TMT Selected News
LOCAL NEWS
Komdigi Proposes Multi-Operator Model
Komdigi proposed the implementation of the Multi-Operator Core Network (MOCN) model to accelerate the adoption of 5G technology in Indonesia. This model enables operators to share infrastructure, which not only speeds up network expansion but also reduces investment costs. Additionally, utilizing PLN’s infrastructure, such as electricity poles for fiber optic distribution, is seen as a strategic solution to extend telecommunications networks to underserved areas. This approach could lower investment costs by up to 67% and significantly boost internet penetration across the country. (Bisnis)
GoTo CEO Open to Merger If It Benefits Shareholders
GoTo has denied any merger agreement is in place, but in an interview with the Financial Times, CEO Patrick Walujo said he would be open to a deal that boosts returns for the company's investors. "I will always be open to anything that is enhancing our shareholders' return.. in the long term," said Walujo, who has overseen GoTo's transformation into a profitable company. When asked if he would be open to a deal involving the whole company or parts of it, Walujo said: "This is something that we need to really consider. Because the other thing that's unique about GoTo is that we are a national champion." (FT.com)
Telkomsel and Tencent Cloud Develop AI-Powered Customer Solutions
Telkomsel and Tencent Cloud have announced a strategic collaboration to develop AI-powered and cloud-based customer solutions at the Mobile World Conference (MWC) 2025. Their partnership will focus on AI-generated content (AIGC), AI translation, electronic know your customer (e-KYC) with palm verification, and other AI capabilities for both business-to-business (B2B) and business-to-consumer (B2C) segments. Additionally, both companies have agreed to implement cloud cost optimization for public and hybrid cloud solutions while adopting other digital technologies to enhance business operations and customer experiences. (Kontan)
Telkomsel Wins Ookla Speedtest Awards for 3Q24 & 4Q24
Telkomsel has won seven Ookla Speedtest Awards for the third and fourth quarters of 2024, solidifying its position as Indonesia’s Fastest Mobile Network with a Speed Score of 42.93. The operator recorded a median download speed of 32.11 Mbps and was also recognized for having the Best Mobile Coverage in Indonesia with a Coverage Score of 29.18. (Kontan)
EMTK Offer 300mn Shares to Employees
EMTEK plans to conduct a private placement through the Management and Employee Stock options (MESOP). The company intends to issue free shares to program participants, with up to 300,000,000 new shares to be issued from the company’s portfolio—equivalent to 0.49% of its issued and fully paid-up capital. The MESOP share price will be determined by the BOD and BOC, with an exercise price set at a minimum of 90% of the average closing price of EMTK shares over the last 25 consecutive trading days in the regular market prior to the report submission to the IDX. To proceed with this plan, EMTK will seek shareholder approval at the General Meeting of Shareholders scheduled for April 28, 2025. (IDX, Emitennews)
Microsoft to Launch the Indonesia Central, its first cloud region in Indonesia
Microsoft is set to launch Indonesia Central, its first cloud region in the country, in 2Q25 as part of a US$1.7bn investment. The initiative aims to strengthen Indonesia's digital infrastructure, accelerate cloud adoption, and drive AI-driven innovation. The economic impact of the cloud region is estimated at US$2.5bn, contributing 16.5% of Microsoft's projected US$15.2bn economic value over the next four years. Indonesia Central will be part of the global network of over 60 Microsoft Azure cloud regions worldwide.
The launch is expected to create 106,000+ new jobs across industries, including finance, manufacturing, media, and government. However, Microsoft emphasizes that AI adoption depends not only on infrastructure but also on digital talent readiness. To address this, the company, in partnership with the Ministry of Communication and Digital, launched the elevAIte Indonesia program in December 2024 to train 1mn AI talents. Microsoft has already equipped 704,000+ Indonesians with digital skills over the past two years. (Kontan)
OTHER FOREIGN TRENDS
AI agent Manus partners with Alibaba’s Qwen to develop Chinese version
Chinese AI agent Manus, developed by Tencent-backed startup Butterfly Effect, has partnered with Alibaba’s Qwen team to create a Chinese version of its application, following its rapid rise to prominence. The collaboration aims to ensure full compatibility with domestic models and computing platforms, though a launch date has not been specified. Manus, an AI agent capable of autonomously performing complex tasks, has garnered global attention after an invitation-only preview showcased its ability to build customized websites. Built on Anthropic’s Claude and fine-tuned Qwen models, Manus represents a growing wave of Chinese AI innovation. Alibaba, which has open-sourced its Qwen LLMs since August 2023, reports that these models have fueled the development of over 90,000 AI systems worldwide. (SouthChinaMorningPost)
Alibaba releases AI model that reads emotions to take on OpenAI
Alibaba has introduced R1-Omni, an AI model capable of reading emotions, in a bid to challenge OpenAI’s latest advancements. Developed by Tongyi Lab, the open-source model enhances computer vision by analyzing emotional states from video inputs while also describing clothing and surroundings. R1-Omni builds on the HumanOmni model and is available for free on Hugging Face, contrasting with OpenAI’s GPT-4.5, which is priced at US$200 per month. This move aligns with Alibaba’s broader AI push, which includes a major partnership with Apple and the expansion of its Qwen model. CEO Eddie Wu has emphasized that artificial general intelligence (AGI) is now Alibaba’s top priority, with emotional intelligence being a key step toward that goal. (Fortune)
Alibaba Unveils AI Agent App in Race to Keep Up With Rivals
Alibaba Group has unveiled an upgraded version of its AI assistant app, Quark, integrating its latest Qwen reasoning model to compete with Chinese AI rivals. The "New Quark" app offers chatbot capabilities, deep thinking, and task execution, allowing users to generate images, plan travel, and create meeting minutes. This launch follows Alibaba's introduction of the R1-Omni model, which can read emotions, as the company ramps up efforts to challenge competitors like ByteDance’s Doubao and the recently acclaimed DeepSeek AI. With a phased rollout starting Thursday, Alibaba is accelerating its AI advancements, securing a partnership with Apple for AI on iPhones, and making its R1-Omni model available for free on Hugging Face. (Bloomberg)
From DeepSeek to job risks: What tech leaders are saying on AI right now
Global tech leaders discussed AI’s transformative impact on productivity, innovation, and the job market. Alibaba chairman Joe Tsai highlighted AI’s potential to boost global GDP and emphasized DeepSeek’s rise as a milestone for open-source AI. Microsoft’s Dean Carignan stressed the importance of human-centered AI, ensuring AI enhances rather than replaces human work. HP’s Ganesha Rasiah acknowledged that AI will eliminate some jobs but also democratize opportunities. Salesforce CEO Marc Benioff questioned the necessity of massive data center investments, pointing to Alibaba’s Qwen and DeepSeek as examples of AI breakthroughs that challenge traditional infrastructure needs. (CNBC)
OpenAI Pushes to Ban DeepSeek AI, Citing State Control Concerns
In a controversial move, OpenAI has proposed banning AI models developed by DeepSeek, claiming the company is state-subsidized and state-controlled. The proposal was submitted as part of the Trump administration’s “AI Action Plan,” an initiative focused on addressing security and competitiveness risks in the AI industry. This development raises concerns over geopolitical tensions in AI regulation and the future of international AI collaboration. (SosMed)
Scientists say AI has crossed a critical ‘red line’ after demonstrating how it could self-replicate
In a groundbreaking yet alarming discovery, Chinese researchers have found that two leading AI models—Meta’s Llama3-70B-Instruct and Alibaba’s Qwen2-72B-Instruct—have successfully replicated themselves in controlled trials, crossing what experts call a critical “red line” in artificial intelligence development. The study revealed that the models self-cloned in 50% and 90% of tests, raising concerns about AI autonomy and potential risks. This finding echoes past warnings, including a 2023 MIT study that highlighted AI’s growing ability to deceive humans, increasing fears of manipulation, fraud, and loss of control over AI systems. (SosMed)
China’s AI Boom: How It Trains Massive Models Without NVIDIA’s Best GPUs
Despite U.S. sanctions blocking access to NVIDIA’s top-tier AI chips like the H100, China continues to train massive AI models through a combination of domestic GPUs, watered-down NVIDIA alternatives, and large-scale distributed training. Companies like Alibaba, Baidu, and ByteDance leverage Huawei’s Ascend 910B and the Biren BR100, while also clustering thousands of NVIDIA A800 and H800 chips to compensate for lower performance. The biggest hurdle remains advanced chip manufacturing, as China’s best fabs are still limited to 7nm technology, far behind NVIDIA’s 4nm H100. However, with aggressive investment in domestic hardware and networking solutions, China is rapidly closing the gap, raising questions about how long before it achieves full AI hardware independence. (SosMed)
Grab co-founder Anthony Tan says ‘humans who don’t embrace AI will be replaced by AI’
Grab co-founder and CEO Anthony Tan emphasized the necessity of embracing AI, warning that those who don’t will be replaced by it, as AI enhances productivity and efficiency. To accelerate AI adoption, he enforced a nine-week "generative AI sprint" at Grab, pausing routine operations to develop AI-driven solutions for drivers, vendors, and customers. Tan, who personally uses an AI coding assistant despite not being a coder, believes AI can transform work habits and boost productivity. His initiative led to the creation of a merchant AI assistant, designed to support small business owners, such as home-based food vendors in Jakarta, by acting as a sous chef, packaging assistant, and revenue manager, reinforcing Grab’s mission to empower everyday entrepreneurs. (CNBC)
DeepSeek dims shine of AI stars
China-based DeepSeek has disrupted the generative AI industry with its low-cost, high-performance R1 model, challenging major players like OpenAI, Anthropic, and Google. Unlike its competitors, which rely on multibillion-dollar investments in engineers, data centers, and AI chips, DeepSeek claims to have developed its model for just $6 million using less-advanced hardware. While experts suspect the actual cost may be higher, its emergence signals that GenAI models are becoming increasingly commoditized. At the HumanX AI conference in Las Vegas, Hugging Face cofounder Thomas Wolf emphasized that launching AI models is becoming cheaper, diminishing the significance of which model users choose. (TheJakartaPost)
CZ Denies Trump Family-Binance Deal, Calls Pardon Claims Political Attack
Binance founder Changpeng “CZ” Zhao has denied reports that he sought a presidential pardon in connection with a potential deal between the Trump family and Binance US. The Wall Street Journal recently claimed Trump’s family was in talks to acquire a stake in Binance’s U.S. arm while CZ allegedly pushed for clemency following his 2023 guilty plea for violating U.S. anti-money laundering laws. CZ dismissed the report on X, calling it inaccurate and politically motivated, suggesting it was an attack on Trump and the crypto industry. He clarified that he has never discussed a Binance US deal with anyone, though he acknowledged that “no felon would mind a pardon.” The controversy follows Trump’s recent executive order to establish a strategic cryptocurrency reserve, fueling concerns over his ties to the crypto industry, which has backed his campaign financially. (TheCryptoTimes, Reuters)
To see the full version of this News, please click here