Tokenmaxxing: Why AI Token Burn Isn’t a Productivity Metric

According to Tech in Asia, a pattern is emerging in AI adoption where organizations prioritize burning through as many tokens as possible daily across prompts, agents, and automated workflows—a practice termed “tokenmaxxing.” Glenn Kaonang, journalist at Tech in Asia, argues that high token usage does not necessarily correlate with high output or improved business outcomes.

The Gym Analogy: Effort vs. Results

Kaonang draws a parallel to gym culture, where spending more hours at the facility does not guarantee better training results. Similarly, token consumption signals effort exerted within an AI system but does not immediately indicate whether outputs have improved. The article notes that most enterprises still measure AI adoption by usage signals such as logins, prompt counts, and seat utilization.

Metrics That Actually Matter

According to the source, the metrics that should be prioritized include:

• How much time AI can save on specific workflows
• Whether AI outputs survive without heavy rework
• Where the recovered time is allocated (described as “where most teams go silent”)

The framework outlined in the piece recommends a practical sequence: efficiency first, then output quality, then capacity reallocation. Kaonang emphasizes that the core question organizations should ask is not “are we using AI?” but rather “are we any better because of it?”

SoBanHang Secures Funding for AI-Native Financial Platform

Vietnamese B2B SaaS startup SoBanHang has raised fresh funding to build an AI-native financial operating system targeting small and medium-sized enterprises (SMEs), according to Tech in Asia. The startup is eyeing regional expansion as early as next year.

Market Thesis and Business Performance

SoBanHang co-founder and CEO Hai Nam Bui has proposed a market thesis that Vietnam’s SMEs may bypass traditional enterprise resource planning systems and move directly to AI-native tooling during a “compressed digitalization” phase. The startup currently serves 800,000 merchants and has achieved over 100% revenue growth.

Application-Layer AI Strategy

The startup represents a case study in application-layer AI strategy, as it is explicitly not building foundation models. Instead, SoBanHang is betting on workflow automation built on top of existing AI models. This approach contrasts with organizations attempting to develop proprietary foundation models.

SkyeChip IPO Oversubscribed 95x Ahead of May 20 Listing

Malaysia-based semiconductor company SkyeChip is set to list on Bursa Malaysia’s main market on May 20, with unprecedented investor demand. According to Tech in Asia, the company’s initial public offering was oversubscribed 95 times—the largest response since Petronas Chemicals’ 2010 listing—representing total demand exceeding 3 billion ringgit (US$773 million).

Financial Performance and Business Focus

SkyeChip reported FY2025 financial results of 119.5 million ringgit (US$30.4 million) in revenue and 35.9 million ringgit (US$9.1 million) in net profit. The Penang-based company designs intellectual property for semiconductors, with a specialty in memory interface IP, including high-bandwidth memory standards that support AI data centers.

Geographic Concentration and Geopolitical Risk

The source notes that over 90% of SkyeChip’s revenue originates from customers in China and Taiwan. The company has acknowledged that tightening U.S. export controls could disrupt its operations, representing a material risk factor despite strong IPO demand.

Trending Tech News

Kuaishou Seeks $2 Billion for Kling AI Video Unit Douyin’s rival Kuaishou is reportedly seeking to raise US$2 billion at a US$20 billion valuation for its Kling AI video generation unit, according to reports cited by Tech in Asia. Kuaishou’s share price rose as much as 10% following the report.

Google Expands Singapore Mosquito Lab with AI and Robotics Google’s R&D lab in Singapore will use AI and robotics to increase yield and improve release precision of mosquitos carrying Wolbachia bacterium, which reduces their reproductive ability and suppresses mosquito populations over time.

OpenAI Launches Deployment Unit with $4 Billion OpenAI has established the OpenAI Deployment Company, which will embed forward-deployed engineers into organizations seeking to build and deploy AI systems. The new unit includes engineers from Tomoro, an AI consulting firm that OpenAI has agreed to acquire.

Indian Microfinance Startup Sindhuja Raises $5 Million Sindhuja operates 366 branches with approximately US$116 million in assets under management. The company has claimed to have disbursed microloans to over 500,000 business owners across 12 Indian states over the past eight years.

TSMC Joins $5 Billion Applied Materials AI Chip Hub Taiwan Semiconductor Manufacturing Company (TSMC) has joined Samsung, SK Hynix, and Micron in a US$5 billion R&D project. TSMC will focus on materials engineering, equipment design, and process integration for data center chips.