted operating expenses. We had operating income of $31.2 million in the third quarter, exceeding the high end of our operating income guidance range of $20 million to $24 million.

As mentioned, this upside was a combination of higher revenue and resulting gross profit flow-through as well as less-than-anticipated operating expenses. We believe this reflects the inherent operating leverage in our business model. This is demonstrated by our operating margin expanding from 7% in the third quarter of last year to 16% this quarter, expansion of approximately 900 basis points year-over-year.

This is underscored by our incremental operating margin flow-through of 79% of revenue on a trailing 12-month basis, significantly above our long-term target of 30% to 40%. In the third quarter, we reported a net profit of $31.6 million or $0.17 per diluted share compared to a net profit of $11.1 million or $0.07 per diluted share in Q3 2025. Our adjusted EBITDA was $43.5 million or 23% of revenues in the third quarter compared to $25.7 million or 16% of revenues in the third quarter of 2025.

Turning to the balance sheet. We ended the quarter with approximately $365 million in cash, cash equivalents, marketable securities and investments, including those classified as long term, a sequential increase of $28 million over Q2 2026. We also ended the quarter with a positive net cash balance of $41 million compared to $5 million one year ago. Our cash flow from operations was positive $40 million in the third quarter compared to positive $29 million in Q3 2025. Our free cash flow for the third quarter was positive $25 million representing our seventh consecutive quarter of positive free cash flow.

Our infrastructure capital expenditures were approximately 6% of revenue in the third quarter. As we discussed in prior quarters, we front-loaded our 2026 CapEx to ensure we have adequate equipment given supply chain constraints. We anticipate our CapEx spend will remain at a lower level in the fourth quarter, as I will discuss in a moment. In closing, the results from our first 9 months reflect disciplined execution and expanding platform reach. We believe this strengthens our underlying financial model and reinforces our conviction in our 2026 as longer-term goals.

I will now discuss our outlook for the fourth quarter and full year 2026. I'd like to remind everyone again that the following statements are based on current expectations as of today and include forward-looking statements. Actual results may differ materially, and we undertake no obligation to update these forward-looking statements in the future, except as required by law. Our revenue model is primarily based on customer consumption, which can lead to variability in our quarterly results.

Our revenue guidance reflects these dynamics in our business and is based on the visibility that we have today. Customers who trust us with one business problem bring us to next. That's the engine behind our platform strategy, which as Kip discussed is foundational for success. Over time, we expect this to mean deeper relationships and a more durable revenue model. The strategy is working, and it gives us added assurance in the opportunity ahead.

In the fourth quarter, we expect revenue in the range of $198 million to $202 million, representing 16% annual growth at the midpoint. We anticipate our gross margins for the fourth quarter will be 65.5% plus or minus 50 basis points. As a reminder, our gross margin performance is highly dependent upon incremental revenue increases or decline relative to our infrastructure costs.

We anticipate OpEx to increase mid-single-digit millions on a sequential basis due to increased go-to-market spend on incentives and seasonal market activities as well as increased IT spend. As I mentioned earlier, we incurred a $1.5 million nonrecurring tax benefit in the third quarter. For the fourth quarter, we expect a non-GAAP operating profit of $30 million to $34 million, reflecting an operating margin of 16% at the midpoint. .

We expect the non-GAAP net earnings per diluted share of $0.16 to $0.18. For calendar year 2026, we are raising our revenue guidance to a range of $744 million to $748 million reflecting annual growth of 20% at the midpoint. We anticipate our 2026 gross margins will be 65.5%, plus or minus 50 basis points. We are increasing our non-GAAP operating profit expectations to a range of $107 million to $111 million, reflecting an operating margin of 15% at the midpoint and highlighting our improved profitability compared to 2025 operating margin of 4%.

We expect our non-GAAP net earnings per diluted share to be in the range of $0.61 to $0.63. We continue to closely monitor supply chain dynamics, particularly regarding memory components and have taken strategic actions in the end of 2025 and through the first half of 2026 to mitigate potential impact. Our software-defined infrastructure is continuously improving, typically with lower capital requirements for expansion than legacy competitors.

We are also implementing server component upgrades in our fleet to efficiently expand our capacity. This structural efficiency underpins our expanding gross margins, positioning us to stay ahead of global traffic trends while maintaining strict capital discipline. For 2026, we anticipate our infrastructure capital spend will be approximately 10% of revenue compared to our previous guidance range of 10% to 12% of revenue as we are at the tail end of our front-loaded 2026 capacity expansion. We will continue to have a rigorous planning process to ensure that our capital investments align with demand.

As a result, we are raising our 2026 free cash flow guidance in the range of $50 million to $60 million. To recap, we're seeing continued evidence disciplined execution and growing platform adoption are translating into stronger financial performance. We believe our refined strategy to focus on the power of our platform is working and we are evaluating ways to better align our financial disclosures to our success.

Before we open the line for questions, we would like to thank you for your interest and your support in Fastly. Operator?

Operator

[Operator Instructions] Our first question will come from the line of Jackson Ader with KeyBanc Capital Markets.

질의응답

Aidan Daniels

This is Aidan Danes on for Jackson. Last quarter, you talked about AI-driven traffic growing materially faster than human traffic. In this quarter, you ended the 7-figure deal that to secure agentic traffic workloads. Can you maybe just walk us through just how agentic traffic shows up in the revenue for Fastly. Or maybe using this deal as an example, our revenue is solely showing up in security. We're also in more traditional CDN revenues, Nick as a function of bytes toward? And then just to follow up on that, just maybe as this template. Is this maybe a template for additional deals you're seeing in the pipeline?

Kip Compton

Yes. Thanks for the question. I think there's a few ways that genetic traffic shows up in our revenue. Sometimes it's helping the agents and protecting their traffic. And that's what that deal that I mentioned in the prepared remarks was about. And that's certainly a source of revenue for us and I would not be surprised if we saw further interest in that from other AI companies. Of course, you mentioned the security side that's right now a broader set of concerns with our broader customer base around managing the agents and the bots that are coming to their sites.

And that drives some volume in our CDN business, although I think people in the industry have noted that actual amount of traffic related is relatively low, given that these are text-based things, but it also drives security revenue in AI bot detection, regular bot detection and things like our new AI run-time controller and AI firewall that we launched during the quarter.

Operator

[Operator Instructions] Our next question will come from the line of Paramveer Singh with Oppenheimer & Co.

Paramveer Singh

I want to kind of dive a little bit more into your go-to-market strategy, obviously, losing Scott is a big loss, but I want to hear think about how you want to address this in terms of bringing on an outside go-to-market leader or kind of looking internally? And also, how does your go-to-market change into the next year as more and more of the security portfolio, the gins to cross-sell into your customer base?

Kip Compton

Yes. Now this is the top of mind for us today with sort of Scott's retirement. Scott's been obviously foundational to the reinvigoration of our go-to-market engine over the last few years and has helped us drive to now, I think, 4 quarters of 20% or greater growth. So we'll miss him. We're very happy, though, that he's going to be sticking around for a while in the next year and helping us very much with this transition. He and I will be leading a search for his successor.

In terms of the search, we will be looking at internal candidates. Scott has built a strong team as well as external candidates to make sure that we find the best leader for the next phase of our go-to-market transformation. You mentioned how well things change. I mean, security cross-sell, you're right, will continue.

And as you've seen us launching additional security products like run time control and AI firewall, those will continue to provide us with additional cross-sell opportunities. We also have a lot of momentum with our existing security products. We had a triple-digit growth again this last quarter in bot and DDoS.

I think the thing that will change will be just the progress we make on the investments we've already talked about. So whether it's strategic partnerships or what our new CMO has been on the marketing side or the international expansion of our go-to-market motion that we've talked about. Scott will continue to drive those, and those will be key priorities for his successor as well.

Operator

[Operator Instructions] Our next question will come from the line of Fatima Boolani with Citi.

Fatima Boolani

My first question is for Kip. I was hoping you could opine on the PHP headlines that took over the new cycle a couple of weeks ago as it relates to server side, website vulnerabilities. How is the platform positioned to monetize some of these newer on the web security stack landscape? And how broadly would you characterize kind of pipeline and demand for the security portfolio in direct response to that?

And then a follow-up for Rich. Just wondering if you can talk to some of the gating factors for the non-top segment of the business. So that's been decelerating. So just hoping you can put some finer points on why the momentum is trending in the opposite direction and what you can do to rehabilitate that especially in the context of Scott's transition?

Kip Compton

Yes. With respect to PHP to Shell, our hypothesis for a while has been that the emergence of AI bringing accelerated vulnerability, discovery and exploitation would actually increase the value of security products, particularly like our next-gen WAF at the edge. That was our reaction earlier this year when Mythos came out and we've seen that a number of times, and I think PHP to sell is just the most recent example of that. Customers can use the NextGen WAF to virtually patch all of their components because we can detect that exploit and block it before it reaches their components.

And that gives them time to remediate the different components in their state. And that can take some time, especially if some of the things in their IT estate are older, patches may not be immediately available. So I don't know that I can point to specific demand specifically from the PHP to shell exploit. But we do see it as an overall trend that reinforces the value of having sort of one control point like our NextGen WAF for customers to deal with this accelerated threat environment. I'll let Rich take your second question.

Richard Wong

Fatima, thank you for that second question. So I think the go-to-market transformation that Scott has done with the organization has been phenomenal. I mean, for 2025 and 2026, we are really focused on the largest customers and focusing on where we succeed, which is where performance matters. We basically have talked about the platform approach and cross-selling opportunities. And we've made a lot of progress on that. That said, the go-to-market transformation is still continuing. His focus right now has also been acquiring new customers.

We've mentioned on the prior earnings call that we have a new CMO with Jon, who started. Joan is very focused on top of funnel and expansion value for all customers. And we've also talked about our international expansion, specifically in APAC and even the LatAm regions. And so we're basically investing that -- these investments take out at different points in time. And so I think that we should be seeing improvement on the non-top 10 to come.

Operator

[Operator Instructions] Our next question comes from the line of Frank Louthan with Raymond James & Associates.

Frank Louthan

For the new contract or the AI company, what kind of drove that business to you? And I assume there's no sort of GPU commitment, but is there any CapEx involved with that? And then I had a follow-up regarding some of the discussions around AI agents as a potential opportunity or threat, depending on where you are with subscription-based businesses and so forth. How do you play in that? Do you benefit that? Is it more on the delivery side or the security side of that play?

Kip Compton

Great question. On that AI deal, no -- you're correct in your assumption that GPUs are not required, and there's no specific additional capital investment required. We've talked about in the past, and this is another example of our unified platform. So we have one network, one infrastructure that runs all of our services, and that gives us the flexibility to onboard new workloads like this one without additional specific investments. In terms of what drove them to work with us. I think with the combination of the quality of some of the security products that we could offer with -- as often is the case with our exceptional performance where they really wanted to make sure that they had a high-performance solution and we're excited to be working with them on that.

On the bread to subscription models, it's an interesting phenomenon. I don't know that it -- just to be clear, I don't think Frank, you are suggesting this, but we don't see it as an effect to our subscription model, but we do see the consumer dynamics where agents can help people manage their subscriptions more effectively, and that could certainly affect those consumer business models. For us, agentic traffic is attended traffic, and we process that traffic, as I said earlier, a relative modest impact on delivery revenue, more significant opportunities in security. And we help our customers govern net traffic and it depends on their business strategy as far as how they want that traffic treated.

Operator

Our next question will come from the line of Peter Levine with Evercore.

Peter Levine

I think the highest we've seen for a while since maybe 2022, just based on my model. But I guess how much of that impact or improvement reflects just broader-based expansion across your customer base versus maybe a handful of larger customers just kind of increasing consumption. Can you maybe just break down kind of the acceleration in net retention this quarter. What drove it?

Richard Wong

Yes, Peter, thank you for the question. Yes, so NRR, we reported was 119%. That was our sixth consecutive quarter of increasing NRR -- as you can imagine, of course, very obviously, the NRR expansion is both cross sells and upsells. I think that for us, the addition of multiple products and new products over the past kind of 12 to like 18 months has really helped the cross-selling efforts on that front.

But the more we release, I think we talked about the product suites and having 6 product suites out there. The more that we've released, we've actually noticed a general correlation that as customers buy more products suites, the NRR tends to increase. And so what we're seeing here and we're benefiting from is really a broader product portfolio.

And as they adopt more and more, that kind of NRR tends to rise with our customer base. We are seeing kind of broad-based across -- I think your question was larger or smaller customers. We do see strong across both our top customers and our nontop customers as well. So it's pretty uniform across that group.

Peter Levine

And I know -- I thinking the right way to phrase the question is Scott me a board are going to be going through this transition, you're looking for a new CRO, and you're not gone in for '27, but maybe help us or investors, how are you going to mitigate any kind of risk, but there's always going to be disruption with the new CRO. We've seen it.

Maybe it doesn't happen, but should we be prepared for the '27 guide to maybe adjust for potential disruption to go to market and maybe that's going to impact numbers when you report next -- when you give us the guide next quarter. Just help us understand how you're thinking about any possible disruption in how you're going to mitigate that and when you do guide for '27, should we be prepared now just to, I guess, just assuming that maybe that is baked into the numbers?

Kip Compton

Yes. I mean it certainly understand the concern. I'll comment and Rich may have some more detail. Our guide that we've offered today incorporates our view of the impact of that transition. We're very pleased with the transition plan that we've worked out with Scott, and we think it affords us more than enough time to find a fantastic leader and then 5 for Scott to continue to be involved in the business. even as we're ramping that leader up.

We're very pleased with the momentum that we've established with 4 quarters now in a row of 20% or greater revenue growth, and we're very happy with the strength of the leadership team that Scott has assembled as well as the investments that I mentioned earlier that are playing out now, and we think will help in the future.

I don't know that we can comment on our '27 guide until we have it to offer. If we develop concerns one way or the other about any aspect of the business, including this transition, obviously, we'll incorporate that entire guide. But for now, the guidance that we've offered incorporates our view of the situation. I don't know, I don't know if we have anything more specific.

Richard Wong

Yes. Nothing else to add. I've been through a number of transitions in prior career, and I would say that the best transitions are ones where there's a next level bench strength that has been formed and developed. I think that Scott has just done a fantastic job in building his next level bench strength. And it's not just Scott's org. I think that the rest of the leadership team has a good bench strength that kind of brings the continuity and the continuity that's needed in a regular business as public.

Operator

[Operator Instructions] Our next question will come from the line of Rudy Kessinger with D.A. Davidson.

Rudy Kessinger

I want to go back to the Fatima's question on the non-top 10. Just any further color you can add there as to why the growth has really deos significantly the last couple of quarters. Is it just weak overall traffic growth in that top 10? Or is there anything else you can share that can help us kind of reconcile the 10-point in that growth the last 2 quarters.

Unknown Executive

Yes. I think the only other thing I would add is that I think that we probably over focused a little bit on the top 10, I think that -- or even top 50 actually because I think one of the areas that Scott has done a fantastic job on has been really focused on the customers that matter to -- that -- where performance matters to them.

And so as a result, we spent a lot of time on that. And it's really timing, right? I mentioned that these initiatives -- different initiatives take different times, building a brand beyond top 50, making investments on kind of getting out there and doing this I guess selling beyond the top 50. That takes a little bit more time than it does with the top 50.

Kip Compton

Yes, I think that's right. I mean, obviously, our engagement with those top customers has been incredibly effective, and that's something we were able to put in place early on in Scott's tenure. And so we're seeing the results of that now. Engaging more tightly with those customers is something we could get in place fairly quickly. When you think about transforming our marketing engine with bringing in a new marketing leader. Obviously, that's going to take time to show results.

You think about international expansion with the new leadership team in APJ and changes in Latin America as well, where both areas where we see significant opportunities going forward that takes some time as well. And with partnerships, we talked about the Comcast 1 on the call today, but we have a number of other partnerships in the pipeline, and those can -- have a huge impact on the business, but they take time to develop. So I think it really is just these investments have different time scales, and we're pretty confident in the progress we're making in those investments, and we think that they will generate results over time.

Rudy Kessinger

Yes. Okay. Got it. And then secondly, everybody has been asking GTA 6, Q4. Any kind of commentary you can provide maybe just comparing it to the Fortnite download traffic you had Q4 last year, the World Cup just in terms of how much it could potentially generate in Q4 and just what all you have assumed in your guide at this point for it?

Kip Compton

Yes. I mean, GTA 6 a bunch of us are looking forward to it. But I'll turn it over to Rich to comment on how his team incorporated into the guide offered.

Richard Wong

Yes. The way we are looking at it, basically, we work pretty closely with our infrastructure team, making sure that we have the capacity to support the customers we have here. We do model it out based on a number of assumptions. So the assumptions that really drive kind of what goes into our model is really around initial download size, kind of the number of subsequent patches and the size of those patches kind of our share of wallet with those customers and their share of wallet on the GTA 6 distribution you can tell that there's a lot of assumptions that go behind it.

And we basically work very closely with our customers to really make that model and put it in. But as you can imagine, it is very episodic and it could be highly variable. I appreciate the question and the drill down here. I think that's all I can really say on the GTA 6 beyond GTA 6, I would say that other episodic areas that we think about and are factoring in. It's a little bit more known because it's happened in the past, which would be holiday shopping, NFL and then midterm elections, of course.

Yes, I think the other part of your question was GTA relative to like Fortnite and others. I would say that GTA from a download kind of assumption perspective, it is bigger. And of course, the number of downloads is a lot bigger, too. So it's definitely assumed to be bigger in our Q4 guide than Fortnite has been.

Operator

[Operator Instructions] And our next question is going to come from the line of Jeff Van Rhee with Craig-Hallum Capital Group.

Daniel Hibshman

Rich, congrats on the quarter. This is Daniel on for Jeff. Maybe just in terms of the component upgrades you talked about in the network in terms of efficiency that you're thinking you can put through that way. I'm assuming what you're talking about there is that said, replacing servers wholesale, you can sort of selectively replace individual components and save on CapEx. Just is that something you've done before? Any thoughts on how meaningful that could be? Is that part of what already saved here in terms of the 2026 CapEx guidance coming down? Just any further thoughts or elaboration on that would be great.

Kip Compton

Yes. I'll comment, and then Rich can maybe comment on how it fits into the CapEx guide. Generally speaking, the capital that we spend on our network is to expand it. We generally do not have significant capital going to repair or upgrade existing servers. We're typically expanding it. There are cases, especially given the current supply chain issues around things like memory where we are looking at component upgrades that enable us to get more out of memory that we already have.

But in general, you should interpret the CapEx that we spend on our network as expanding our capacity. And now I'll also comment, though, that beyond the capital and growing the server and the network base, we regularly improve our software. And we've seen some significant improvements in our efficiency for various services this year. As services scale up, our engineering team looks at how they can make them more efficient and often that provides us with an advantage as well. But Rich, maybe you can comment on the CapEx guide.

Richard Wong

Yes. Our CapEx guide for the full year is now at 10%. I think in the prior, we had guided to 10% to 12%. I would say that from a network perspective, I think we're -- because we have one network that supports all of our products, we're just like a lot more efficient, and we can use a lot of off-the-shelf hardware. That gives us a lot of more flexibility in terms of how we kind of want to build out our network. And so whether it's going to be components or whether it's going to be wholesale servers across the board, we have flexibility in how we scale and grow.

And that's actually playing out. If you look at what we did for this past year, almost 70% of our CapEx was in the first half and we had chosen to do a combination of like regular wholesale purchases of server, we also chose to do some component upgrades. That actually helps us a lot on and I think it's because we have a software configured kind of infrastructure, and it makes us more efficient than our peers.

Daniel Hibshman

Yes. And then just one follow-up in terms of AI traffic and this use case that was discussed in particular on this call with a 7-figure win. Just when we talk about that as agentic traffic, just -- how does that work? So if this is primarily security, not delivery or compute, if I'm correct, that's what I got it from the call. If it's security, that primarily, I think it is on sort of a software subscription basis. Is it price based on volume of traffic. To what degree does sort of the volume of traffic drive the pricing relative to the sort of just being a flat subscription for this?

Kip Compton

I don't want to comment on any specific deal. Broadly speaking, the majority of our security revenue is based on subscriptions, but those subscriptions do increase in price based on the amount of volume of traffic that's covered by the subscription. So that does provide us with growth of revenue with the volume of traffic. But I don't know, Rich, if you have further insight.

Richard Wong

No, I think that's absolutely right. I mean I think it is pseudo subscription based on consumption, how much RPS request that we expect to kind of process for our customers.

Operator

[Operator Instructions] Our next question comes from the line of [indiscernible] with Piper Sandler.

Unknown Analyst

I know you guys aren't guiding out to 2027, but kind of given the transition here, how should we start thinking about kind of balancing investments with the go-to-market and increasing the product velocity with kind of profitability just kind of moving forward?

Richard Wong

Yes. I think that we had Investor Day in September, we laid out a framework out to 2029 that where we're targeting. I think with Scott's announcement today, nothing really kind of changes from -- or deviates from what we've talked about. We do think that we're going to continue to make investments in the right areas around product but also go to market and we want to continue the expansion in APAC and LatAm that we've talked about. So I don't think, from my perspective, anything really changes. We're looking longer term for that 2029 objectives and seeing how we get there. And so we're keeping kind of line in the ball and continue the same path that we're on.

Unknown Analyst

Got it. And really, I appreciate kind of the history of the multiproduct adoption here, some kind of big increases, especially with the 4-plus products. Can you talk about what's kind of driving that? And what are some ways that you can keep pushing that higher?

Unknown Executive

Yes. I think that what's driving it, I would say it really starts with a real good product offering. I think that from -- over the past kind of even 24 to 36 months, we've launched a number of products at parity or better than our peers. And so having that is very helpful in terms of existing customer base. But that alone doesn't do it, right? I think the investments we made on the go-to-market side, bringing in sellers who know how to sell security was really important, bringing in people who know how to do the platform approach and changing the way our commissions plan.

So I would say it's both the combination of having a broader product portfolio, combined with a real go-to-market motion adjusted for a platform approach is what's really driving it. I think that as I look forward, we've talked and announced some AI products as well. And I just think that we're on that trajectory where as we launch more products and as we get deeper with our customers and as we add more value, that opportunity just gets bigger and bigger for us.

Operator

Thank you. I would now like to hand the conference back over to Kip Compton for closing remarks.

Kip Compton

Thank you. I want to thank our Fastly employees for all of their contributions, our customers for their trust and partnership and our investors for their continued support and interest. Thank you, and have a great day.

Operator

This concludes today's conference call. Thank you for participating. You may now disconnect. Everyone, have a great day.