The useful question is not whether Suiji is "real"
The cheap way to misunderstand Shanghai Suiji Networks Tech. Co. Ltd. is to ask whether it is a real network operator. The public record answers yes and no at the same time. Yes, it has a Chinese company identity under the name Shanghai Suiji Networks Tech. Co. Ltd., with the Chinese legal name Shanghai Suiji Network Technology Co., Ltd. Yes, APNIC records AS146813 under the SUIJI name, a Lingang New Area address in the China (Shanghai) Pilot Free Trade Zone, administrative and technical contacts using the suijinetworks.com mail domain, and an abuse contact that was still being maintained in late 2025. Yes, Suiji's own site presents a menu of enterprise internet products: IP Transit, data-centre hosting, virtual machines, DIA dedicated internet access, bare metal and overseas native IP or ISP bandwidth. The site carries an ICP filing number, a public-security filing number and a support address under the same domain. Those are not the signals of a domain parked for speculation.
No, the same public record does not show a scaled carrier network. RIPEstat currently sees AS146813 announcing one IPv4 prefix, 23.236.111.0/24, first visible in July 2025, with one observed neighbour and no visible IPv6 route. Hurricane Electric's BGP view shows the same basic shape: one originated IPv4 /24, RPKI-valid, one observed v4 peer, no v6 prefixes, and Zenlayer Inc. as the visible peer. ARIN's RDAP record for the /24 points to Zenlayer as the address-space registrant, not Suiji. PeeringDB has no public network record for AS146813. The official sales pages repeatedly send the buyer to a Zenlayer console for ordering. In other words, Suiji is visible as a permitted Chinese-facing enterprise-network vendor, but its visible autonomous-system footprint looks like a very small edge of someone else's much larger platform.
That contradiction is the point. Suiji's operating value is not best measured by counting prefixes. It is best measured by the credibility spread between the permissions it can show and the infrastructure it appears to rent. In China's enterprise internet market, that spread can be commercially valuable. A buyer that needs an invoiceable Chinese supplier, ICP and filing support, local-language operations, a licensed path for IDC, CDN, ISP, IP-VPN or domestic data-transmission work, and some ability to assemble China-plus-overseas connectivity may not care whether the ultimate IP Transit fabric belongs to Suiji, Zenlayer, a state carrier or another upstream. It cares whether the supplier can legally stand in the contract, route traffic without embarrassing the buyer, handle abuse and filing problems, and keep the service working under Chinese regulatory and carrier constraints. Suiji's public record is therefore not a miniature version of China Telecom. It is a due-diligence case in how small companies sell trust at the edge of a heavily licensed market.
A young company with old-fashioned licensing economics
Suiji's own company-introduction page says it was founded in October 2020 in the Lingang New Area of the China (Shanghai) Pilot Free Trade Zone. It describes the business in plain terms: providing enterprise users with internet access services and internet business acceleration to improve digital experience. A third-party recruiting/company profile gives the same incorporation month, a registered capital of RMB 10 million, the unified social credit code 91310000MA1H3A8N7A, legal representative Chen Shuo, the Lingang address at No. 1881 Zhengbo Road, and a business scope that includes ordinary technology services plus licensed telecom activities. The profile also classifies the company as a private enterprise with 1-49 people. Treat that headcount as a recruiting-platform signal rather than audited fact, but it fits the rest of the public record: Suiji looks like a specialist enterprise-network company, not a national infrastructure owner.
The licence signals are unusually important. Suiji's official qualifications page lists six categories: internet information service, domestic internet virtual private network, fixed domestic data transmission, internet access service, internet data centre/cloud computing and content delivery network. A public repost of the MIIT value-added telecom licence issuance list identifies Shanghai Suiji Network Technology Co., Ltd. with licence number A2.B1-20214126 in the 2021 batch, covering fixed domestic data transmission, IDC, CDN, domestic internet virtual private network and internet access service. A later 2022 list shows a legal-representative-change item for the same company and licence number. These are not equivalent to owning international gateways or national backbone facilities, but they materially change the risk profile. A small Chinese company without such licences can be dismissed quickly when it sells network access, CDN or IDC-like services. A small company with them deserves a narrower question: what exactly can it provide itself, what does it lawfully resell, and where does the dependency sit?
That question matters because China's network-access market prices permission separately from engineering. Since MIIT's 2017 clean-up of the internet access service market, IDC, ISP and CDN providers have been under explicit pressure not to use unlicensed network infrastructure or resell access resources through informal chains. Cross-border business is even more sensitive: the notice says unapproved entities may not establish or lease dedicated lines, including VPN channels, to conduct cross-border business, and that international leased lines rented from basic telecom operators are for internal office use rather than connecting domestic and overseas data centres for telecom operations. Separately, China's international communications gateway regime limits who can build and operate international communication gateways. That means a company such as Suiji can create value by helping customers navigate permitted services, but it cannot turn a website promise of "global" reach into an independent right to carry regulated cross-border telecom traffic.
The practical economics follow. Licences are fixed-cost assets: they require corporate filings, compliance systems, personnel, security procedures and regulator-facing discipline. They do not automatically produce bandwidth. Bandwidth comes from carriers, data centres, cloud platforms, IP transit providers and local access partners. A small licensed company can make money when it buys those inputs better than a customer can, packages them under a local contract, handles filing and support friction, and accepts the operational blame when something fails. The gross margin is an intermediation margin. It is not the margin of owning fibre across China.
The website is a sales surface and a warning label
Suiji's website is valuable evidence precisely because it is imperfect. It lists product categories that make commercial sense for a Shanghai enterprise-access intermediary: bare metal cloud, IP Transit, data-centre colocation, virtual machines, DIA dedicated internet access, and overseas native IP addresses with ISP bandwidth. The home page also lists resource locations in Wuhan, Hangzhou, Chengdu, Jinan, Xi'an, Beijing, Shanghai, Guangzhou, Shenzhen and Zhengzhou. The company page calls Suiji a leading edge-cloud service provider. The contact page says Suiji provides 7-by-24 Chinese-English technical support. The colocation page says it can offer customized data-centre hosting and managed host solutions, active monitoring, procurement, transport, delivery, customs assistance, installation and system configuration. Those are exactly the work items that a light-asset network-services company would sell around larger suppliers.
The warning label is in the details. Some site copy is generic, duplicated or awkwardly templated. The home page repeats the company introduction three times with numbered headings. Product pages contain broken-looking placeholders such as missing superlatives and repeated "bare metal versus dedicated server" links. Several ordering links point to console.zenlayer.com rather than a Suiji checkout. The IP Transit page claims a "130+ Tbps" network, nationwide capacity above 500 Gbps, more than 20 connected data centres in more than 10 cities, local ISP direct connections, a dedicated backbone and a 99.95 percent SLA. Those numbers may be describing the service fabric available through partners rather than Suiji's own AS146813 footprint. The bare-metal and virtual-machine pages also resemble translated or adapted platform copy, and both direct the buyer toward the same Zenlayer ordering environment.
For a procurement officer, the website therefore helps and hurts at once. It helps because the company publishes a concrete service catalogue, filing numbers, support email, office address and sales posture. A totally opaque network name would not give even that much. It hurts because the sales surface does not clearly separate owned resources from partner resources. If the buyer is purchasing "Suiji IP Transit," is Suiji providing BGP transit under AS146813, reselling Zenlayer IP Transit, bundling a carrier line with a Zenlayer cloud node, or acting as local contract and support intermediary? Those are different risk allocations. The public pages do not answer them. They reveal a product wrapper, not the underlying bill of materials.
The addresses add another clue. The APNIC record and corporate registry signal point to the Lingang New Area address. The official website footer gives an office address in Baoshan District, at Hulan West Road. The privacy policy gives a Jing'an District address on Jiangchang Third Road. Multiple addresses are not suspicious by themselves; young Chinese technology companies often use registration, office and service-contact addresses differently. But they reinforce the central due-diligence problem. Suiji's value is distributed across legal identity, licence, office presence, website, partner platform and routing records. No single public surface should be treated as the whole company.
AS146813 proves operational literacy, not scale
The cleanest technical evidence is AS146813. APNIC's RDAP record identifies the autonomous system as SUIJI, country CN, active, registered on 23 June 2022, with a description naming Shanghai Suiji Networks Tech. Co. Ltd. at the Lingang address. Administrative and technical contacts are listed under suijinetworks.com email addresses, and the abuse contact is gnoc@suijinetworks.com. APNIC also shows a portable IPv6 allocation, 2403:6980::/32, under the same SUIJI name and address. In formal registry terms, Suiji has enough network identity to be more than a white-label website. It has a registered ASN, an IPv6 allocation and maintainable abuse contact data.
The live route table is much smaller than the registry identity. RIPEstat's routing-status API shows one visible IPv4 prefix, 23.236.111.0/24, first seen on 11 July 2025 and still visible in early July 2026, with 322 of 324 RIPE RIS IPv4 peers seeing it. The same API shows no visible IPv6 route and one observed neighbour. The announced-prefixes view for the two weeks ending 3 July 2026 likewise shows only 23.236.111.0/24. IPinfo counts 256 IPv4 addresses, zero IPv6 addresses routed, seven hosted domains, one upstream, one peer and zero downstreams. Hurricane Electric shows the /24 as RPKI-valid, one v4 peer and no v6 originated prefixes. PeeringDB's API returns no public network record for the ASN.
That is a very narrow technical footprint. It does not support a story that AS146813 is the fabric behind a 130+ Tbps enterprise network. It does support a story that Suiji understands enough BGP, registry operations and route-origin hygiene to maintain a small public routing identity. That distinction is commercially important. For a buyer, a valid ASN with a single routed /24 can be enough for test infrastructure, branded service endpoints, control-plane experiments, customer-facing performance probes, small hosted services or an abuse-isolated slice. It is not enough to prove owned nationwide data-centre capacity, independent international reach or a deep transit business.
The prefix provenance makes the dependency explicit. ARIN's RDAP record for 23.236.111.0/24 names the network as ZL-LAX-ZENWORKS-0001 and the registrant as Zenlayer. The parent block sits under Zenlayer's ARIN allocation. BGP views show Zenlayer Inc. as the visible peer and upstream for AS146813. IPinfo's traceroute data reaches a Suiji address through AS21859 before entering AS146813. This looks like Suiji originating a small Zenlayer-assigned block behind Zenlayer transit, not Suiji bringing its own IPv4 inventory to the table. Because IPv4 addresses are scarce and expensive, that matters. A company that owns a clean /24 has a small balance-sheet asset and more routing autonomy. A company that announces a partner's /24 has useful service control but less asset depth and more supplier risk.
The APNIC IPv6 allocation is also revealing. A /32 is more than enough address space for a serious IPv6-enabled service platform. Yet public route collectors show no visible IPv6 prefix originated by AS146813. That may be a temporary operational choice, a filtering issue, a dormant allocation or simply a service that does not depend on public IPv6 routing. Whatever the explanation, it means the company's registry potential exceeds its live routed surface. A buyer should not read the existence of 2403:6980::/32 as proof that Suiji is serving IPv6 customers at scale unless route visibility, reverse DNS, service endpoints and customer references corroborate it.
Zenlayer is not just a supplier; it is the operating shadow
Zenlayer's presence runs through almost every important technical clue. The Suiji product pages use Zenlayer ordering links. AS146813's visible IPv4 route is tied to Zenlayer address space and upstream transit. AbuseIPDB hostname listings around a Digital Pulse Technology address show Suiji subdomains among many Zenlayer and CDN-related operational hostnames, including test and client API names. Zenlayer's own public pages describe a far larger platform: hundreds of edge nodes, global cloud on-ramps, a multi-hundred-Tbps network, IP Transit, bare metal, China optimization and private cross-border connectivity. PeeringDB lists AS21859 with global scope, heavy outbound traffic and 10-20 Tbps traffic levels. In plain commercial terms, Zenlayer has the global infrastructure story that Suiji's pages appear to sell into.
That does not make Suiji meaningless. In reseller, channel and managed-service markets, the local wrapper can be the product. A multinational buyer can buy from Zenlayer directly if it knows exactly what it wants, has cross-border legal review, can manage Chinese filings and does not need a smaller local integrator. A Chinese enterprise or media customer may prefer a domestic supplier that speaks the local procurement language, carries relevant licence categories, offers ICP filing support, can blend domestic access with overseas nodes, and absorbs front-line operational friction. The economic value of Suiji is therefore closer to a broker-dealer or systems integrator than to a backbone carrier: it converts global infrastructure into a local enterprise service.
The dependency still changes the risk. If Zenlayer's pricing, account access, route policy or China-facing partnership terms change, Suiji's public product bundle may change with it. If a customer has an abuse issue, geolocation problem, VPN tag, blocked domain or route leak, the resolution path may require Zenlayer action even when the customer contract is with Suiji. If a customer buys "overseas native IP addresses and ISP bandwidth," the actual reputation of those addresses will depend on the local supplier and address history, not on Suiji's Chinese licence alone. If a buyer needs provable diversity, it must ask whether Suiji can deliver alternative suppliers beyond Zenlayer, and whether those alternatives are pre-contracted or improvised after failure.
The positive side is that dependence on a large supplier can reduce execution risk. A 1-49 person company does not need to own hundreds of PoPs to deliver a usable enterprise service if it has disciplined access to a larger network. The question is not whether Suiji uses Zenlayer. The question is whether Suiji is transparent about what Zenlayer provides, where Suiji adds operational control, and which obligations Suiji can satisfy without waiting on the upstream.
The licence stack is real value, but it does not erase the boundary
Suiji's licence categories explain why the company can be more valuable than its AS suggests. Internet access service, IDC/cloud, CDN, domestic IP-VPN and fixed domestic data transmission are the regulated boxes that enterprise-network procurement teams care about in China. A supplier that can show these categories can pass screens that a generic hosting reseller cannot. It can speak to ICP filing, public-security filing, network access compliance, domestic VPN service boundaries and data-centre service procurement. The official site also offers "filing support" in the service menu. These are not decorative items in China. They are part of the cost of doing business.
But the licence categories also define the limit. Domestic data transmission and domestic IP-VPN are not the same as unrestricted international private-line operation. IDC and CDN licences do not make a company a basic telecom carrier. An ICP filing on suijinetworks.com proves that the website is registered for mainland service; it does not prove that every overseas IP or cross-border acceleration product is operated under a dedicated international telecom approval. MIIT's clean-up notice and the international-gateway rules create the perimeter: value-added providers can operate licensed services within scope, but cross-border telecom operation remains tightly controlled and often depends on the basic carriers or approved channels.
This is why Suiji's "overseas native IP and ISP bandwidth" page is the most interesting product page. The page says the company works with multiple local suppliers to help customers achieve global internet coverage, and describes use cases such as overseas market expansion, load balancing, network testing and privacy protection. That is a commercially plausible product: Chinese firms need overseas endpoints, testing locations, app-delivery paths and local-looking connectivity. It is also the product most likely to attract reputational and regulatory scrutiny. "Native IP" can mean a legitimate local ISP address allocation for performance testing or content localization; it can also sit uncomfortably close to proxy, VPN or traffic-masking markets if not governed carefully. IPinfo's ASN page tags at least one Suiji-addressed IP as VPN and notes BitTorrent observation in the past 30 days. A Reddit Quad9 discussion listed a Suiji test-performance subdomain among alleged DNS false positives. Those are not proof of misconduct. They are market-chatter signals that IP reputation management is part of the job.
That reputation work is costly. The customer buying overseas IP coverage does not want an address that triggers banking risk controls, streaming anti-proxy systems, DNS filtering or spam heuristics. The supplier has to know where the address came from, which upstream can fix WHOIS and geolocation records, how abuse desks respond, and whether the block has been burned by previous users. Suiji's public AS uses only 256 visible IPv4 addresses, so the company's larger overseas IP story almost certainly depends on partners. That makes partner selection, abuse response and transparency the heart of the product.
Customer proof is thin, but one procurement trace matters
Most small network companies have a public customer problem. They may have real customers, but the customers do not want their suppliers publicized, and the operator's website fills the gap with generic case studies. Suiji is no different. Its "customer success case" page is mostly a title and image. The home page mentions a case in which a customer deployed across multiple public clouds and needed SaaS acceleration, but the visible snippet does not name the customer or give measurable outcomes. The product pages list target customers - multinational enterprises, cloud service providers, network operators, games, video, banking, insurance, government and large enterprises - but target sectors are not customer proof.
The strongest public customer signal is a government-procurement notice reported as sourced from China Government Procurement Network and reposted by Sohu and aggregated by Jobui. It says China News Service awarded Suiji a domestic and overseas data-centre and internet services project, project number ZYZB-2025-1053, with a winning amount of RMB 3.845 million, service period from 15 February 2026 to 14 February 2027, and a comprehensive score of 88.64. The stated scope was to support an integrated service system, domestic optimization and overseas data localization management. That is exactly the type of contract Suiji's public positioning would seek: a media organization needing domestic and overseas infrastructure under one supplier relationship.
Because the currently accessible copies are reposts and aggregations, this signal should not be overstated. Before a buyer treats it as a reference, the original government-procurement page and the tender requirements should be pulled directly. Still, the signal is too specific to ignore. A RMB 3.845 million one-year contract would be meaningful for a small specialist supplier. It suggests that Suiji is not merely a website translating Zenlayer copy; at least one public procurement trail places it as a named supplier for a real institutional buyer. That does not prove service quality. It does show a possible revenue mechanism: Suiji packages domestic and overseas data-centre/internet services for customers whose procurement process values a Chinese licensed counterparty.
The same signal also frames the risk. Media, SaaS, gaming and cross-border enterprise customers are sensitive to outage, DNS filtering, routing asymmetry, content-delivery geography, data-localization commitments and public-sector scrutiny. If Suiji can manage those issues, the company can earn a premium over simple pass-through resale. If it cannot, the customer will eventually buy direct from a larger cloud, a state carrier or a global edge provider.
Unit economics: licence, supplier spread and support labour
Suiji's unit economics are likely driven by three layers. The first is licence and corporate overhead. A company with value-added telecom licences, ICP filings, public-security filings, support contacts, privacy-policy obligations and regulated service categories has a minimum compliance cost. APNIC's published fee schedule shows that a new member with a /32 IPv6 allocation and minimal IPv4 resources would face annual fees in the low thousands of Australian dollars, plus sign-up costs for new members. Registry fees are not the main burden. The larger costs are people, systems, audits, filing support, security controls, legal renewals and regulator-facing maintenance.
The second layer is supplier cost. If Suiji uses Zenlayer for bare metal, IP Transit, cloud networking or edge nodes, the gross margin depends on discount, volume, account terms and how much support Suiji performs itself. If Suiji buys domestic access or DIA from carriers, the margin depends on last-mile availability, installation fee, commit period, bandwidth burst model and whether the line is bundled with managed service. If Suiji arranges overseas IP addresses from multiple local suppliers, the margin depends on the scarcity and reputation of those addresses. Cheap addresses with poor reputation are not a bargain if they break customer workflows.
The third layer is support labour. The contact page's 7-by-24 Chinese-English support claim is commercially ambitious for a small company. True around-the-clock bilingual support requires staffing, rota discipline, escalation processes, monitoring, vendor access and incident documentation. A small supplier can fake the claim during quiet periods and fail during real incidents. It can also deliver it by combining its own first-line coverage with upstream support from larger providers. The buyer needs to know which model applies. Support is where the reseller either earns the margin or exposes the customer to double escalation: first to Suiji, then to Zenlayer or a carrier.
The revenue side is easier to sketch than to verify. Bare metal and virtual machines can be sold as monthly recurring infrastructure. DIA and IP Transit can be sold by committed bandwidth, port, location and term. Colocation can be sold by rack unit, cabinet, power, cross-connect and remote-hands services. Overseas IP and ISP bandwidth can be sold as a managed access or testing package. CDN and acceleration services can be sold by traffic, region, commit and SLA. In each case, Suiji's defensible margin is not the cheapest bandwidth. It is the combination of licence comfort, Chinese-language contract/support, supplier aggregation and faster deployment than a customer could achieve alone.
Switching cost is the hidden economic asset. Once Suiji handles an ICP filing, a domestic DIA line, an overseas endpoint, a CDN acceleration path, an abuse desk and a bilingual escalation chain, the customer is not merely changing bandwidth provider when it leaves. It has to retest latency, replace whitelisted addresses, re-document security controls, adjust routing, renegotiate procurement and explain to its risk team why the new path is still acceptable. That stickiness can make a thin resale margin attractive if incidents remain rare. It also makes underperformance more dangerous. A customer that has embedded Suiji into delivery, testing and support workflows may tolerate small price increases, but it will not tolerate ambiguous responsibility during a cross-border outage. The customer will ask one simple question: can the supplier restore service faster than the customer can rebuild the stack elsewhere? For a small company, that is the real economic test.
The price pressure is severe. State carriers can undercut or out-authorize smaller players on domestic circuits. Alibaba Cloud, Tencent Cloud, Huawei Cloud and UCloud can bundle compute, CDN, security and compliance at platform scale. Zenlayer can sell directly to sophisticated customers. Global CDNs and SASE providers can sell performance and security to multinationals. Small Chinese integrators can copy the same website language. Suiji's customer has to believe that Suiji knows the messy middle - the place where a cloud node, a Chinese licence, an ICP filing, a domestic line, an overseas test address and a bilingual support ticket all have to meet.
What the market chatter says, and what it does not say
Public chatter around Suiji is thin. There are no obvious waves of customer complaints, no widely discussed outage record, no community forum built around the brand and no large public employee footprint. That silence has two readings. The positive reading is that the company serves enterprise customers through private channels and does not need a loud retail presence. The negative reading is that the company has little independent brand pull and relies on partner infrastructure plus direct sales. Both can be true.
The security and DNS signals are noisier. IPinfo labels AS146813 as hosting and tags at least one assigned IP with VPN and BitTorrent observations. A Reddit user discussing Quad9 false positives listed test-perfops.idevops.suijinetworks.com among several domains they believed had been falsely blocked. AbuseIPDB's WHOIS page for a Digital Pulse Technology IP lists multiple Suiji and Zenlayer operational hostnames among a long reverse-DNS or co-hosted-domain set. None of this proves customer harm. It does show the kind of reputation environment Suiji is operating in. Performance-test domains, acceleration platforms, overseas IP pools, CDN hostnames and proxy-like use cases often sit near threat-intelligence noise even when the service is legitimate. The supplier that sells those services has to be good at explaining and cleaning up that noise.
The website's language around overseas IP use cases also requires discipline. "Market expansion" and "network testing" are clean enterprise needs. "Privacy protection" and identity-hiding language can attract customers whose behaviour damages address reputation. Suiji's long-term value depends on keeping the former without becoming a magnet for the latter. In IP markets, one abusive customer can consume more operational time than ten good ones, and a small routed footprint leaves little room to quarantine bad reputation.
The competitive map: carriers own permission, clouds own scale, intermediaries own friction
Suiji sits between three stronger classes of competitor. The first is the state carrier group. China Telecom, China Mobile and China Unicom own the deepest domestic network authority and the international-gateway position. They are slower and more bureaucratic, but their permission stack is stronger. Any customer whose core need is regulated international connectivity or mission-critical national access will treat them as the reference point, even if a smaller integrator handles part of the implementation.
The second group is the cloud and CDN platforms. Alibaba Cloud, Tencent Cloud, Huawei Cloud, UCloud, Baidu AI Cloud and global providers solve many enterprise needs at scale: compute, CDN, security, DNS, monitoring, identity, logs and compliance support. They are not always flexible for mixed-vendor, cross-border, nonstandard route or overseas-native-IP work. But they set the procurement baseline. If Suiji cannot explain why its bundle is faster, cheaper, more flexible or more China-specific than a cloud-native bundle, it loses.
The third group is the global edge-network providers, especially Zenlayer itself. Zenlayer has the public infrastructure story, global PoPs, direct cloud connectivity and broad emerging-market pitch. Suiji's pages appear to borrow from that gravity. The competitive question is whether Suiji adds something Zenlayer does not want to do directly for every Chinese customer: local licence posture, sales access, filing help, domestic data-centre bundling, bilingual support and small-project customization. If the answer is yes, Suiji can survive as a channel and managed-service specialist. If the answer is no, the company is exposed to disintermediation by its own upstream.
This is the normal shape of an intermediary market. The supplier that owns friction earns money until the platform product removes that friction. Suiji's defence is that China's telecom, filing, cross-border and procurement frictions are durable. They are not just bad UX. They are regulatory and operational facts. That gives a small licensed company room to exist even when it does not own much routed space.
The practical diligence test is therefore contractual, not rhetorical. A buyer should ask for the service map behind each quoted line item: which ASN originates the route, whose address block is used, which carrier installs the domestic circuit, who holds the cross-connect, who can open the upstream trouble ticket, who approves abuse remediation, and who pays service credits after an outage. If Suiji can answer those questions cleanly, the small routed footprint is a manageable limitation. If the answers collapse into vague partner language, the customer is buying opacity wrapped in licence comfort. That is the difference between a useful integrator and a risk-transfer illusion.
Public evidence that anchors the judgement
The company identity and self-description are anchored by Suiji's own pages at https://suijinetworks.com/ and https://suijinetworks.com/?list_22%2F=, which describe the company as founded in October 2020 in Shanghai's Lingang New Area and focused on enterprise internet access and acceleration. The product catalogue is visible across the official pages for IP Transit, DIA, colocation, virtual machines, bare metal and overseas native IP/ISP bandwidth, including https://suijinetworks.com/?list_10%2F=, https://suijinetworks.com/?list_13%2F=, https://www.suijinetworks.com/?list_11%2F=, https://www.suijinetworks.com/?list_12%2F=, https://www.suijinetworks.com/?list_9%2F= and https://suijinetworks.com/?list_14%2F=. The qualifications page at https://www.suijinetworks.com/?list_23%2F= lists the licence categories, while the site footer shows the ICP and public-security filing numbers.
The network identity is anchored by APNIC RDAP for AS146813 at https://rdap.apnic.net/autnum/146813 and APNIC RDAP for 2403:6980::/32 at https://rdap.apnic.net/ip/2403:6980::. RIPEstat's API views at https://stat.ripe.net/data/routing-status/data.json?resource=AS146813 and https://stat.ripe.net/data/announced-prefixes/data.json?resource=AS146813 show the live public route footprint. Hurricane Electric's AS page at https://bgp.he.net/AS146813 and IPIP's page at https://whois.ipip.net/AS146813 corroborate the one-prefix, one-peer picture. ARIN RDAP for https://rdap.arin.net/registry/ip/23.236.111.0 ties the visible /24 to Zenlayer. IPinfo at https://ipinfo.io/AS146813 adds hosted-domain, upstream, peer, VPN and BitTorrent reputation context. PeeringDB's API at https://www.peeringdb.com/api/net?asn=146813 returned no public record during research.
The regulatory frame is anchored by the MIIT 2017 clean-up notice republished by the Cyberspace Administration at https://www.cac.gov.cn/2017-01/23/c_1120366809.htm and by the international gateway rules published by Shanghai's economic and information technology authority at https://www.sheitc.sh.gov.cn/bmgzjxgwj/20020620/0020-656687.html. The licence-number signal is supported by public reposts of MIIT issuance lists, including https://m.tlkuaiban.com/zixunzxxq.html?id=962 for the 2021 issuance batch and https://www.9mayi.com/news/18963.html for the 2022 legal-representative-change item. The company-registry and employment-scale signal is supported by Liepin's company profile at https://www.liepin.com/company/gs67956737/.
The supplier-dependency evidence is anchored by Zenlayer's own pages at https://www.zenlayer.com/, https://www.zenlayer.com/ip-transit/, https://www.zenlayer.com/bare-metal/ and https://www.zenlayer.com/china-connect/, plus PeeringDB's AS21859 page at https://www.peeringdb.com/net/1234. The customer/procurement signal is supported by the China News Service project repost at https://m.sohu.com/a/969589965_122434053 and the Jobui procurement aggregator page that attributes the notice to China Government Procurement Network. Reputation-chatter context is supported by the Reddit Quad9 thread at https://www.reddit.com/r/Quad9/comments/1otbgyc/fps_on_9999/ and AbuseIPDB's hostname listing at https://www.abuseipdb.com/whois/103.148.58.203.
What would change the judgement
The bullish case would improve materially if Suiji published or could provide under NDA a clearer network map separating owned infrastructure from Zenlayer, state-carrier and other partner facilities. Public PeeringDB records, exchange ports, multiple visible upstreams, customer-visible route diversity, an announced IPv6 plan for 2403:6980::/32, and direct evidence of non-Zenlayer address resources would all show more control. So would a verified original government-procurement page for the China News Service award, audited revenue or tax data, named customer references, support metrics, and licence records pulled directly from MIIT's official system. Evidence that Suiji holds or lawfully accesses specific cross-border telecom approvals would also matter for the overseas IP and enterprise connectivity claims.
The bearish case would strengthen if AS146813 disappears, if the /24 is withdrawn or reassigned, if Zenlayer ordering links break without replacement, if Suiji's ICP or public-security filings lapse, if licence categories cannot be verified in official systems, if customer references cannot be produced, or if IP reputation deteriorates around Suiji-branded domains. A pattern of proxy-abuse complaints, unresolved DNS blocking, or unsupported "native IP" claims would be especially damaging because it would attack the company's most valuable asset: trust.
The final reading
Suiji should be priced as a licensed enterprise-network intermediary with a small public routing identity, not as an independent infrastructure carrier. That is not a dismissal. In China's enterprise internet market, the scarce asset is often not fibre. It is the ability to combine permission, local support, filing knowledge, upstream procurement, cross-border caution and service accountability into something a customer can buy. Suiji has enough public evidence to belong on a procurement longlist for certain narrow work: domestic access, IDC/CDN-adjacent services, managed acceleration, Zenlayer-backed edge services, overseas testing or small enterprise connectivity bundles. It does not have enough public evidence to be treated as a standalone global network owner.
The most important number is therefore not one /24, 130+ Tbps or RMB 3.845 million. It is the spread between them. One /24 shows the visible network Suiji controls directly enough to announce. The 130+ Tbps-style language shows the scale of the partner platform its sales surface wants to invoke. The procurement number shows what a customer might pay for a local company to make domestic and overseas data-centre/internet service usable under a single Chinese contract. Suiji's business, if it works, lives in that spread. The risk is that buyers mistake the spread for ownership. The opportunity is that many buyers do not need ownership; they need a credible operator who knows where the boundaries are.