Goldman Sachs maintains a Buy rating on Biren with a HK$54 target price, expecting rapid AI chip revenue growth
AI summary card
Goldman Sachs maintains a Buy rating on Biren with a HK$54 target price, expecting rapid AI chip revenue growth
Goldman Sachs is positive on the growth driven by Biren's new-generation AI chip BR20X, domestic cloud capex, and product mix upgrades, maintaining a 12-month target price of HK$54, implying about 87.4% upside from the current price.
- BR20X is expected to start contributing revenue from the second half of 2026 and supports FP8/FP4 computing, higher compute density, memory capacity, bandwidth, and interconnect capabilities.
- The company achieved mass production and scaled shipments of BB106 and BR166 GPGPUs in 2025, and delivered multiple computing clusters, including a 2,048-GPU SuperPod cluster.
- Goldman Sachs raises its 2026-2030E revenue forecasts, mainly reflecting higher AI chip ASPs and a stronger new product line; 2028-2030E net profit forecasts are raised by 12%-26%.
- The valuation uses a 2030E EV/EBITDA method, with a target multiple of 37.3x and a 12.7% cost of equity discounted back to 2027E, resulting in a 12-month target price of HK$54.
Report interpretation
Overview
This report is Goldman Sachs' company research update on Biren(6082.HK). The core view is that Biren's AI chip business is in a strong ramp-up phase, benefiting from the new BR20X product, growth in domestic cloud capex, rising domestic market share, AI application diffusion, and product upgrades toward higher-margin compute boards. Goldman Sachs maintains its Buy rating and HK$54 12-month target price.
Core views
Goldman Sachs believes Biren's future growth will mainly come from five areas: first, new products with higher compute power, better efficiency, and higher ASPs; second, rising capex by China-based cloud vendors; third, higher domestic market share; fourth, growth in AI applications driving demand for AI infrastructure; and fifth, a product mix shift toward higher-margin compute boards. The report expects revenue to achieve a CAGR of about 104% from 2025-2030E, reaching about RMB 15.431 billion in 2028E and about RMB 36.648 billion in 2030E.
Analysis framework
The report updates 2026-2030E revenue, gross margin, expense ratio, EBITDA, and net profit by combining the company's 2H25 results, GPU shipment and ASP trends, product-level revenue forecasts, income statement forecasts, peer EV/EBITDA vs. EBITDA growth relationships, and a discounted valuation framework.
Methodology notes
Multiply 2030E EBITDA by the target EV/EBITDA multiple, then discount back to 2027E to derive the target equity value and target price.
Goldman Sachs uses a 37.3x 2030E EV/EBITDA multiple. The target multiple is derived from the correlation between peers' forward EV/EBITDA and EBITDA growth, and after discounting at a 12.7% cost of equity, it arrives at a 12-month target price of HK$54.
Adjust medium- and long-term profit forecasts based on product mix, ASP, and R&D investment changes.
The report raises its 2026-2030E revenue and gross margin forecasts, driven by higher AI chip ASPs and a larger share of compute boards; it also raises expense ratios, mainly reflecting higher R&D spending required for continued AI chip development and upgrades.
Compare individual stocks relative to the market and industry peers from the perspectives of growth, financial returns, valuation multiples, and combined factors.
This framework uses standardized rankings of sales, EBITDA, and EPS growth, ROE, ROCE, CROCI, and metrics such as P/E, P/B, and EV/EBITDA to form percentile comparisons.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- 6082.HKReport-covered name
- Strengths
- The new-generation AI chip BR20X is expected to improve compute density, memory, and interconnect capabilities; the company already achieved mass production and scaled shipments of GPGPUs in 2025; customers cover AIDC, foundation models, enterprises, and national-level AI computing platforms.
- Weaknesses
- The company is still in a stage of heavy R&D investment, and it is expected to remain loss-making in 2026E and 2027E; the expense ratio remains high, and the profitability inflection point depends on continued revenue scale expansion.
- Comparison
- Valuation multiples reference the relationship between EV/EBITDA and EBITDA growth for peers such as SMIC, JCET, Naura, Kematek, Hygon, and Cambricon.
- Risks
- AI chip demand in China is lower than expected, competition is stronger than expected, and wafer supply constraints affect GPU board shipments.
Key data
- RatingBuyGoldman Sachs maintains a Buy rating.
- 12-month target priceHK$54.00Target price unchanged.
- Current priceHK$28.82Price disclosed in the report.
- Implied upside87.4%Calculated from the target price and current price.
- 2026E revenueRMB 2.585 billionRaised by about 35% versus the previous forecast.
- 2028E revenueRMB 15.431 billionThe report expects revenue to grow 117% year over year.
- 2030E revenueRMB 36.648 billionThe revenue forecast corresponds to rapid growth from 2025-2030E.
- 2025-2030E revenue CAGR104%The chart shows that growth is mainly driven by scale-up in training and inference GPUs.
- 2030E EBITDARMB 3.904 billionUsed in the 37.3x target EV/EBITDA valuation.
- Cost of equity12.7%Valuation discount assumption, including Beta 1.5, risk-free rate 3.0%, and market risk premium 6.5%.
Impact & implications
If Goldman Sachs' forecasts are realized, Biren will gradually move from an early-stage, high-investment, loss-making phase into one of revenue scale-up and profitability improvement. For investors, the key implication is that the company's valuation depends more on medium- to long-term AI computing demand, domestic GPU product iteration, cloud vendor procurement pace, and product mix upgrades than on near-term earnings.
Risks
- AI chip demand in the China market is lower than expected.
- Competition is stronger than expected.
- Wafer supply constraints affect GPU board shipments.
- The pace of BR20X contribution or customer ramp is slower than expected.
- Heavy R&D spending causes losses to narrow more slowly than expected.
What to watch
- The actual revenue contribution from BR20X and customer validation progress in 2H26.
- Capex from China-based cloud vendors and the procurement pace of domestic AI infrastructure.
- Whether GPU shipments can increase from about 20,000 units in 2025 to about 750,000 units in 2030E.
- Whether the higher mix of high-margin products such as compute boards can keep gross margin above 50%.
- Whether net profit turns positive in 2028E as forecast.