Bingo Software reported first-half 2026 revenue of 182 million yuan, up 10.04 percent year over year, and net profit attributable to shareholders of 4.4244 million yuan, versus a loss in the prior-year period. The company also disclosed that the Pinyuan AI appliance series generated about 45 million yuan in first-half revenue. These figures, cited by the Securities Times, indicate a return to profitability alongside a contribution from packaged hardware, with details available in the Securities Times report.
The data does not assert a direct causal link between the Pinyuan line and the shift to profit. It presents a snapshot of a company adding a hardware stream to a software-and-cloud business, without breaking out margins by product or assigning profit drivers. That framing allows readers to see the reported numbers as a mix that now includes appliances, while stopping short of implying that any single element alone produced the outcome.
The same report describes how Pinyuan appliances use Jiangyuan’s domestic AI accelerators and notes a financial connection between the firms. Bingo owns 14.53 percent of Jiangyuan Technology after cumulative investment of 400 million yuan. Company materials list a D10 configuration at 64GB memory and 72W power consumption, characterized here as company-reported specifications. These details show how the appliance line incorporates domestic components but do not specify performance targets, efficiency metrics beyond the listed power figure, or workload guidance.
First-Half 2026 Metrics and Revenue Mix, With Caveats
The first half of 2026 combines revenue growth with a move back to profitability. Total revenue reached 182 million yuan, up 10.04 percent year over year, and net profit attributable to shareholders was 4.4244 million yuan compared with a loss a year earlier. Within that period, the Pinyuan AI appliance series generated about 45 million yuan. The report does not quantify margin by product, and it does not state that appliances alone caused the profit shift. Based on the disclosed numbers, the presence of a packaged hardware line is part of the mix, but its specific margin contribution is not identified.
This kind of revenue composition can be useful context. A packaged system introduces a defined bill of materials, a fixed configuration, and a supportable deployment pattern that differs from custom builds. Procurement teams can evaluate a known configuration against site standards, while IT can map deployment requirements to existing facilities. The company-reported D10 specification at 64GB and 72W communicates a baseline envelope for memory and power, even though it stops short of performance claims. As a result, buyers have a reference point for planning, but they still need to assess suitability for their own workloads.
For additional perspective on how vendors position physical AI systems, readers can consult EastFrontier’s look at China’s physical AI hardware push. In that context, Pinyuan sits within a broader move toward preconfigured boxes that pair software stacks with locally sourced accelerators. That framing does not change the specifics of Bingo’s reported results, and it does not ascribe causality to any single factor in the profit outcome.
Inside Pinyuan’s Company-Reported Specs and Sourcing
The sourcing note is straightforward: Pinyuan appliances use Jiangyuan’s domestic AI accelerators. Alongside that, Bingo owns 14.53 percent of Jiangyuan Technology after cumulative investment of 400 million yuan. Taken together, the sourcing choice and the ownership stake align a system line with a chip supplier. The report does not enumerate joint roadmaps or exclusive arrangements, so implications for pricing, allocation, or availability are not specified.
On the configuration side, the D10 entry lists 64GB memory and 72W power consumption in company materials. Those figures provide a descriptive baseline rather than a performance profile. Without throughput data, target model sizes, or latency ranges, the numbers serve mainly to define the appliance’s stated memory capacity and power draw for planning purposes. Prospective buyers would still need to benchmark intended applications to confirm service levels under real workloads.
Revenue disclosure for Pinyuan at about 45 million yuan in the first half offers a measure of traction without implying a unit count or margin. That figure sits alongside 182 million yuan in total revenue and net profit attributable to shareholders of 4.4244 million yuan versus a prior-year loss. The proportional impact of appliances on margin is not derivable from the disclosed data, and no allocation by line item is provided in the cited report.
Why Unified AI Appliances Appeal to Enterprise Buyers
Packaged AI appliances can help a software-and-cloud vendor present a unified offer that combines hardware, models, and deployment into a single deliverable. When a system ships with a specified accelerator, a validated memory configuration, and a supported software stack, the path from pilot to production may be more predictable. IT teams can examine a fixed configuration against security baselines and compliance controls, while operations can plan around known power and space requirements. Procurement gains clarity on what is being purchased and how it will be supported, which simplifies scoping and risk management.
A unified appliance also creates a single surface for software delivery and updates. In enterprise environments, model deployment is often governed by repeatability and documentation. A preconfigured box allows a vendor to certify versions, manage regression risk, and describe performance envelopes under stated conditions. That can reduce integration work across sites, since each installation follows the same configuration and support terms. For a vendor with software and cloud roots, such packaging turns intangible services into an on-site deliverable that can be evaluated in production settings.
Domestic accelerators add another consideration. Some organizations prefer locally sourced components, and the report establishes that Pinyuan uses Jiangyuan’s accelerators while Bingo holds a stake in that supplier. This combination can streamline sourcing conversations by aligning product and component choices under a disclosed relationship. The report does not describe special coordination arrangements, so conclusions about cost or supply are not drawn here.
For readers who track how product and capital decisions intersect in China’s technology sector, a Kechuangban Daily market report offers broader context. It does not modify the specifics of Bingo’s results, and it is noted here for perspective on how appliance strategies can be situated within market narratives.
The first half of 2026 thus provides a limited but clear set of data points. Revenue rose 10.04 percent year over year to 182 million yuan, net profit attributable to shareholders reached 4.4244 million yuan compared with a loss a year earlier, and the Pinyuan AI appliance series contributed about 45 million yuan in revenue.
The report also states that Pinyuan uses Jiangyuan’s domestic AI accelerators, that a D10 configuration is listed at 64GB memory and 72W, and that Bingo owns 14.53 percent of Jiangyuan Technology after cumulative investment of 400 million yuan. The figures do not single out one driver for the return to profit, and they are presented here as a snapshot of a company adding a domestic appliance line alongside existing software and cloud offerings.