The 2025 annual results report of GD-HKGBA HLDGS (01396) ostensibly documents a deep adjustment in its traditional business, but in essence, it serves as a "declaration of transformation" and a "certificate of potential" for the AI era. A thorough analysis of this financial report reveals a robust core of growth, diluted by a "consolidation timing difference," much like the "tip of the iceberg." Behind it lies the "Token Super Factory," which holds sufficient orders to support high-certainty revenue streams for the coming years, now coming clearly into view.
The truth, diluted by the "time window," shows three months of contribution versus a full year of potential. According to the report, for the 2025 fiscal year, GD-HKGBA HLDGS recorded total revenue of approximately RMB 1.0025 billion, a significant decrease compared to RMB 2.6027 billion in the 2024 fiscal year. This figure directly reflects the challenges faced by its traditional infrastructure business, primarily the development and operation of trade and logistics parks, within the industry cycle, as well as a pragmatic strategic restraint. Revenue from this segment plummeted by 85.2% year-on-year to just RMB 385.8 million, recording a negative gross profit. This is the contraction picture presented on the "surface" of the report.
However, the "internal" logic of the report points to a completely different story—a historic turnaround to profitability. The Group's net profit swung from a massive loss of approximately RMB 2.0413 billion in the 2024 fiscal year to an attributable net profit of about RMB 73 million in 2025. The engine for this fundamental shift came entirely from the acquisition of Shenzhen Tiandun Data Technology Co., Ltd., a leading AI computing power service provider, completed in October 2025.
Although Tiandun Data's performance was only consolidated starting from October 23, 2025, it contributed approximately RMB 616.7 million in revenue within just over two months, accounting for 61.5% of the Group's total revenue. It also generated a gross profit of about RMB 177.3 million, with an overall gross profit margin of 28.7%. The core AI computing power service segment achieved an even higher gross margin of 39.6%.
This highlights the first key "overlooked data point": a severe asymmetry in consolidation timing. The weakness of the traditional infrastructure business is reflected over a full 12 months, while the strong growth of the future-oriented AI computing power business is represented by only slightly over two months. This significantly dilutes the boosting effect of the new business on the Group's overall revenue scale, revealing merely the "tip of the iceberg." A simple linear extrapolation of the consolidation period data suggests the annualized revenue potential of the Tiandun Data business.
Furthermore, the Group's cash and cash equivalents surged from approximately RMB 13.1 million at the end of 2024 to about RMB 170.4 million by the end of 2025, an increase of over 1205.1%. This robust improvement in cash flow is the most direct evidence of the AI business's ability to generate cash.
The full picture of the "new engine" lies in Tiandun Data's strength and its annualized potential. To truly assess the future of GD-HKGBA HLDGS, one must look beyond the report's time window and examine the complete standalone performance of its core acquisition, Tiandun Data. Supplementary disclosures in the report reveal Tiandun Data's independent full-year 2025 performance as the key metric for assessing its true scale: revenue of approximately RMB 2.0253 billion, a year-on-year increase of 757.9%, and an attributable net profit of about RMB 207.1 million, achieving strategic profitability.
This means the current consolidated report reflects only about 30% of Tiandun Data's annual energy. A straightforward projection is striking: if Tiandun Data had been fully consolidated from the beginning of 2025, GD-HKGBA HLDGS's total revenue for the year would have easily surpassed the RMB 2.5 billion mark, rather than the reported RMB 1.0025 billion, and the net profit base would have been substantially stronger. This "hypothetical" scenario reveals that the reported revenue decline is essentially an "accounting phenomenon" during the business structure transition period, with the core reality being a complete replacement of the profit engine and a quantum leap in growth momentum.
Therefore, the "low profile" of the current report is, in substance, a case of "delayed gratification" under accounting rules. This is not a negative factor but rather reserves a significant expectation gap and value discovery opportunity for the market.
Looking at future strategy and market valuation, the ambition of becoming the "first AI computing power stock in Hong Kong" is becoming a reality. As we stand at the beginning of 2026, the picture for GD-HKGBA HLDGS is clear. With the ongoing disposal of traditional infrastructure assets and debt—the interest-bearing debt ratio has already improved from 45.3% to 28.2%—the company is poised to advance lightly and focus entirely on the AI computing power track.
More importantly, 2026 will be a critical year for the large-scale realization of Tiandun Data's order book. Stable operational billing orders exceeding RMB 15 billion, supported by ultra-large-scale computing power of over 42,000 PetaFLOPs FP16 (dense)—approximately ten times the total existing computing power scale of Hong Kong—will be substantially converted into operating revenue and profit. Compared to the just over two months of consolidated contribution in 2025, the full-year consolidation of Tiandun Data's performance in 2026 is expected to drive explosive growth in the Group's overall results, making the financial reports significantly more impressive.
The company plans to focus on building an "100,000-card level" ultra-large-scale computing cluster and advance dual-track expansion domestically, such as in green computing centers in Northwest China, and internationally, such as computing power expansion into Southeast Asia. This is expected to foster new growth points atop the existing business.
Recent corporate developments further reinforce this outlook. Beyond being selected for the MSCI Index, the company has planned a name change to "GD-HKGBA Intelligent Computing Technology Co., Ltd.," which will thoroughly refresh its capital market image.
Transformation requires "fuel." The company has secured cumulative intended credit facilities exceeding RMB 30 billion from financial institutions. Particularly significant is the strategic investment of RMB 800 million from a Futian state-owned capital entity, creating a model of "state capital guidance coupled with market-oriented operation." Armed with ample cash, massive credit lines, state-backed support, and international endorsement, GD-HKGBA HLDGS is solidifying the "profit inflection point" shown in its financial reports into the starting point of a long-term upward trajectory.
The 2025 annual report of GD-HKGBA HLDGS is a "treasure script" requiring in-depth excavation. It trades the pragmatic strategic contraction of traditional business for greater growth space for the AI business. The "tip of the iceberg" represented by the short-term consolidated performance conceals a powerful core of intelligent computing technology. The current market valuation may still partially reflect the impact of the old business and the timing mismatch in consolidation. However, with the full-year consolidation of Tiandun Data's performance in 2026, the company's revenue and profit scale are set for a qualitative leap. At that point, the market is likely to reassess the company more definitively through the lens of a pure-play AI computing infrastructure operator—the "Token Super Factory."