CHAINTECH — Interim / Quarterly Report 2019

Dec 16, 2019

Chaintech Technology Corporation

Investment orumF

2019.12.16

Declaration

The information in this document won’t contain financial forecasts .

The information in this document was acquired from TSE MOPS and sources available to the company.

Product Portfolio

==> picture [720 x 270] intentionally omitted <==

----- Start of picture text -----

3 2019/12/16
----- End of picture text -----

Financial Statement

Statement of Comprehensive Income(QoQ)

==> picture [720 x 270] intentionally omitted <==

----- Start of picture text -----

5 2019/12/16
----- End of picture text -----

Statement of Comprehensive Income2019(YoY)

==> picture [720 x 270] intentionally omitted <==

----- Start of picture text -----

6 2019/12/16
----- End of picture text -----

2019Q3 Consolidated Condensed Balance Sheets-1

==> picture [720 x 270] intentionally omitted <==

----- Start of picture text -----

7 2019/12/16
----- End of picture text -----

2018 Q3 Consolidated Condensed Balance Sheets-2

==> picture [720 x 270] intentionally omitted <==

----- Start of picture text -----

8 2019/12/16
----- End of picture text -----

2018 Q3 Consolidated Condensed Balance Sheets-3

==> picture [720 x 270] intentionally omitted <==

----- Start of picture text -----

9 2019/12/16
----- End of picture text -----

Financial Ratio

		2019.9.30	2018.12.31
	Debt Ratio	31.10%	13.92%
	Current Ratio	276.60%	619.59%
	Quick Ratio	209.72%	584.66%
	AR Turnover	5.21	4.08
	Days sales in AR	70.05days	89.46days
	Inventory Turnover	10.24	35.28
	Average days in sales	35.64days	10.34days
	Cash Flow Ratio	2.40%	159.19%

Consolidated Condensed Cash Flow Statements

==> picture [720 x 270] intentionally omitted <==

----- Start of picture text -----

11 2019/12/16
----- End of picture text -----

Consolidated Condensed Cash Flow Statements

==> picture [720 x 270] intentionally omitted <==

----- Start of picture text -----

12 2019/12/16
----- End of picture text -----

==> picture [720 x 270] intentionally omitted <==

2019/12/16

13

2020 Prospect

• Expand the AI server market continuously

• Application Technology Dep Invest in software application technology development and AI software and hardware system integration continuously

Invest in AI Focus on AI SERVER SI & Application Technology

==> picture [720 x 270] intentionally omitted <==

----- Start of picture text -----

15 2019/12/16
----- End of picture text -----

XIDIAN  UNIVERSITY -AI computing  acceleand technology-a  computing  acceleration  platform  project

Project background ：

XIDIAN UNIVERSITY is a national key university with information and electronic subjects as the main subject and coordinated development of engineering, science, management and culture. It is directly under the Ministry of education. It is one of the key universities of national advantageous discipline innovation platform project and "211 Project", one of the national innovation and entrepreneurship demonstration bases, the first 35 demonstration software colleges, the first nine demonstration microelectronics colleges and the first nine approved ones One of the first batch of demonstration projects for the construction of first-class network security college Facing the major national strategic development and international cutting-edge development needs, the College of artificial intelligence of XIAN UNIVERSITY deeply implements the spirit of the report of the 19th National Congress and the development plan of new generation artificial intelligence, practices the construction of "Internet + belt and road" and innovation oriented country, and strives to

build a training base for high-end talents in the field of artificial intelligence, a research and Development Center for innovation achievements and a high-level team cultivation platform. Solution plan：

Based on the construction of intelligent education computing acceleration platform of Artificial Intelligence College of XIAN

UNIVERSITY , and combined with the practical experience of similar customers before, siton Heli proposes a complete cluster solution

of management node + computing node (several NVIDIA dgx-1) + storage + Infiniband network.

XIDIAN  UNIVERSITY -AI computing  acceleand technology-a  computing  acceleration  platform  project

NVIDIA DGX-1：

• -The computing power can reach petaflops

• -8 ×Tesla V100 nvlink interconnection technology

• -NVIDIA CUDA core quantity 40960

• -Number of NVIDIA sensor cores 5120

Through the GPU cluster solution, siton Heli has successfully built a computing acceleration
Pingsheng for   XIAN   UNIVERSITY

==> picture [720 x 270] intentionally omitted <==

PING AN TECHNOLOGY - DGX-1

Project background:

Founded in 2008, PING AN TECHNOLOGY is a wholly-owned subsidiary of Ping An Group Guokang. Your company has branches in Shenzhen, Beijing, Shanghai, Chengdu and Nanjing. Ping An technology develops and operates key pinghao and services of PING AN TECHNOLOGY negative voice, which supports the efficient development of the group's insurance, banking, investment and Internet business. At the same time, it is also a technology incubator of Ping An group. It has strong cloud, artificial intelligence and big materials Research and development capabilities

In the financial industry, time is money, and millisecond determines profitWith lightning insight and decisive execution, profits can be made. The focus is on making informed decisions faster than competitors, which will ultimately be achieved by using big data, and obtaining analysis results faster is a big advantage. With the statistical calculation gradually approaching the limit, the financial industry is focusing on GPU, and banks and investment companies are gradually switching to NVIDIA GPU NVIDIA DGX-1 (the first system developed for in-depth and a-accelerated analysis in the world) to meet the real-time analysis needs, including fraud analysis, risk management and algorithmic transactions.

Core requirements:

1. In portfolio risk management, a trader must extract information and input it into a special system to perform advanced analysis and modeling. When calculating risk, a large number of calculations are needed, which usually takes all night to divide into "lines", and it is difficult to make adjustments in time with market changes. 2. Transaction execution involves figuring out how to find out the best price of a stock in the limit order book. Whether the trading time is hundreds of milliseconds or a minute from now, in the trading gap of more and more specific stocks, we all want to know when the stock price is the highest, now or a few seconds later.

Solution:

In response to the needs of users, siton Heli puts forward the NVIDIA DGX-1 super Thunderhead solution. With the help of advanced GPU and in-depth event processing technology, traders can perform arduous tasks such as data exploration, model development scoring and model consumption on the computing platform

PING AN TECHNOLOGY – DGX-1

Secondly, in the process of trading execution, by mastering the quantitative data supported by the deep learning framework, we can understand the future trend of this threshold stock through the millions of trading data in the past. After the period training of massive data, we can make real-time reasoning on these data, and judge whether we should trade in a few hundred milliseconds, a second or a minute. This kind of intelligence cap improves the potential of algorithmic trading

NVIDIA DGX-1：

• The computing power can reach 1 PetaFLOPS

• -8 × Tesla V100 nvlink interconnection technology

• -NVIDIA a CUDA core quantity 40960

• -Number of NVIDIA sensor cores 5120

Nvlink is a high bandwidth and low energy consumption interconnection technology. With this technology, the interconnection between nvd|agpu and the same generation GPU or other devices in the node can be realized. The total bandwidth of each GPU can reach 300gb / 5, which is about 9 times of the current interconnection of pcle gen3x16 and the hybrid cube GPU topology of DGX-1 orange. The highest bandwidth can be achieved between a group of 8 Tesla v100 of data exchange.

==> picture [720 x 270] intentionally omitted <==

==> picture [720 x 270] intentionally omitted <==

----- Start of picture text -----

20 2019/12/16
----- End of picture text -----

==> picture [720 x 270] intentionally omitted <==

==> picture [720 x 270] intentionally omitted <==

----- Start of picture text -----

21 2019/12/16
----- End of picture text -----

Product Information

==> picture [720 x 270] intentionally omitted <==

----- Start of picture text -----

22 2019/12/16
----- End of picture text -----

SITONHOLY Software product

==> picture [720 x 270] intentionally omitted <==

----- Start of picture text -----

23 2019/12/16
----- End of picture text -----

SITONHOLY Hardware products

==> picture [720 x 270] intentionally omitted <==

----- Start of picture text -----

24 2019/12/16
----- End of picture text -----

Colorful GeForce RTX 2080 SUPER Gaming ES

	Chipset	GPU Manufacturing Process CUDA cores	GeForce RTX 2080 SUPER 12nm 3072
	Core Clocks	Base/Boost Clock (Turbo Model) Base/Boost Clock Buswidth	1650MHz/1815MHz N/A 256Bit
	Memory Clocks	Memory Clock Memory Config Memory Interface Memory Bandwidth	15.5Gbps 8GB GDDR6 496GB/S
	Display and	Video Output Maximum Digital Resolution*	3DP+1HDMI 7680x4320
	Connectors	PCI Express NVLink/SLI	3.0 Yes
	Thermal and Power Specs	Maximum GPU Temperature Graphics Card Power Power Phase Power connector	89 C 250W 8+2 8+8Pin
	Cooling	Type Intelligent star-stop fans Heatpipe size & Q’ty	290+180mm N 3*φ8
	3D API	Fan Power Connector DirectX	DirectX 12 OpenGL 4/5
	Others	OpenGL Supported NV Technologies Form Factor	Real-Time Ray Tracing, Ansel, GPU Boost Over Dual slot 31012653mm

Colorful GeForce RTX 2070 SUPER Gaming ES

	Chipset	GPU	GeForce RTX 2070 SUPER
		ManufacturingProcess	12nm
		CUDA cores	2560
	Core Clocks	Base/Boost Clock	1605/1770MHz
		(Turbo Model) Base/Boost Clock	N/A
		Buswidth	256Bit
	Memory Specs	MemoryClock	14Gbps
		MemoryConfig	8GB
		MemoryInterface	GDDR6
		MemoryBandwidth	448GB/S
	Display and Ct	Video Output	3DP 1.4 +1HDMI 2.0
		Maximum Digital Resolution*	7680x4320@60Hz
	onnecor s	PCI Express	3.0
		NVLink/SLI	YES
	Thermal and Power Specs	Maximum GPU Temperature	88 C
		Graphics Card Power	215W(NV)
		Power Phase	8+2
		Power connector	8+8Pin
	Cooling	Type	3 * Fan(180+290mm)
		Intelligent star-stopfans	YES
		Heatpipe size & Q’ty	3*φ8
		Fan Power Connector	8-pin, PWM
	3D API	DirectX	DirectX12
		OpenGL	OpenGL 4/5
	Others	Supported NV Technologies	Real-Time Tracing, Ansel ,GPU Boost
		Form Factor	Dual Slot
		Dimensions	31012653mm
		Back Plate	Yes

iGame GeForce GTX 1660 Ti Ultra 6G

	Chipset	GPU	GeForce GTX 1660 Ti
		ManufacturingProcess	12nm
		CUDA cores	1536
	Core Clocks	Base/Boost Clock	1500MHz/1770MHz
		(Turbo Model) Base/Boost Clock	1500MHz/1845MHz
		Buswidth	192Bit
	Memory Specs	MemoryClock	12Gbps
		MemoryConfig	6GB
		MemoryInterface	GDDR6
		MemoryBandwidth	288GB/S
	Display and	Video Output	DP+HDMI+DVI
		Maximum Digital	7680x4320@60Hz
	Connector s	l i Reso ut on*
		PCI Express	3.0
		SLI	NO
	Thermal and Power Specs	Maximum GPU Temperature	89℃(NV)
		Graphics Card Power	120W
		Power connector	8 PIN
	Cooling	Type	3*90mm Fan
		Intelligent star-stopfans	NO
		Heatpipe size & Q’ty	Φ6*2
		Fan Power Connector	4 PIN PWM
	3D API	DirectX	DirectX12
		OpenGL	OpenGL 4/5
	Others	Supported NV Technologies	Ansel ,GPU Boost
		Form Factor	Dual Slot
		Dimensions	31012642mm
		Back Plate	YES

iGame GeForce RTX 2080 Ti Advanced

	Chipset	GPU Manufacturing Process CUDA cores	GeForce RTX 2080 Ti 12nm 4352
	Core Clocks	Base/Boost Clock One-key OC	1350/1635MHz N/A
	Memory Specs	Memory Clock Memory Config Memory Interface Memory Bandwidth	14Gbps 11GB GDDR6 616GB/S
	Display and Connecto rs	Video Output Maximum Digital Resolution PCI E*	3DP 1.4 1HDMI 2.0 1USB Type-c 7680x4320@60Hz 30*
		xpress SLI	. Yes
	Thermal and Power Specs	Maximum GPU Temperature Graphics Card Power Power supply	89℃(NV) 250W (NV) 8+8pin
		Power Phase	13+3
	Cooling	Type Intelligent start-stop fans Heatpipe size & Q’ty Fan Power Connector	3 * Fan (90mm) YES 5*φ8 8pin, PWM
	3D API	DirectX OpenGL	DirectX12 OpenGL 4/5
	Others	Supported NV Technologies Form Factor Dimensions Back Plate	Real-Time Tracing, Ansel ,GPU Boost Over Dual Slot 30411852mm YES

THANK YOU