Why crypto's best infrastructure companies stopped looking like crypto?

Every infrastructure sector in modern history developed using the same playbook. Someone builds raw capacity ahead of demand. The raw resource commoditizes and gets cheaper forever. And the money moves to whoever sells guaranteed, contracted service on top of it. So far, cloud ran it, telecom ran it, even chip makers did it to an extent. Crypto infrastructure is attempting to repeat it right now. Motivation to continue this research came from a news article a couple of weeks ago. Storj, one of the actually working service businesses in decentralized storage, went into Chapter 11 Bankruptcy in July while its own restructuring announcement called the business underneath “strong and right-sized,” held back by “legacy obligations from an earlier chapter,” and said it expects to keep operating without interruption [1]. A crypto company restructuring like a normal business, to keep serving customers like a normal business. And that is not a single case. At least three different companies from the legacy DePIN space were doing some form of restructuring, selling or splitting the business to get decoupled from the token component just this summer [1]. That’s what I want to talk about in Part 2 of a series. Part 1 argued the old DePIN model is dead. DePIN thesis was in building Networks that incentivise ordinary people to crowdsource hardware, bandwidth, and storage, trying to turn resources none of us can negotiate with individually into services someone will actually buy. Part 2 is the evidence of the OG model stopping to work. The pattern, the winners, and where we are on the curve. I. The Pattern The cloud ran the script Amazon launched EC2 in 2006, selling raw compute by the hour. Rent a server, configure it yourself. Table 1 shows the infrastructure layer financials for the past twelve years. The elephant in the room is the fact that AWS is 18% of Amazon’s revenue but 57% of its operating income, and Amazon is about to rebuild the infrastructure layer all over again with roughly $200B of CapEx in 2026, driven by AI [2][3]. Table 1. AWS by year: the infrastructure layer compounding. | Year | AWS net revenue | AWS operating income | AWS share of Amazon operating income | |---|---|---|---| | 2013 | $3.1B | n/a | n/a | | 2015 | $7.9B | $1.9B | ~84% | | 2017 | $17.5B | $4.3B | >100% (international losses exceeded North America retail profit) | | 2019 | $35.0B | $9.2B | 63% | | 2021 | $62.2B | $18.5B | 74% | | 2023 | $90.8B | $24.6B | 67% | | 2024 | $107.6B | $39.8B | 58% | | 2025 | $128.7B | $45.6B | 57% | Source: Amazon 10-K segment reporting, 2013-2025 [2][3] The services built on top are the ones getting the value accumulated from the infrastructure, not the infrastructure companies themselves. Netflix pays Amazon an estimated $1B+ a year for effectively all its computing, about 2 to 3 cents of every dollar Netflix earns (an estimate; Netflix delivers video over its own CDN, AWS runs the compute and storage) [4][5]. Snowflake books $4.5B in product revenue and owns zero data centers. Datadog $3.4B, Zoom $4.9B, Airbnb $12.2B, all running on rented infrastructure their customers never see [6][7]. The whole market is renting raw servers by the hour (Infrastructure as a Service or IaaS) against finished software delivered over the internet (Software as a Service or SaaS). Total public cloud spending hit $595.7B in 2024, is forecast at $723.4B for 2025, and passes $1 trillion in 2027 on Gartner’s projection [8]. The service layer is structurally larger and the absolute gap keeps widening, but the infrastructure layer is currently the faster-growing one. GPUs and power are scarce again, so profits are flowing down to the infrastructure. Likely, if we were able to isolate the GPU demand growth, we would see overall cloud commoditisation even further. Also, if the trend continues, GPUs and power will be an example of the same commoditisation in a couple of years. Telecom runs the same script, just different wires T-Mobile paid up to $1.35 billion for the parent company of Mint Mobile, a phone brand that owns no towers, no spectrum, no radios, but with Ryan Reynolds [9]. Over 2,100 such carriers operate in about 100 countries, an $89-99B global market of phone companies that rent everything (the industry calls them MVNOs) [10]. Table 2. Towers vs wireless services, US, 2010-2024. | Year | Big-3 tower operators combined revenue (AMT + CCI + SBA) | US wireless service revenue (CTIA) | ARPU (CTIA) | Ratio (service : towers) | |---|---|---|---|---| | 2010 | $4.5B | $159.9B | $47.53 | 36x | | 2015 | $10.1B | $191.9B | $44.65 | 19x | | 2020 | $16.0B | $189.9B | $35.31 | 12x | | 2024 | $19.4B | $224.9B | $33.36 | 12x | Sources: company annual reports [11][12][13]; CTIA Annual Wireless Industry Survey [14]. Tower revenues include international operations, CTIA is US-only; treat the ratio as illustrative. Crown Castle 2024 as reported; $4.46B continuing operations after the 2025 fiber divestiture [12]. The same story as in cloud infrastructure, but with a twist. Raw resources get commoditised. The price per user fell 30% since 2010 while data traffic grew 341x. Per unit, infrastructure gets cheaper forever. But the premium sat with contracted, guaranteed capacity. Tower master lease agreements are the original SLA business, and the tower landlords quadrupled revenue on long-term contracts while the carriers fought a price war underneath them. Commoditized layer means price war. Contracted layer means pricing power. That is the lesson DePIN needs. The rule has a name Clayton Christensen formulated this rule down in 2003. When one layer of a stack commoditizes, the attractive profits migrate to an adjacent layer that still offers unique value or solve hard problem [15]. In infrastructure examples so far, that means moving capital to services on top of infrastructure. Different literature has known versions of it for decades, from Perez’s technology cycles to Carr’s big switch [16][17][18][19]; Perez even predicted the sequencing, with the deployment phase arriving after a financial crash, which maps a little too well onto the 2022 and now 2026 crypto bear. Is the same script running inside crypto today? The stage where infrastructure turns invisible is already observable in crypto at large. Kalshi users trade egg prices and NFL games. PYUSD holders see PayPal. The chain underneath is of interest of tech enthusiasts and blockchain purists. Even Hyperliquid is the adjacent case. Its users might know exactly what chain they are on, but they experience an exchange-grade product first and a chain second, and its $844M of 2025 revenue outearned Ethereum itself, though not Solana [20]. Table 3. Products where crypto is a feature, not the product. | Product | What the user experiences | What runs underneath | Killer metric (2025-2026) | |---|---|---|---| | Kalshi | CFTC-regulated prediction market app | Crypto deposit rails via Zero Hash (USDC, BTC, SOL) | $263.5M revenue in 2025; $22B valuation (May 2026); ~2M MAU [21] | | Hyperliquid | CEX-grade perp trading app, chain invisible in the UX | Its own purpose-built L1 | $2.95T volume and $844M revenue in 2025 [20] | | LayerZero / PYUSD | Tokens that just work across chains; PYUSD holders see PayPal | Omnichain messaging, 90+ chains | PayPal’s PYUSD runs on it; Google Cloud operates a verifier node [22] | | Stablecoins | Dollars that settle in seconds | Public blockchains | $310B+ supply; $10.2T adjusted 12-month transfer volume (Visa/Allium, Aug 2026 pull); GENIUS Act signed July 2025 [23][24] | Every infrastructure sector that matured ran this same script. Crypto is not the exception. It is the newest re-run, and that is a real signal that the industry is maturing and going mainstream. II. The Evidence Can we even see the real revenue in DePIN today? If we want to make a fair comparison, we need to compare apples to apples. With all crypto on-chain transparency, we are still horrendous at reporting companies’ revenues. And especially bad if revenue comes from off-chain sources. The table below is the closest thing that is possible to pull together for the revenue trajectory analysis. DePIN has no audited reporting standard. As projects move to enterprise deals, revenue becomes a black box. Contracts settle off-chain, disclosures are blog posts, and public narratives can be shaped by whoever writes the listicle. During the audit for this article, one widely circulated figure turned out to be a project’s 2022 funding round recycled as its 2026 revenue. So be prepared to take any crypto revenue number today with a grain of salt. The data here is for educational purposes, not for diligence. No audited standard exists in this sector! Table 4. DePIN revenue “leaders”, what they sell, and how much to trust each number (2025-2026). | Project | Reported revenue | Data grade | What counts as revenue | Service and buyer | Traditional alternative | |---|---|---|---|---|---| | Aethir | $127.8M FY2025 [25] | Self-reported, unaudited; net-vs-gross undisclosed | Off-chain enterprise fiat contracts | GPU-as-a-service for AI and gaming, “150+ partners & customers” (company claim) | H100 pricing, see Table 6 [26] | | Storj | ~$13M ARR 2025 [27] | Company statements; Chapter 11 July 2026 [1] | Fiat contracts with Web2 companies | S3-compatible storage, ~$4/TB/mo vs S3 ~$23/TB/mo | AWS S3 | | GEODNET | $8.3M ARR 2026 claim; verified base Q3 2025 $1.23M, +216% YoY [28] | Claim above a verified Messari base | Fiat/USDC subscriptions, 80% funds token buyback | Centimeter-accurate positioning for tractors, drones, and robots; DroneDeploy, Quectel, Propeller | $40/mo vs Trimble-style networks ~$600-1,000+/yr | | io.net | Q1 2025 $5.7M; ARR est. ~$12.5M (third-party annualization) [29] | Messari quarterly + third-party annualization | Marketplace; GMV vs revenue must be flagged (GMV: total flowing through the marketplace, not the company’s take) | GPU marketplace for AI teams | See Table 6 | | Hivemapper / Bee Maps | Six and seven-figure enterprise deals; no verified total [30] | CEO statement; the “$18M annualized” listicle figure failed audit | Map-data contracts | Street-level imagery for VW robotaxi program, Lyft, Mapbox, NBC | Google Street View | | Helium Mobile | Consumer brand acquired by Noble Mobile, June 2026; terms undisclosed, reportedly not profitable [31] | Fortune reporting + Helium’s own announcement; prior ~$23.5M ARR estimate retired | MVNO subscriptions in fiat; Helium team pivots to selling network access to carriers | Mobile plans $15-30/mo vs US average unlimited $65-75/mo | Big-3 carriers | | Akash | FY2025 $3.15M; Q1 2026 lease revenue $253K, declining [32] | Messari on-chain data | On-chain lease payments | Permissionless compute for developers | Cheapest tier, no enterprise SLA | | Livepeer | ~$190K quarterly (Q4 2025, paywalled Messari figure) [33] | Messari on-chain data | On-chain fees | Video transcoding + AI video | AWS MediaConvert | The table could be read in the following way. The winners sell subscriptions and contracts in dollars, the decliners sell raw capacity on-chain. Two points on the companies. Render is absent because it discloses no USD revenue at all. Their “$38M in January” figure floating around listicles was not verifiable anyhow else [34]. And Helium’s acquisition reads both ways. A DePIN consumer brand getting acquired is the Mint Mobile play from Section I completing itself, the buyer is Noble Mobile, Andrew Yang’s carrier startup, which committed to keep running subscribers on the Helium Network while the Helium team moves up the stack to sell network access to other carriers [31]. The brand found a buyer, while the network became the supplier. Yet, reportedly it was never profitable. The measurement problem Here is my favorite piece of statistics from this research. Messari, doing careful work with the data that exists, counts about $72M of on-chain revenue for the entire DePIN sector in 2025. Aethir alone claims $127.8M, because enterprise contracts settle off-chain where dashboards cannot see them [25][35]. The problem here is that nobody can audit the exact number. The industry’s own transparency feature cannot help us to get the numbers right. We built the tool, but forgot to use it for our own sake. Real contracts will decide the winners in the space Why do contracts decide who wins? Because of what an enterprise buyer actually asks for before a dollar moves. The compliance alphabet enterprises live by, including SOC 2, ISO 27001, DPA, SLA, MSA, is the wall every vendor climbs, and DePIN 1.0 could answer none of it. Belief before was that crypto just needed to educate people on how to run things in a blockchain way, and then the adoption would come. Today the answer is the opposite. Projects have to adapt and provide industry-standard contractual obligations, or they cannot exit the Web3 bubble, which is extremely small for infrastructure demand. For scale, enterprise cloud spend most commonly peaks between $100K and $500K per month, and cost efficiency is the top success metric for 81% of organizations [36]. Table 5 covers everything else the buyer asks first. Table 5. The procurement wall: what the buyer asks, what DePIN 1.0 had, what the winners built. | Buyer requirement | DePIN 1.0 answer | The winners today | |---|---|---| | SLA with uptime targets and financial remedies (a service level agreement: uptime guaranteed, credits paid when it fails) | None; best-effort network | AWS-league targets: Storj 99.95% availability, Titan Network 99.95% success rate (author’s company, see disclosure), Aethir advertises enterprise SLAs with 24/7 support; hyperscaler benchmark 99.99% with tiered credits [37][27][38][39] | | SOC 2 Type II / ISO 27001, GDPR DPA, data residency | None | Aethir KYC on GPU providers; DeStor (Seal) SOC 2 and HIPAA [37][40] | | MSA with an accountable legal entity (a master service agreement: a contract with someone you can actually sue) | Anonymous node operators | Service entity signs [31][28] | | 24/7 support with escalation tiers | Discord | Enterprise support desks [37][27] | | Predictable fiat billing | Hold and spend a volatile token | Invoices in stablecoins, on-ramps in dollars or direct dollar settlement [28] | Sources: enterprise procurement criteria and cloud-spend data, Flexera 2026 State of the Cloud [36]; SLA figures from provider pages [38]. We had a token and a dashboard; they needed an MSA and a support line. Disclosure. I co-founded Titan Network. The Titan figures here are mine. Treat them with the same skepticism as any founder’s. I watched the wall win, from the inside In 2021 and 2022, I was a Startup Operator in the Filecoin ecosystem at Protocol Labs. I onboarded more than a hundred teams. The technology worked. The storage proofs were real. And every enterprise conversation hit the same wall. “Do you have an SLA and who would be liable in case of service failures?” Nobody wanted to depend on a network of miners to store their data in a decentralized way, without single org responsibility. Nobody wanted to hold FIL. Nobody wanted to negotiate with anonymous storage providers. They wanted a contract, a responsible entity, and an invoice in dollars. The enterprises that did store data on Filecoin went through service wrappers: Seal Storage, Estuary, FilSwan, companies with compliant structures and MSAs [40]. Some wrappers did not survive: Estuary shut down in 2023, FilSwan exited the storage-provider business in February 2023, Textile’s buckets were discontinued in January 2023, the original Web3.Storage API was deprecated in January 2024 [41]. The survivors, Seal via DeStor with SOC 2 and HIPAA compliance, Akave with S3-compatible enterprise storage, survived because they sold services, not infrastructure [40][42]. The ecosystem keeps trying: DeStor has since been folded into FIL One, an S3-compatible storage service at $4.99/TB per month pitched at the AI era. As of August 2026, its site lists SOC 2 Type II as still in progress, so whether the relaunch captures enterprise adoption remains to be seen [43]. The Storj’s new playbook? In Part 1, I held Storj up as one of two projects in the whole sector with real revenue. Cleanest hybrid model in decentralized storage, roughly $13M in ARR by company statements, sevenfold growth in 2024, real Web2 customers, a published 99.95% availability figure [27]. Sixteen weeks after Part 1 went out, Storj Labs filed for Chapter 11 [1]. The network keeps operating. The company says it expects no service interruptions. The restructuring plan proposes sharing ownership of the reorganized company among management, investors, and token holders, who normally get nothing in a Chapter 11. And the framing in Storj’s own restructuring announcement is the probable pathing for many infrastructure startups in DePIN space: “The business underneath is strong and right-sized. What holds it back are legacy obligations from an earlier chapter” [1]. Helium Mobile is the second datapoint in the same shape. Real service, real subscribers, sold without ever reaching profit [31]. The biggest loss landed on retail, who bought tokens on the promise of product-market fit. Now even the projects that achieved product-market fit and signed the deals are trying to get out from under their token-era obligations. That is what “legacy obligations from an earlier chapter” means in plain language [1]. So why does the hybrid model win anyway? Because the token part solved a problem nothing else could. Token incentives are a real mechanism-design innovation. They crack the cold-start problem of two-sided markets by bootstrapping supply without CapEx, something no Web2 marketplace ever managed. DePIN 1.0’s failure was not the subsidy. It was subsidizing one side forever with no demand loop, and never making the market safe for buyers: no SLA, no accountable counterparty, no quality guarantee. The solution was to repeat the industry standards and bring an accountable service entity standing between the network and the customer. The intermediary as quality guarantor, which crypto tried to eliminate in the first place. GEODNET even closed the loop DePIN 1.0 never had, pushing 80% of revenue into token buybacks [28]. And notice what none of the winners do: lead with decentralization in the sales call. Crypto had a token and a chain. Enterprises needed an SLA and a support line. The winners built the second thing on top of the first. III. The Inflection Point Capital is moving up the stack, and the MOAT question DePIN raised a record ~$1B in private funding in 2025, and the funded deals look different. Bee Maps’ $32M round, co-led by Pantera, was for map-data contracts with VW and Lyft, not hotspot subsidies [30][35]. Sector leaders now trade at 10-25x revenue against 1,000x+ in 2021. Multiples compressing toward SaaS norms means the market has started pricing these as businesses, not tokens, while equity becomes the driver of the fundraising terms [35]. Which forces a question to the industry: What is the defensible advantage of decentralized supply? Table 6. What an H100 GPU-hour costs, by provider type (accessed August 9, 2026; every row publicly linkable, see [26]). | Provider type | Provider | H100 price per GPU-hour | Basis | |---|---|---|---| | Hyperscaler | AWS (p5.48xlarge, derived per GPU) | ~$6.88 | On-demand, 8-GPU instance only | | Hyperscaler | Azure (NC40ads H100 v5) | ~$6.98 | On-demand, single-GPU VM | | Hyperscaler | GCP (a3-highgpu-8g, derived per GPU) | ~$11.06 | On-demand, own pricing page | | Neo-cloud (centralized) | RunPod | $2.99 | H100 SXM, community cloud, own pricing page | | Neo-cloud (centralized) | CoreWeave | HGX H100, 8-GPU instance, own pricing page | | | Neo-cloud (centralized) | Lambda | $3.99 | H100 SXM on-demand, own pricing page | | DePIN | Spheron | $2.98 (live marketplace rate) | Own pricing page, moves with supply | | DePIN | Fluence | $1.50-1.73 (H100 SXM5 VM configs) | Company blog, Nov 2025, not a live pricing page | | DePIN | Aethir | $1.25 | Own enterprise pricing page; no commitment terms disclosed | Sources and per-row links in [26]. Prices are on-demand list rates on the stated date; marketplace rates move. Against hyperscalers the discount is real, roughly 50-75%. Against centralized neo-clouds the ranges overlap, and the lowest DePIN stickers come with undisclosed terms. Previously, DePIN competed on price, and the price was subsidized by token emissions. That era is now over. The token subsidies are dead and the prices have largely equalized. Time to look for more defensible MOATs. My set of durable candidates: supply elasticity, geographic distribution for edge and data-residency workloads, coordination of crowdsourced infrastructure and jurisdiction diversity. Where none of those bind, the winners’ moat is the same as any cloud vendor’s, contracts and customer relationships, with a token-incentivized supply chain underneath. Where are we on the commoditisation curve? I believe DePIN in 2026 sits roughly where cloud sat in 2010-2012, after the raw layer proved itself and before the service explosion on top. The naming problem Right now “DePIN” covers both of these: a project paying people to plug in hotspots with no paying customers, and a project selling enterprise GPU compute under contracts at a self-reported $127.8M a year [25]. That is a practical problem, and it costs everyone. Investors cannot price the difference. Enterprise buyers cannot find the serviceable projects. Builders aim at the wrong target because the celebrated label rewards the wrong behavior. And sector statistics become garbage. The measurement problem above already showed the headline number missing the largest self-reported player entirely [25][35]. I have one prediction. By the end of 2027, the majority of DePIN demand-side revenue will settle off-chain under enterprise contracts, and the on-chain dashboards will be tracking a minority share. What I will not do here is coin the new definition for the winners in the space with real contracts. The sector needs one, with hard inclusion criteria, and that deserves its own article. The stablecoin parallel In every infrastructure cycle, the money ends up with whoever sells contracted service on commoditized supply, and crypto is no different. Stablecoins are crypto’s first killer feature to actually reach the mainstream. More than $310B in supply, $10.2T in adjusted annual transfer volume, a US federal law in the GENIUS Act, Circle on the NYSE, Stripe paying $1.1B for Bridge and co-building a settlement chain, Tempo, with Paradigm [23][24]. The users of these rails increasingly do not know they are using crypto. That is what winning looks like. DePIN can become the second crypto killer feature, but on one condition. Projects keep building services for real customers instead of running for vanity metrics. The time of monetizing hype without traction is gone. The discipline that stablecoin law now demands audited reserves and real reporting is coming for DePIN sooner or later. Long term, only projects that can show orders and revenue on the books will survive; the self-reported revenue announcement will die out as a genre. If the CLARITY Act passes, crypto gets a legal playbook that will transfer to the rest of the industry, and legal framing will lift the projects doing things right today and kill the rest. The sector needs a category with hard inclusion criteria for the projects that crossed the line. The next article will name it, draw the line, and show what it means for capital allocation. Stablecoins proved crypto can disappear into a product people actually use. DePIN is next in line, but only for the projects that sell services, not stories. References [1] Storj Labs Chapter 11 filing, U.S. Bankruptcy Court, Northern District of West Virginia. Storj press release, “Storj Announces Voluntary Financial Restructuring to Resolve Legacy Liabilities and Position the Business for Growth,” GlobeNewswire, July 26, 2026 (quote “The business underneath is strong and right-sized. What holds it back are legacy obligations from an earlier chapter,” attributed in the release to Kaloyan Raev, Director of Software Engineering); CoinDesk, “Cloud Data Firm Storj Files for Chapter 11,” July 27, 2026 (no expected service interruptions; reorganized ownership shared among management, investors, and token holders; STORJ fell 16% to ~$0.06, ~98% below its 2021 peak). https://www.globenewswire.com/news-release/2026/07/26/3333224/0/en/ and https://www.coindesk.com/business/2026/07/27/cloud-data-firm-storj-files-for-chapter-11-extending-a-week-of-crypto-failures-token-slides-16 [2] Amazon.com Inc., Form 10-K filings, FY2015-FY2025, AWS segment reporting, SEC EDGAR. (2017 note: AWS exceeded 100% of Amazon operating income because International segment losses of $3.06B exceeded North America retail profit of $2.84B.) [3] Amazon.com Inc., Q4 2025 earnings release, February 5, 2026, SEC EDGAR. AWS FY2025: $128.7B net sales, $45.6B operating income (57% of Amazon’s $80.0B total; AWS 18% of $716.9B total net sales). CEO guidance: “about $200 billion in capital expenditures across Amazon in 2026.” https://www.sec.gov/Archives/edgar/data/1018724/000101872426000002/amzn-20251231xex991.htm [4] CloudZero, “Inside Netflix’s AWS Strategy.” AWS spend estimated at $1.0-1.3B per year; estimate only, Netflix does not disclose. Caveat: video delivery runs on Netflix’s own Open Connect CDN; AWS runs compute and storage. https://www.cloudzero.com/blog/netflix-aws/ [5] Netflix Q4 2025 earnings report and shareholder letter, January 2026. FY2025 revenue $45.2B. [6] Snowflake Q4 FY2026 earnings release, February 2026. Product revenue $4,472.3M (+29% YoY). [7] Datadog Q4/FY2025 earnings release, February 10, 2026 ($3.43B); Zoom Q4 FY2026 earnings release, February 25, 2026 ($4,868.8M); Airbnb Q4/FY2025 earnings report, February 2026 ($12.2B). [8] Gartner, “Gartner Forecasts Worldwide Public Cloud End-User Spending to Total $723 Billion in 2025,” November 19, 2024 (total 2024: $595.7B; 2025 forecast: $723.4B; SaaS 2024: $250.8B; SaaS 2025 forecast: $299.1B). The $1 trillion in 2027 projection is from Gartner, “Gartner Says Cloud Will Become a Business Necessity by 2028,” November 29, 2023. Diagram 1’s IaaS line uses Gartner’s separate vendor-revenue actuals series, 2016-2024: “Gartner Says Worldwide IaaS Public Cloud Services Market Grew 22.5% in 2024,” August 6, 2025 ($171.8B in 2024; 2023 restated to $140.2B). The two Gartner series use different methodologies (vendor-revenue actuals vs end-user spending); the diagram keeps one vintage per line. https://www.gartner.com/en/newsroom/press-releases/2024-11-19-gartner-forecasts-worldwide-public-cloud-end-user-spending-to-total-723-billion-dollars-in-2025 and https://www.gartner.com/en/newsroom/press-releases/2025-08-06-gartner-says-worldwide-iaas-public-cloud-services-market-grew-22-point-5-percent-in-2024 [9] T-Mobile acquisition of Ka’ena Corporation (parent of Mint Mobile, Ultra Mobile, Plum), up to $1.35B, earnout-contingent, closed May 1, 2024. Ryan Reynolds was a part owner and remains the brand’s pitchman. [10] GSMA Intelligence, “A closer look at MVNOs” (2,100+ MVNOs across ~100 countries). Market size: IMARC $88.7B (2025), Fortune Business Insights $98.7B (2025); definitions vary. https://www.gsmaintelligence.com/blogs/a-closer-look-at-mvnos-what-2100-mvnos-tell-us-about-the-trends-shifts-and-the-outlook [11] American Tower Corporation annual reports and Q4/FY2024 results. Revenue: 2010 $1.99B; 2015 $4.77B; 2020 $8.04B; 2024 $10.13B. [12] Crown Castle Inc. annual reports. Revenue: 2010 $1.88B; 2015 $3.66B; 2020 $5.84B; 2024 $6.57B as reported ($4.46B continuing operations after the 2025 fiber divestiture). [13] SBA Communications Corporation annual reports. Revenue: 2010 $0.63B; 2015 $1.64B; 2020 $2.08B; 2024 $2.68B. [14] CTIA, Annual Wireless Industry Survey, 2025 edition (data through 2024). Service revenue 2010 $159.9B, 2015 $191.9B, 2020 $189.9B, 2024 $224.9B; ARPU $47.53 / $44.65 / $35.31 / $33.36; data traffic grew 341x since 2010. https://api.ctia.org/wp-content/uploads/2025/08/2025-CTIA-Survey-Summary-and-Background.pdf [15] Christensen, C.M. and Raynor, M.E. (2003). The Innovator’s Solution: Creating and Sustaining Successful Growth, ch. 6 (law of conservation of attractive profits). Harvard Business School Press. [16] Perez, C. (2002). Technological Revolutions and Financial Capital: The Dynamics of Bubbles and Golden Ages. Edward Elgar. [17] Mudambi, R. (2008). “Location, Control and Innovation in Knowledge-Intensive Industries.” Journal of Economic Geography, 8(5), 699-725. [18] Isenberg, D.S. (1997). “The Rise of the Stupid Network.” Computer Telephony, August 1997, 16-26. [19] Carr, N.G. (2008). The Big Switch: Rewiring the World, from Edison to Google. W.W. Norton. See also Carr (2003), “IT Doesn’t Matter,” Harvard Business Review, 81(5). [20] Hyperliquid FY2025: $2.95T cumulative volume, ~$844M revenue (ASXN data, December 2025/January 2026). Chain comparison: Solana ~$1.3-1.4B, Hyperliquid ~$844M, Ethereum ~$524M in 2025 (BlockEden, January 10, 2026); revenue concepts differ (exchange fees vs L1 fees/burn). https://blockeden.xyz/blog/2026/01/10/hyperliquid-revenue-dominance-onchain-trading-solana/ [21] Kalshi: 2025 revenue $263.5M and ~2M MAU (Forbes, June 1, 2026); $1B Series F at $22B valuation, May 2026, led by Coatue (Bloomberg, May 7, 2026); crypto deposits via Zero Hash (USDC, BTC, SOL) (The Block, May 23, 2025). https://www.forbes.com/sites/digital-assets/2026/06/01/kalshis-ceo-wants-to-turn-every-question-into-a-market/ [22] PayPal developer blog, “Building the Future of Stablecoin Interoperability: PYUSD and LayerZero” (PYUSD on LayerZero’s OFT standard); Fortune, September 12, 2023 (Google Cloud operates a LayerZero verifier); Messari, “Understanding LayerZero” (90+ chains; paywalled). https://developer.paypal.com/community/blog/pyusd-layerzero/ [23] Visa Onchain Analytics (Allium data), visaonchainanalytics.com: trailing 12-month adjusted stablecoin transfer volume ~$10.2T (August 2026 pull); total supply $310-322B across mid-2026 snapshots. [24] GENIUS Act, Public Law 119-27, signed July 18, 2025. Circle IPO June 5, 2025, NYSE: CRCL. Stripe: Bridge acquisition ~$1.1B, closed February 2025 (CNBC, February 4, 2025); Tempo L1 co-incubated with Paradigm, announced September 2025. [25] Aethir, “Aethir’s 2025 Wrap-Up” (January 2026): FY2025 revenue $127.8M+, “150+ partners & customers,” 439K GPU containers. Unaudited, self-reported; net-vs-gross undisclosed. No 2026 figures published as of August 2026. https://aethir.com/blog-posts/aethirs-2025-wrap-up-decentralized-gpu-cloud-milestones [26] H100 GPU-hour pricing, all accessed August 9, 2026: AWS p5.48xlarge $55.04/hr on-demand = $6.88/GPU-hr (instances.vantage.sh/aws/ec2/p5.48xlarge; AWS’s own page loads prices dynamically); Azure NC40ads H100 v5 $6.98/hr (instances.vantage.sh/azure/vm/nc40adsh100-v5); GCP a3-highgpu-8g $88.49/hr = $11.06/GPU-hr (cloud.google.com/products/compute/pricing/accelerator-optimized); RunPod H100 SXM $2.99 (runpod.io/pricing); CoreWeave HGX H100 $49.24/hr = $6.16/GPU-hr on-demand, spot $19.71/hr = $2.46 (coreweave.com/pricing); Lambda H100 SXM $3.99 (lambda.ai/service/gpu-cloud); Spheron H100 $2.98 live marketplace rate (spheron.network/pricing; rental page lists from $3.38); Fluence H100 SXM5 VMs $1.50-1.73 (fluence.network/blog/nvidia-h100-deep-dive, November 2025); Aethir $1.25 (enterprise.aethir.com/Pricing, no commitment terms disclosed). Directional cross-check only (its snapshot dates differ): CloudZero, “H100 GPU Cost In 2026,” May 20, 2026. [27] Storj: “Storj Increases Revenue Sevenfold In 2024, Enters 2025 In High-Growth Mode,” PRNewswire, December 10, 2024 (sevenfold ARR growth; release contains no dollar figure). ~$13M ARR: Storj company statement, August 2025 (x.com/storj). 99.95% published availability figure (storj.io/benefits/availability-and-durability). [28] Messari, “State of GEODNET Q3 2025”: quarterly revenue $1.23M, +216% YoY (paywalled; corroborated in secondary coverage). $8.3M ARR 2026 is a secondary third-party claim above that verified base. Pricing $40/month (store.geodnet.com, billed in USD) vs Trimble-style RTK correction services ~$600-1,000+/yr. 80% of revenue funds GEOD buyback. Customers: DroneDeploy (Businesswire, February 2025), Quectel, Propeller. [29] Messari, “State of io.net Q1 2025”: Q1 2025 revenue $5.7M (paywalled; corroborated in secondary coverage). ARR ~$12.5M is a third-party annualization. [30] Bee Maps (Hivemapper): $32M raise co-led by Pantera Capital (with LDA, Borderless, Ajna), October 2025; customers include Volkswagen’s robotaxi program, Lyft, Mapbox, NBC (PRNewswire, October 6, 2025; release contains no revenue figures). CEO Ariel Seidman: “six and seven-figure deals” (Blockworks). The “$18M annualized” figure circulating in listicles failed verification and likely echoes Hivemapper’s 2022 $18M funding round (CoinDesk, April 5, 2022). [31] Helium Mobile consumer brand acquired by Noble Mobile (founded 2025 by Andrew Yang), announced June 2, 2026 (blog.heliummobile.com/noble). Noble commits to using the Helium Network; subscribers keep plans and token rewards; Helium team pivots to a carrier-facing network platform. Fortune, June 2, 2026: terms undisclosed; “Helium Mobile is not profitable” (COO Frank Mong); ~600K figure is cumulative sign-ups including churned users. Prior ~$23.5M ARR third-party estimate retired. [32] Messari, “State of Akash” series: FY2025 revenue $3.15M, +128% YoY; Q1 2026 lease revenue $253K, down 45% QoQ (paywalled; corroborated in secondary coverage). Akash’s separately self-reported ~$5M Q1 2026 “compute spend” is a different metric; do not mix. [33] Messari, “State of Livepeer Q4 2025”: demand-side fees ~$190K (paywalled figure, not independently corroborated; order of magnitude consistent with Livepeer’s historical quarterly fees). [34] Messari, “Understanding the Render Network”: USD revenue not disclosed; OTOY publishes burn counts, not dollar revenue. The “$38M January 2026” claim circulating in listicles is unverified. [35] Messari, “State of DePIN 2025,” January 29, 2026: ~$72M FY2025 on-chain revenue (on-chain only; off-chain enterprise contracts not counted), leaders at 10-25x revenue multiples, record ~$1B private funding in 2025 (up from $698M in 2024). The 2021 multiples comparator (1,000x+) is from the report itself; it does not appear in free coverage. [36] Flexera, 2026 State of the Cloud Report: cost efficiency is the top cloud success metric for 81% of organizations; enterprise monthly cloud spend most commonly peaks between $100K and $500K. https://www.flexera.com/blog/finops/flexera-2026-state-of-the-cloud-report-the-convergence-of-cloud-and-value/ [37] Aethir enterprise and documentation pages (accessed August 9, 2026): enterprise.aethir.com (“24/7 support and robust SLAs”); docs.aethir.com KYC/KYB verification via Sumsub (required for Cloud Hosts to claim or withdraw ATH); Cloud Host operational requirements (99% monthly uptime service level objectives with slashing penalties). Note: contractual SLA terms are marketing-page claims, not published contracts. [38] Hyperscaler SLAs, provider pages (accessed August 9, 2026): AWS Compute SLA 99.99% region-level with credits of 10/30/100% below 99.99/99.0/95.0 (aws.amazon.com/compute/sla); Azure Virtual Machines 99.99% with instances across 2+ availability zones; Google Compute Engine 99.99% multi-zone (cloud.google.com/compute/sla). [39] Titan Network: 99.9% uptime SLA published at titannet.io/web-services; the 99.95% success rate figure in Table 5 is the author’s current founder-sourced number. Customers named on the public product page: iQIYI, Volcengine, Tencent; TikTok, Baidu, NetEase, Bilibili appear in company-published materials only (Chainwire, May 2025). Cloudflare is a reseller partnership, not a customer. Founder source (author); see in-body disclosure. [40] Seal Storage / DeStor: SOC 2 and HIPAA compliance (sealstorage.io); clients include UC Berkeley (neutrino physics research data, PRNewswire) and Starling Lab (Filecoin Foundation case study). destor.com now redirects to fil.one; see [43]. [41] Filecoin service-wrapper shutdowns: Estuary discontinued 2023 (GitHub repository notice; original announcement no longer publicly accessible); FilSwan exited the storage-provider business February 7, 2023, migrating clients to Nebula Block (Swan Chain announcement); Textile buckets discontinued January 9, 2023 (deprecation notice); original Web3.Storage API deprecated January 9, 2024 (repository archive notice). [42] Akave: S3-compatible enterprise object storage on Filecoin, no egress fees, $6.65M seed (SiliconANGLE, March 2, 2026; investors include Protocol Labs and Filecoin Foundation). [43] FIL One (fil.one, accessed August 9, 2026): S3-compatible storage on Filecoin, $4.99/TB/month, no egress fees; successor brand to DeStor (destor.com redirects to fil.one); no customers named on site; SOC 2 Type II and ISO 27001 listed as in progress; not mentioned in Filecoin’s “2026 Filecoin Network Strategy” (filecoin.io blog, February 19, 2026).