Welcome back to Asia Tech Review, your curated digest to keep up to date with tech news across Asia.

Sea showed its ambition with news that it is planning to spinout and list its fast-growing fintech business, and also launch a cloud computing business. These two strategies show why it is Southeast Asia’s top tech company, but they worried investors and sent the share price tumbling.

To keep up with our issues, follow us on LinkedIn or WhatsApp. Please do share Asia Tech Review with people you know who want to get smart about Asia tech.

Sea is making major moves after reports emerged that it is launching a cloud computing business and is planning a public listing for its fintech arm, Monee, according to reports from Reuters and Bloomberg.

Less than one day after Southeast Asia notched its largest fintech IPO ever, as Philippines-based Mynt raised nearly $1 billion from a domestic listing, Sea could be the next. We’ve written about Monee’s impressive growth. The business covers payments and lending, and its nearly 60% annual revenue growth has made it double the size of Sea’s trusty Garena games business when it comes to sales.

Monee is clearly Sea’s next growth lever, as we wrote in August:

Shopee continues to account for over 60% of all revenue, and it is clearly the big focus, but Monee has the scope to grow strongly.

“Only a fraction of the users across our ecosystem are using Monee’s financial products today, and credit penetration remains low across our markets,” Sea chairman and CEO Forrest Li said in a statement.

But rather than capture the value fully inside Sea, the company believes Monee needs its own space to breathe. Monee will get “resources and the room to go further and faster than it can today,” Sea founder Forrest ⁠Li wrote in a memo to employees. This is in contrast to Grab’s recent move to acquire fintech lender Atome for $1.5 billion and integrate its business into its company. [Disclosure: Atome is an Ellerton & Co. client]

Investors don’t agree, for now at least. Sea’s share price fell by over 9% on the news.

That may also be linked to the second revelation, that Sea is creating its own cloud computing business. The company is copying the Amazon-AWS and Alibaba-Alibaba Cloud playbook. That has huge potential, but it will come at cost.

AWS was famously an expensive venture for Amazon when it launched in 2006, while Alibaba joined three years later at considerable cost. But these two business units have been strong growth engines for both companies:

• AWS now has a $169 billion annualised revenue run rate

• Alibaba Cloud is the centrepiece of the Chinese firm’s AI and compute services division, which saw revenue rise 45% year-on-year to reach 48.44 billion CNY ($7.15 billion) in the June 2026 quarter

Sea has never been shy to invest in long-term strategies, and this is clearly the ultimate one. Despite being decades late to the cloud, it is betting that there’s room to make money as AI expands demand for affordable and quality compute services.

That’s a major gamble, and certainly the most ambitious plan a Southeast Asia-based tech company has ever made. That’s not exactly what I meant when I called Sea “Southeast Asia’s true internet giant,” but it certainly deserves it based on this newfound ambition.

Meanwhile in Indonesia, GoTo released a major LLM for the local market. Sure, this isn’t cloud computing or a spinout IPO, but it is meaningful for developing a powerful and local model that’s specifically for Bahasa Indonesia and regional languages in Southeast Asia.

“GoTo-AI is a collection of Large Language Models (LLMs) which have been pretrained and instruct-tuned for Indonesian culture and languages. GoTo-AI-Gemma-3-27B-IT leverages the Gemma-3 architecture alongside specialized local models,” reads its description on popular AI database Hugging Face.

GoTo worked with universities, government agencies and media organisations to build the dataset, the company told Business Times. It claims it achieved its highest overall score across five Indonesian languages, and it outperformed its previous model despite being less than half the size.

We’ve written before that Southeast Asia is worryingly ill-equipped to deal with the digital sovereignty crisis that is emerging from using international products, including LLMs, and computing services that are not native to the region. Sea’s cloud computing business and this GoTo AI model are promising responses, but much more needs to be done.

Foxconn will pay 115 billion yen ($727 million) for a 50% stake in Mitsubishi Electric’s automotive-components business, pairing Mitsubishi’s expertise in electric powertrains and in-car systems with Foxconn’s electric-vehicle ambitions [Nikkei Asia]

Indian healthcare startup Even Healthcare raised $50 million in a Khosla Ventures-led Series B that values it at $300 million, three times its valuation 18 months ago, as it expands from insurance into running hospitals [TechCrunch]

Chinese autonomous-delivery company Neolix raised $400 million in a Boyu Capital-led Series E, taking its total funding past $1.4 billion [Bloomberg]

South Korean neocloud provider Elice raised 201 billion won ($151 million) in the year’s biggest Kosdaq IPO, it fell as much as 32% in its trading debut before rebounding the next day [Bloomberg]

Mynt, the operator of Philippine payments app GCash, rose as much as 2.1% in its market debut after selling 8.03 billion shares and raising 53 billion pesos ($840 million) in the base offering [Reuters]

Shares in Chinese memory-chip maker ChangXin Memory Technologies fell 7.8% in a global chip sell-off despite its forecast of up to $19.8 billion in net income for the first nine months of 2026 [SCMP]

Chinese AI startup MiniMax updated its M3.1 Flash model, claiming nearly three times faster inference and 70% lower costs for agent, coding and cybersecurity work [MiniMax]

Researchers found that AI agents powered by DeepSeek and Claude Code were used to breach more than 100 companies, with one campaign targeting more than 1,200 ecommerce sites for an estimated $5,000 in tokens [Reuters]

DeepSeek is said to be working with Huawei on AI chips due next year, with Huawei engineers embedded at the startup to shape designs around how its models process and move data [The Information]

In more official news, DeepSeek open-sourced DeepInfer, the inference framework behind its V4.1-Flash model, to help developers reduce latency, computing costs and energy use [GitHub]

Indian chemical companies are looking to enter semiconductor supply chains as Tata targets chip production by mid-2028, but many still lack the ultra-high-purity technology and customer access chipmakers require [Nikkei Asia]

South Korea plans to invest 1,350 trillion won ($1 trillion) in its southwest, including four semiconductor plants costing 800 trillion won ($593 billion) in Gwangju, with production targeted for 2030 [Bloomberg]

Alibaba’s Damo Academy has shipped more than 6 billion chips based on its XuanTie RISC-V architecture as China looks for alternatives to Arm and Intel under US export controls [SCMP]

Alibaba, Tencent and Huawei are investing in data centres in Shaoguan, Guangdong, where cheaper electricity, ample land and a cooler climate are drawing computing capacity inland [FT]

ByteDance, Alibaba and Tencent have reportedly ordered more than 3 million Nvidia RTX PRO 5500 graphics cards worth over $35 billion as they seek export-compliant chips [Reuters]

ARTEX, the Chinese AI tool linked to breaches at South Korean banks last month, was also used in cyberattacks on about 20 Japanese companies, with eight reporting data leaks [Bloomberg]

The US sanctioned Chinese venture firm MiraclePlus and three affiliates over allegations that they financed a startup accused of stealing Google AI trade secrets [Bloomberg]

The US Department of Homeland Security ordered federal civilian agencies to remove DeepSeek technology from their IT systems and barred future purchases of the Chinese company’s products [WSJ]

Japan will certify the security of data centres and cloud services as it tries to encourage their use in sensitive sectors while curbing risks linked to China and Russia [Nikkei Asia]

Four Z.ai founders worked on a research project with an institute the UK government says has close links to China’s main spy agency, according to records seen by Bloomberg [Bloomberg]

Japan is months away from exhausting supplies of yttrium, a metal used in automotive, defence and semiconductor products, and has asked Washington to press Beijing to ease export controls [FT]

China launched the state-owned National Data Group in Xiongan to develop data resources and promote sharing across industries and regions [Reuters]

Apple has made more money than Samsung in India for the first time after the US firm’s operating revenue rose 31% to ₹1.04 lakh crore ($12.5 billion) in FY26, while Samsung’s mobile-phone revenue in India fell 1.4% to ₹81,472 crore ($9.8 billion) [Economic Times]

Flipkart is testing a food-delivery service with more than 8,000 employees ahead of a potential challenge to Zomato and Swiggy, using its Ekart logistics arm and the government-backed ONDC network [TechCrunch]

The US-China AI race is becoming a contest over computing infrastructure, technological innovation and military power, with the outcome likely to shape who writes the rules for the AI era [Foreign Affairs]