Every time you tap play, open a site, or launch a game and it just “appears,” there's a quiet network behind it putting content — and now the computing too — as close to you as possible, instead of sending every request halfway around the world to one central data center. This is the story of CDNs and edge computing — infrastructure that's shifting from a “move bits cheaply” business into a battleground over security and AI.
Contains
Theme index· base 100 · USD total return
No index history for this theme yet.
News & notes movingEdge & Content Delivery
Edge & Content Delivery▲impact 4
Cloudflare Says AI Agent Traffic Tops 50% as Q2 Revenue Jumps 36%
Cloudflare said more than 50% of the traffic flowing across its network was nonhuman in the second quarter of 2026, reflecting rapid growth in AI agents and machine-to-machine traffic. The company ended the quarter with more than 7.4 million developers on its Workers platform, adding nearly two million in the quarter alone, and management said Workers remains one of the fastest-growing parts of its business. Cloudflare also introduced Monetization Gateway, Wallets and cloudflare.pay during the quarter to support transactions between AI agents, businesses and content providers. Second-quarter revenues increased 36% year over year, while dollar-based net retention reached 120%, and management said its total addressable market has expanded to more than $300 billion from $30 billion at the time of its IPO. The Zacks Consensus Estimate points to Cloudflare revenue growth of 32.5% in 2026 and 28.6% in 2027, with 2026 earnings estimated at $1.26 per share, up 35.5% from the prior year.
Xeal Launches Laitent, World's First Edge Inference Compute Network Using Idle EV Charging Capacity
Xeal launched Laitent, which it calls the world's first edge inference compute network using idle EV charging capacity, tapping more than 200MW of permitted, installed electrical infrastructure across 1,600+ properties. Xeal, a member of NVIDIA Inception, plans to deploy over 100,000 NVIDIA GPUs alongside EV charging infrastructure, and has secured partnerships with Rafay Systems for AI infrastructure orchestration, Spectrum Business for dedicated enterprise-grade fiber, dozens of real estate and property managers, and a Tier 1 inference provider for up to 5MW of compute. The first Laitent Pod will be brought online with partner JVM Realty by the end of 2026. Each Laitent Pod is about the size of one parking space, contains up to 48 NVIDIA Hopper or Blackwell Ultra GPUs, requires no water hookup, and runs quiet at less than 65 decibels, offering sub-20ms latency in metro areas. Xeal said EV charging sites typically operate at less than 10% of permitted capacity, and it taps the remaining 90% for compute, with property owners able to add as much as $1m in property value for little-to-no upfront investment. Looking beyond the initial 200MW of installed charging capacity, Xeal plans to unlock over 1GW of existing headroom across real estate and EV charging deployments.
Fastly Posts 23% Revenue Growth as Security Sales Jump 43%
Fastly reported quarterly revenue of $183 million, up 23% year over year, with security revenue climbing 43% and delivery services revenue rising 17%, CFO Rich Wong said at a Piper Sandler conference. Remaining performance obligations increased 38% year over year, and the current portion of RPO rose 44%, which Wong attributed to stronger customer commitments amid component shortages. Wong credited the security growth to an expanded portfolio that now spans five security products, up from a single web application firewall, and to improved sales execution under Scott Lovett, who joined as head of sales in mid-2024. Bot management and DDoS protection each posted triple-digit year-over-year growth, and net retention reached a four-year high of 117%. Fastly also said it will hold an investor day at Nasdaq MarketSite in New York to discuss a longer-term operating model and its outlook over the next three years, though it does not plan to provide fiscal 2027 guidance at the event.
Lenovo and MemryX Sign Saudi Edge AI Memorandum Tied to Vision 2030
Lenovo Group and MemryX signed a memorandum of understanding to expand sovereign edge AI solutions in Saudi Arabia, supporting the country's Vision 2030 program through secure, large-scale deployment of AI at the edge. The two companies already have live edge AI deployments in the country, which form the base for a wider national rollout. Lenovo Group, a HK$419.3 billion hardware and services player, is using the Saudi push to lean further into infrastructure-grade computing rather than just end-user devices, a tilt that lines up with the bull case that its Infrastructure Solutions Group can become a larger, more profitable pillar alongside PCs. The bear-side concern is that heavy AI and infrastructure investment could weigh on profitability if these deployments stay narrow or prove costly to scale, making future disclosures on the Saudi and broader Middle East projects the practical test, with a key focus on whether the Infrastructure Solutions Group moves closer to break-even as the edge deployments expand across more clients and use cases.
Arista Networks Earns Zacks Rank #1 as AI Networking Demand Drives Upgraded Outlook
Arista Networks has been named a Zacks Rank #1 Strong Buy stock after posting a strong beat-and-raise quarter in early August, with its Q3 earnings estimate surging 19%, its FY26 EPS estimate up 11%, and its FY27 outlook 14% higher. The AI infrastructure company, which counts Microsoft and Meta among its largest clients, beat the Q2 earnings estimate by 15% on August 4 and reported its first $3 billion quarter, boosted by its Arista 2.0 platform strategy. Arista more than doubled its revenue between 2022 and 2025, from $4.38 billion to $9.00 billion, and more than doubled its GAAP earnings over the same period, from $1.07 a share to $2.75 per share. The company is projected to grow revenue another 39% in FY26 and 26% in 2027 to reach $15.80 billion, while adjusted EPS is expected to rise 36% in 2026 and 24% next year to $4.99 per share. Arista holds $13.3 billion in cash and equivalents against near-zero debt, and its average Zacks price target implies 30% upside from current levels.
Cloudflare Launches Disallow AI Training Controls for Websites
Cloudflare introduced new "Disallow AI Training" and AI crawler controls for website owners on 15 September 2026, letting site operators block AI training on their content while keeping pages indexable for traditional search engines. The company set criteria for "Accountable" crawlers and said it is participating in IETF efforts on open AI access standards. Cloudflare runs a global cloud network that helps businesses secure and deliver their websites and applications, and the company says the controls fit its role as an internet policy layer rather than just a content delivery utility. The launch does not yet answer how Cloudflare will charge for this layer of AI governance, or whether it becomes a paid tier versus table stakes in a market that also includes AWS and Akamai. Simply Wall St's analysis points to a $334 fair value for Cloudflare.
Cloudflare Launches Disallow AI Training Setting and Accountable Crawler Designation
Cloudflare announced a new Accountable designation for AI crawling and launched Disallow AI Training, a setting that lets any website refuse AI training while staying fully in search results. Apple, Google, and Microsoft are now labeled as Accountable and have shown they meet Cloudflare's criteria or given specific timelines to meet them, while other mixed-use crawlers are blocked if site owners block training. Mixed-use crawlers, which collect content for both a search index and AI training, now make up 36.6% of verified crawler traffic on Cloudflare's network, the single largest category, and fewer than 1% of website owners block search crawlers while 17% restrict AI training. Cloudflare is replacing its Block AI Bots switch with three independent controls for search, AI training, and AI agents, and replacing Managed Robots.txt with Bot Preference Sync, which applies site owners' crawling preferences across all supported crawlers automatically. The company said every operator it designates as Accountable must give site owners a way to opt out of AI summaries, and by early next year its goal is to let site owners control how much of their content is included, in one place.
Harmonic Launches cOS SensAI for Broadband Networks
Harmonic unveiled cOS SensAI on September 8, a vendor-agnostic software layer that pulls scattered network data from routers, modems, and in-home devices into one coherent operational picture for broadband providers. Analysys Mason estimates the addressable market for such a tool at roughly $2.2 billion a year among tier two and smaller operators, though Harmonic has disclosed no revenue attached to SensAI yet and the only customers named publicly, Hotwire Communications and GCI, are early reference accounts. The launch follows a quarter in which Harmonic reported Broadband revenue up 54% year over year for the period ended July 3, with the Rest-of-Market segment growing 44% and supplying about 60% of bookings, while backlog and deferred revenue reached $587.6 million, up 71%, and cash climbed to $231.9 million after the $145 million sale of its Video business to MediaKind on June 16. The cOS platform was running across 161 customers and 48.2 million connected devices by that point. Skepticism remains: the stock fell 6.25% on the launch, the top two customers accounted for 63% of Broadband revenue in the quarter, up from 58% the prior quarter, and the company posted a GAAP net loss of $0.02 per share, weighed down by a $19.4 million loss from discontinued operations tied to winding down the divested Video business.
Comcast Rolls Out Fastly-Powered AI Edge Platform and Rural Broadband Builds
Comcast announced a Fastly-powered edge platform integrated across its nationwide network, targeting ultra-low latency services. The company reported new broadband buildouts in Glades County, Florida and Jackson County, Michigan, reaching thousands of previously under-connected locations. Management highlighted AI-focused edge capabilities, including support for streaming, gaming, and emerging AI services delivered closer to end users. The rollout embeds Fastly's software into more than 200 edge compute centers, extending Comcast's Connectivity & Platforms push into broadband innovation, intelligent WiFi, and bundled services. Comcast operates as a global media and technology group, a dual position that lets it test AI-focused connectivity services on its own network footprint while also selling them to customers.
MIIT releases 15th Five-Year Plan for information and communications; telecom sector bucks the trend to close up 0.63%
The Ministry of Industry and Information Technology released the 15th Five-Year Plan for the development of the information and communications industry, setting the goal of building a new-generation communications network by 2030, with cumulative investment in information infrastructure reaching 3.8 trillion yuan, and the target for intelligent computing power jumping from 1,590 exaflops to 9,800 exaflops. It also proposed launching 6G commercial use at an appropriate time and advancing 6G standard development and technical trials. Driven by this industry catalyst, the Shanghai Composite Index fell 1.18% in the previous trading session, while the communications sector rose 0.63% against the trend. The Southern Fund's Communications ETF saw a turnover rate of 6.93% and turnover of 24 million yuan, while the CSI Communications Services Index it tracks gained 0.63%. At the industry level, institutional data shows that the communications sector rose 73.59% in the first half of 2026, significantly outperforming the CSI 300. Shipments of 800G optical modules continued to expand, and 1.6T entered the early stage of large-scale deployment. Third-party institutions forecast that the global AI optical module market will reach 26 billion US dollars in 2026. In addition, on September 11, the US Federal Communications Commission finalized relevant rules, and several leading domestic optical module makers such as Zhongji Innolight and Eoptolink were not included for the time being, easing overseas policy concerns.
Cloudflare Raises Long-Term Growth Target to 50% on Agentic AI Bet
Cloudflare now sees agentic artificial intelligence as a growth opportunity that will dwarf its previous drivers, and has raised its long-term "North Star" growth target to 50% from 40%. Chief Financial Officer Thomas Seifert, speaking at an investor event, said the company's network architecture is positioned to support low-latency, secure and distributed AI inference workloads, and that Cloudflare expects to exit the year with momentum toward that objective. Demand is showing up through the Workers developer platform, where the company added 2 million developers quarter over quarter, and through variable consumption revenue, with companies including Lovable, Figma, Anthropic and OpenAI building on Workers; Cloudflare estimates that about 80% of significant AI enterprise infrastructure customers use its services. Seifert said Cloudflare is focused on inference rather than training the largest models, and that capital expenditures could rise but any increase would be expected to coincide with higher revenue. The company is also building monetization infrastructure for the agentic economy, including "pay per crawl," a Monetization Gateway, Cloudflare Wallets and identity tools, working with partners including Visa, Mastercard, American Express, PayPal and Coinbase, though Seifert acknowledged that significant technical work remains and that Cloudflare cannot yet process transactions at the millions-per-second scale it believes may be necessary.
Nvidia Partners With Australian Cloud Firms on Up to 2 Gigawatts of AI Capacity
Nvidia is working with Australian cloud partners on up to 2 gigawatts of AI computing capacity by 2027, expanding the chipmaker's ecosystem beyond selling processors into designing AI factories. The program involves sites designed to host several generations of Nvidia DSX infrastructure, with partners including IREN, Sharon AI, Firmus, ResetData, Megaport, CDC, NEXTDC and AirTrunk. IREN plans to apply the DSX reference architecture across its development portfolio, including its 800-megawatt Bundey campus in South Australia, while Sharon AI plans to deploy up to 68,000 Nvidia GPUs and other partners are preparing data centers, powered shells and connectivity for accelerated computing. Atlassian plans to use Nvidia's Nemotron open models, datasets and tools to customize multimodal applications and enhance Rovo, its workplace AI platform. The 2 gigawatts represents an upper target rather than completed capacity, and partners must still secure power, financing, equipment and customers while meeting construction deadlines.
Comcast and Fastly Launch Industry-First Edge Delivery Partnership for Xfinity Streaming
Comcast and Fastly announced a next-generation content and application delivery partnership that embeds Fastly's programmable edge software into Comcast's edge compute platform, pushing live streaming content closer to customers' homes for millions of Xfinity members. The deployment creates an industry-first hyper-local service layer across Comcast's nationwide network, which includes more than 200 AI-powered, data-processing edge compute centers distributed across the country. Initial results have demonstrated substantial gains in data delivery and significant reductions in round-trip latency compared to traditional content delivery mechanisms, with the platform designed to handle massive live streaming traffic, real-time security, edge compute, and AI services. Elad Nafshi, Chief Network Officer at Comcast, said the stakes will never be higher than during the upcoming fall football season, when a single streaming game can drive record network traffic, and that integrating Fastly establishes a new approach for delivering the biggest moments in streaming. Kip Compton, Chief Executive Officer at Fastly, said extending Fastly's platform into Comcast's network brings more local performance and real-time traffic management to streaming, compute, and security applications while infusing intelligence into the network layer for the AI era.
Cloudflare Shares Jump 10.5% After CFO Raises Long-Term Growth Outlook
Cloudflare Inc. raised its long-term growth outlook to 50 percent annually from 40 percent at present, sending its shares up 10.51 percent on Wednesday to close at $10.51 apiece. Speaking at the Goldman Sachs Communacopia + Technology Conference 2026, Chief Finance Officer Thomas Seifert said the company is moving beyond a traditional software-as-a-service model to become an agentic AI platform, and that its future opportunity could be much larger than its business, with revenues projected to grow 36 percent year-on-year. In the second quarter, Cloudflare grew revenues 36 percent to $696.1 million from $512 million a year earlier, while its net loss widened 237 percent to $169.98 million from $50.4 million amid a 206 percent higher operating loss. Earlier this month the company partnered with OpenAI on the early launch of Vulnerability Discovery and Remediation, available in early access through Cloudflare Managed Defense, which CEO Matthew Prince said shifts defense strategy from chasing patches one vulnerability at a time to an automated approach. Institutional interest also rose in the second quarter, with 87 hedge funds holding positions in the stock, up from 84 in the prior quarter, and their combined holdings climbing 11.8 percent to $3.97 billion from $3.55 billion.
Penguin Solutions to Deploy AI Platform for Lektra's Distributed Data Centers
Penguin Solutions, Inc. (Nasdaq: PENG) announced it has been selected by Lektra to deploy and optimize its AI micro data centers, supporting Lektra's pioneering distributed cloud network powered entirely by carbon-free energy. Penguin Solutions will provide validated reference designs, AI-optimized compute, and expert services to scale Lektra's EdgeScale AI data centers, which operate at capacities of 20 MW or less and have eight sites already deployed. The collaboration leverages Penguin Solutions' Full-Stack AI Factory Platform, including NVIDIA-based hardware and engineering expertise, to enable incremental GPU capacity expansion for cloud providers and energy hosts. Lektra's CEO Karl Andersen highlighted the partnership's role in advancing distributed energy production, or Grid 2.0, while Penguin Solutions' CEO Kash Shaikh emphasized the need for new deployment approaches as AI demand accelerates.
Sceye's Stratospheric Platform Returns to US After Record Flight
Sceye, a U.S. aerospace company, announced the successful return of its Service Test 1 (ST1) mission to the United States after a record-breaking trans-Pacific stratospheric flight from New Mexico to Japan and back. The journey, which lasted 30 days and covered nearly 30,000 km, included operating for over a week in Japanese airspace, where Sceye and partner SoftBank Corp. verified Direct-to-Device connectivity to unmodified devices, HAPS-based communications with drones, and edge computing from the stratosphere. ST1 launched on August 9, 2026, arrived in Japan 13 days later, and re-entered the continental U.S. on September 4, 2026, before landing near Sceye's headquarters in Moriarty, New Mexico. The mission demonstrated mobile broadband services including text, voice, internet, and video streaming, and achieved a world-first by installing a mobile core network and web server on a HAPS, cutting latency by over 40% compared to cloud processing. Sceye's platform, designed to cover an area equivalent to about 500 terrestrial towers, marks a milestone toward commercial deployment of stratospheric infrastructure.
Qualcomm shares rose 6.02% intraday after announcing a multi-generation collaboration with Amazon on customized silicon for large-scale AI data centers, focused on inference. The two companies will also work on optical connectivity extending up to 1.6T, leveraging Qualcomm's SerDes and optical DSP technologies to support bandwidth demands inside AWS networks. Qualcomm plans to deepen its use of AWS infrastructure, including Amazon Bedrock, for electronic design automation workloads to shorten chip design cycles. CEO Cristiano Amon said data center infrastructure will require advances in both computing and connectivity. Qualcomm unveiled the Dragonfly C1000 data center CPU in June and said Meta Platforms would use it from 2028 production, while targeting $15 billion in data center sales in fiscal 2029. Neither company disclosed terms for the Amazon arrangement.
Ambarella Guides Q3 Revenue to $115M-$124M, Unveils X7 Accelerator
Ambarella reported fiscal Q2 revenue of $108.1 million, up 13.2% year-over-year, with non-GAAP EPS of $0.18, and guided fiscal Q3 revenue to a range of $115 million to $124 million, citing favorable seasonality. The company also announced its new X7 stand-alone AI accelerator, now sampling, and two indirect-channel partnerships with CapGemini and a 7-year agreement with Macnica, each expected to contribute about $0.5 billion in revenue over time. Management raised its 5-year rolling serviceable addressable market forecast from $8.5 billion in fiscal year 2027 to $22.9 billion in fiscal year '32, representing a CAGR of about 20%, driven largely by the edge infrastructure opportunity. CFO John Young noted Q2 non-GAAP gross margin of 59.3% and operating expenses of $57.4 million, while CEO Fermi Wang said rising memory costs will be passed to customers to maintain the long-term gross margin target of 59% to 62%. The company also authorized a new $50 million share repurchase program valid through June 30, 2027.
Equinix, NVIDIA, Together AI Launch Distributed AI Inference Program
Equinix has expanded its collaboration with NVIDIA and announced a new partnership with Together AI to deliver Equinix Inference Exchange, a distributed AI inference program for global enterprises. The offering combines NVIDIA Enterprise Reference Architectures, Together AI's inference platform, and Equinix's global infrastructure to improve deployment speed, flexibility, and cost efficiency as enterprises move AI workloads from experimentation to production. Equinix's footprint includes more than 280 data centers across 77 metros, 230 cloud on-ramps, and over 10,500 interconnected businesses, with eight of the top 10 AI model providers and nine of the top 10 AI cloud providers already deployed. The service is expected to become available in the first quarter of 2027, and Equinix has not disclosed financial targets for the offering.
ASUS Expands AI Infrastructure Ecosystem with Intel and AMD
At ASUS AI Tech 2026, ASUS announced the expansion of its AI infrastructure ecosystem through strategic collaborations with Intel, AMD, and partners including Aleria, Foxlink, IBM, Samsung, Schneider Electric, and WD. The company showcased new servers powered by Intel Xeon 6 and AMD EPYC 9006 Series processors, as well as AMD Instinct MI350P accelerators, designed for enterprise and AI workloads. For industrial edge applications, ASUS introduced the RUC-2000 series embedded computer with Intel Core Ultra Series 3 processors, offering up to 180 AI TOPS in a rugged, fanless design. These products aim to support AI deployment from cloud to edge, ensuring data confidentiality and compliance in sectors like healthcare and manufacturing.
iPronics Raises $125 Million to Scale Optical Networking for AI Data Centers
iPronics, a pioneer in silicon photonics-based optical circuit switching for AI infrastructure, has announced $125 million in Series B funding co-led by Maverick Silicon and Light Street Capital, with NVIDIA joining the round. The Series B brings iPronics' total funding to $177 million, and the company will use the funds to scale operations, accelerate commercial deployment, and meet demand for its iPronics Optical Networking Engine (ONE), a rack-mounted optical switch for AI. New investors include Triatomic Capital, Bosch Ventures, Catalight Capital, European Innovation Council Fund, The Tate Family Trust, Fine Structure Ventures, Amadeus Capital Partners, Build Collective, and Criteria Venture Tech. Manish Muthal of Maverick Silicon and Geoffrey Tate, an advisor to Light Street Capital, join the board, while Young Sohn of Catalight Capital becomes an advisor. iPronics has also opened an office in Santa Clara, California, to strengthen its U.S. presence.
Advanced Micro Devices, Cisco Systems, and Saudi Arabia's HUMAIN announced on Monday that AI computing capacity built around AMD's Instinct MI355X chips and Cisco's networking hardware is now live in the kingdom, serving paying customers. The system, which combines AMD's Instinct MI355X graphics processors and EPYC server chips with Cisco's Silicon One networking chips and 800G optics, is being used by HUMAIN, the AI arm of Saudi Arabia's roughly $1 trillion Public Investment Fund, to sell GPU capacity to customers inside and outside the kingdom. The joint venture, announced in November 2025, has moved from announcement to live production workloads in under a year, with HUMAIN CEO Tareq Amin highlighting the milestone as proof of the company's ability to build and deliver advanced AI infrastructure at scale. A second phase of up to 250 megawatts, using AMD's newer Instinct MI400 series chips, is due to begin deployment in 2027, with the joint venture still targeting up to 1 gigawatt of capacity by 2030. However, every future chip shipment still depends on individual U.S. export approval, as Washington authorized HUMAIN to buy the equivalent of up to 35,000 Nvidia Blackwell chips in November 2025 on a case-by-case basis, and Saudi Arabia was left out of a July 2026 reclassification that moved the United Arab Emirates into the top tier of U.S. export trust. AMD's data center revenue hit $5.8 billion in the first quarter and accelerated to $6.7 billion in the second quarter, up 107% year-over-year, though it still trails Nvidia's roughly $89 billion in second-quarter AI chip revenue, and AMD's stock has more than doubled over the past year, pushing its market cap toward $900 billion by mid-2026.
5G Edge Cloud Market to Grow by $12.90B Through 2030
The global 5G edge cloud network and services market is projected to expand by USD 12.90 billion from 2025 to 2030, registering a compound annual growth rate of 20.6%, according to a new report from ResearchAndMarkets.com. Growth is driven by ultra-low-latency applications, the convergence of AI and distributed computing at the network edge, and rapid adoption of private 5G networks for Industry 4.0. The report segments the market by network type, component, business segment, and geography, covering key countries including the United States, China, Japan, and Germany. It profiles approximately 25 vendors, including Oracle, Hewlett Packard Enterprise, Intel, Huawei, Cisco, and 15 other key players. Opportunities center on decentralized generative AI, sustainable infrastructure, and sovereign edge clouds for data residency and regulatory compliance.
On August 28, A-share CPO concept stocks opened lower and moved higher. Star-Net Ruijie, Hefei Metalforming Intelligent Manufacturing, and Qingshan Paper hit limit up, while Shenzhen Microgate Technology, Shengyi Electronics, Optowide Technologies, and Advanced Fiber Resources followed higher. On the news front, the Shanghai Municipal Science and Technology Commission released the 2026 application guidelines for the new-generation optical communications technology project under the Intelligent Computing Optical Network program. Approved projects will receive 40 percent of fiscal funding at launch, another 40 percent after completing milestone targets, and the final 20 percent after meeting all assessment indicators. Overseas, Nvidia reported second-quarter results for fiscal 2027, with revenue of 96.2 billion dollars, up 106 percent year on year, and data center revenue of 89 billion dollars. It also guided to a midpoint of 108 billion dollars in revenue for the third fiscal quarter and raised its fiscal 2028 revenue growth forecast to 70 percent. After the earnings release, Nvidia shares jumped 8.7 percent, giving it a market value of 5.49 trillion dollars, and more than ten investment banks including Morgan Stanley and Citigroup raised their price targets. On the technology side, Nvidia's Spectrum-X Ethernet photonic switching platform has entered full mass production, marking CPO's move into large-scale manufacturing. LightCounting forecasts that global Ethernet optical module sales will grow 73 percent year on year in 2026. Brokerage analysts believe the industry's prosperity has medium- to long-term support, with the four major overseas cloud providers guiding to capital expenditure of 720 billion to 745 billion dollars in 2026.
Cloudflare's Workers Platform Drives Growth with 7.4 Million Developers
Cloudflare's Workers developer platform is becoming a key growth engine, with the company reporting over 7.4 million developers on the platform by the end of the second quarter of 2026, adding nearly two million in that quarter alone—more than the 1.5 million added during all of 2025. The platform is gaining traction among enterprise customers, with more including Workers in pool-of-funds contracts; for instance, an APAC technology company signed a $4 million Workers contract after an $8.7 million Application Services deal, and another signed a $6 million contract for AI agent capabilities. Cloudflare says Workers has moved beyond adoption to meaningfully contribute to revenues, driving consumption alongside security and networking products. The rise of AI agents could further boost the platform, with Zacks Consensus Estimates projecting 32.3% and 28.5% year-over-year revenue growth for 2026 and 2027, respectively. However, Cloudflare faces tough competition from Palo Alto Networks, whose SASE segment grew 40% year over year in fiscal Q3 2026, and Zscaler, which ended its fiscal Q3 with over 700 Zero Trust Everywhere enterprises, up from 550. Cloudflare's stock is up 40.7% year-to-date, trading at a forward price-to-sales ratio of 28.87 versus the industry's 3.91, and carries a Zacks Rank #3 (Hold).
Veea and Totalplay Advance Secure AI Business Services in Mexico
Veea Inc. announced that its VHC25 VeeaHub has completed Totalplay's technical qualification, a key milestone toward offering SecureConnect and VigiLynx to Totalplay business customers in Mexico. The companies have signed a collaboration agreement to combine Totalplay's fiber network and managed services with Veea's edge platform. Totalplay, with approximately 5.7 million subscribers and 19.5 million homes passed as of June 30, 2026, will integrate Veea's platform to extend services beyond network functions to cameras, sensors, and AI workloads. This collaboration complements Veea's existing launch over Telcel's 5G network, where it has begun generating revenue under the Internet Gestionado brand. Veea believes Mexico is a key market for validating this operator-led model and expanding across Latin America.
Baird Strategist Sees Software and Cybersecurity Leading Next Tech Phase
Baird tech strategist Ted Mortonson forecasts that hyperscaler capital expenditures will reach about $800 billion this year and could hit $1.1 trillion by 2027, driven by demand outstripping supply in compute and early-stage agentic adoption in enterprises. He notes that while semiconductor stocks have surged 61% and software indices are flat, the software sector is seeing short covering and institutional interest in system-of-record names. The security industry has been transformed by the introduction of Mythos, boosting names like Cloudflare, Crowdstrike, and Palo Alto, though these stocks now trade at nearly 30 times enterprise value to revenue, with some insider selling on the periphery. Mortonson cautions that changes in security architecture show up in pipeline rather than annual recurring revenue, and the street may not yet grasp how quickly that pipeline converts to orders.
Akamai Cloud Infrastructure Revenue Jumps 39% on AI Demand
Akamai Technologies reported strong growth in its Cloud Infrastructure Services segment, with second-quarter 2026 revenue of $99 million, up 39% year over year. The company said demand is being driven by the rapid expansion of AI workloads, particularly inference and AI-agent applications that require low-latency computing closer to end users. Akamai's existing GPU capacity is completely sold out, and the company plans to invest up to $500 million in additional capital expenditure to meet demand. In fiscal 2026, Akamai has signed more than $2.8 billion in multi-year CIS commitments, including a more than $600 million, four-year agreement with a U.S. tech company for robotics-related infrastructure. The company expects CIS revenues to grow at least 50% in constant currency for full-year 2026.
Enterprise Networking Market to Reach $193.77 Billion by 2030
The enterprise networking market is projected to grow from $124.59 billion in 2025 to $193.77 billion by 2030, a CAGR of 9.2%, according to a new report from MarketsandMarkets. The firm attributes the expansion to enterprise modernization of network infrastructure for hybrid work, multi-cloud environments, and stronger cybersecurity. North America is expected to hold the largest share, while Asia Pacific is the fastest-growing region. Key players include Cisco Systems, Broadcom, Hewlett Packard Enterprise, Juniper Networks, Arista Networks, Huawei, Extreme Networks, Fortinet, Cloudflare, and Alcatel-Lucent Enterprise.
Video Streaming Infrastructure Market to Reach $91.66 Billion by 2031
The global video streaming infrastructure market is projected to grow from USD 48.86 billion in 2026 to USD 91.66 billion by 2031, a compound annual growth rate of 13.41%, according to a new report from ResearchAndMarkets.com. The market was valued at USD 43.32 billion in 2025. Growth is driven by demand for low-latency live video, AI-enabled media workflows, cloud-native platforms, and expanded content delivery network capacity. Non-linear and video-on-demand streaming accounted for 52.47% of the market in 2025, while live video streaming is forecast to grow at a 17.33% CAGR from 2026 to 2031. North America held 36.63% of the market in 2025, and Asia-Pacific is expected to be the fastest-growing region with a 15.98% CAGR through 2031.
Cathie Wood's Ark Invest Pours $27 Million Into Nvidia
Cathie Wood's Ark Invest made another aggressive rotation toward artificial intelligence last week, adding nearly $30 million of Cloudflare and $27 million of Nvidia while unloading almost $60 million of Deere. Cloudflare was Ark's largest purchase, with the firm buying about $29.9 million across its funds after the stock surged more than 16% following stronger-than-expected second-quarter results and guidance. Ark also purchased a combined $26.9 million of Nvidia across ARKK, ARKF, ARKQ, ARKW and ARKX after Nvidia confirmed it was working with lenders on financing tied to as much as $500 billion of AI infrastructure development. The firm added roughly $24.5 million of Cerebras Systems, about $23 million of Rocket Lab, $16.7 million of Broadcom, and $9.7 million of Teradyne. On the selling side, Ark sold approximately $59.9 million of Deere, cut $22.8 million of Shopify and $11.7 million of Palantir, and trimmed holdings including Snowflake, Caterpillar and several biotechnology names.
Atlassian and Cloudflare Earnings Defy Enterprise Software Fears
Atlassian and Cloudflare reported strong quarterly results that pushed back against Wall Street's 'SaaSpocalypse' fears over slowing enterprise software demand. Atlassian posted adjusted earnings per share of $1.87, beating the $1.50 consensus, and grew its Remaining Performance Obligations 44% year-over-year to $4.8 billion, while its CEO announced plans to personally buy up to $250 million in company stock. Cloudflare delivered second-quarter revenue of $696.1 million, up 36% year-over-year and above the $666 million estimate, and raised its full-year 2026 revenue guidance to between $2.864 billion and $2.870 billion. Hedge fund holdings in Atlassian fell from 57 funds in the fourth quarter to 44 in the first quarter, with short interest at 11.20% of the float, while Cloudflare saw institutional accumulation rise from 70 to 84 funds and short interest of just 3.09%.
Nvidia's Spectrum-X Photonics Enters Mass Production
Nvidia has begun full mass production of its Spectrum-X Ethernet Photonics platform, the first co-packaged optics Ethernet switch built for 200G-per-lane volume shipping. The platform delivers 5x lower network power consumption, 64x better signal integrity, and 10x better resiliency at scale by soldering optics directly onto the switch chip. CoreWeave, Lambda, and Oracle are the first customers, with production ramping from May through July before reaching full volume. Nvidia's networking revenue surged 199% to $14.8 billion in fiscal Q2 2026, while Taiwan Semiconductor Manufacturing handles advanced silicon photonics fabrication and Lumentum supplies lasers and optics, with Lumentum's fiscal Q4 revenue more than doubling to $1.01 billion and multi-hundred-million-dollar CPO orders locked in for delivery in the first half of calendar 2027.
Akamai Technologies reported second-quarter revenue of $1.1 billion, beating analyst estimates of $1.09 billion, with adjusted EPS of $1.59 versus the $1.58 consensus. The company highlighted a new $600 million, four-year cloud infrastructure deal with a U.S. technology company and cited AI-driven workloads and security demand, including a recent customer win with CrowdStrike, as key growth drivers. Management reconfirmed full-year revenue guidance of $4.49 billion at the midpoint but lowered full-year adjusted EPS guidance to $6.73, a 0.7% decrease. Operating margin fell to 7.3% from 14.5% a year earlier, while billings rose 4.3% year over year to $1.10 billion. On the earnings call, executives said large cloud deals typically take six to nine months to convert to recognized revenue and that initial colocation and hardware costs can temporarily depress margins before profitability improves within a quarter of deployment.
Spear Invest CIO sees AI opportunities in Cloudflare, CPUs, optics, semi cap equipment
Spear Invest founder and chief investment officer Ivana Delevska says companies that improve GPU and CPU utilization, such as Cloudflare, will benefit as AI demand grows. She also highlights optical components, semiconductor capital equipment, and CPUs as relatively newer bottlenecks, noting that the CPU cycle just turned with AMD and Arm. Delevska expects exponential increase in CPU demand as agentic AI buildout requires more CPUs per GPU than data centers used for model training. She adds that the space theme, including defense, communications, and data centers in space, is another interesting idea for investors.
Akamai Stock Falls 22% in 3 Months as AI Growth Meets Execution Risks
Akamai Technologies shares have declined 22% over three months as investors weigh the company's shift toward AI infrastructure against execution risks. Second-quarter revenue rose 5% year over year to $1.1 billion, with Cloud Infrastructure Services revenue up 39% to $99 million, while Delivery and other cloud applications revenue fell 6% to $395.9 million. Management expects Cloud Infrastructure Services to grow at least 50% in 2026 and overall revenue growth to accelerate into the low teens in 2027, supported by more than $2.8 billion of multiyear commitments signed year to date. However, capital expenditures reached $347 million, or 32% of revenue, in the second quarter, and the company plans third-quarter capital expenditures of $475 million to $525 million, plus up to $500 million to replenish and expand GPU capacity. Non-GAAP operating margin fell five percentage points year over year to 25%, and Akamai currently carries a Zacks Rank #4 (Sell) with a Value Score of D, Growth Score of F, Momentum Score of A, and VGM Score of D.
AI infrastructure stocks surge after strong earnings from CoreWeave, Supermicro
AI infrastructure stocks surged on Wednesday after blowout results from AI server hardware maker Supermicro and neocloud providers CoreWeave and Nebius Group. Nebius stock surged after quarterly revenue topped analyst expectations, while Supermicro reported fourth quarter results that beat analyst estimates on earnings and issued an upbeat forecast that sent shares higher by more than 6%. AI hosting peers Applied Digital and IREN also moved higher after CoreWeave posted quarterly results highlighting accelerating demand and a surging backlog. Lumentum shares gained after the maker of optical and photonic products for data centers posted fiscal fourth quarter revenue that more than doubled to $1.01 billion, and peers Coherent and Ciena both jumped. The memory and storage complex also rose, with the Roundhill Memory ETF up 4%, SanDisk up 4%, and Micron and SK Hynix each up 4%.
Cloudflare reported second-quarter revenue of $696.1 million, exceeding analyst estimates of $664.8 million and representing 35.9% year-on-year growth. Adjusted earnings per share came in at $0.29, beating the consensus of $0.27. The company raised its full-year revenue guidance to $2.87 billion at the midpoint, up from $2.81 billion, and lifted its full-year adjusted EPS outlook to $1.26. Management cited rapid adoption of its Workers developer platform, strong large-customer additions, and growing demand for agentic AI services as key drivers.
Cloudflare Closes $2.175 Billion Zero Coupon Convertible Note Offering
Cloudflare has closed a $2.175 billion offering of zero coupon convertible senior notes due August 2031. The deal provides the company with fresh capital and balance sheet flexibility. The notes carry no interest and convert into equity at a premium, reflecting strong recent stock performance with a 90-day share price return of 66.28% and a one-year total shareholder return of 54.39%. The offering intensifies the debate over Cloudflare's valuation, as the stock last closed at $310.59 while a widely followed fair value estimate pegs it at $167.45, suggesting the shares are overvalued.
Goldman Sachs Says AI Trade Rotates to Optical Connectivity, Picks Lumentum and Coherent
Goldman Sachs Asset Management's Sung Cho says the AI trade is rotating from graphics processors to optical and fiber equipment makers, naming Lumentum and Coherent as key beneficiaries. Cho argues that as AI workloads shift from training to inference, the bottleneck is no longer compute speeds but chip-to-chip and server-to-server communication, driving demand for optical connectivity that is extremely hard to scale. Lumentum reported fiscal Q3 2026 revenue of $808.4 million, up 90.1% year over year, with non-GAAP EPS of $2.37 and operating margin expanding 700 basis points sequentially to 32.2%, while Coherent posted Q3 FY2026 revenue of $1.81 billion, up 20.5% year over year, with its Datacenter & Communications segment contributing $1.36 billion, up 40.6% and now 75% of total revenue. Lumentum shares closed at $813.51, up 599.67% over one year, and Coherent finished at $325.15, up 76.17% year to date.