• Techzi
  • Posts
  • Vietnam's AI Awakening

Vietnam's AI Awakening

Local tech giants pour money into homegrown models and global partnerships

• Vietnam's homegrown AI scene is heating up fast with new local language models.

• Vingroup unveiled PhoGPT while its healthcare AI arm announced Stanford and Microsoft partnerships.

• However, industry insiders say only firms with exceptionally deep pockets can compete in the capital-intensive AI arena.

Summary

ChatGPT and other generative AI models may dominate headlines in the West. Still, Vietnam’s homegrown AI scene is heating up fast with the launch of locally-trained language models like PhoGPT and major investments by leading corporations.

VinAI, the AI research division of Vietnamese conglomerate Vingroup, unveiled PhoGPT earlier this month - a series of natural language processing models tailored specifically for the Vietnamese language.

The launch comes on the heels of VinBrain – Vingroup’s healthcare AI arm – announcing partnerships with Stanford and Microsoft to boost the accuracy of medical imaging AI.

Vingroup’s Up game 

Vingroup is far from the only major Vietnamese firm upping its AI game.

Consumer giant Masan plans to plough up to $105 million into Singaporean credit scoring platform Trusting Social to accelerate the use of AI in retail. Gaming and messaging unicorn VNG is building Vietnamese-language models using open-source codes from US tech firms.

IT services leader FPT forged a strategic partnership with Landing AI to strengthen Vietnam’s AI talent pool.

Industry insiders say AI is a capital-intensive arena that only companies with exceptionally deep pockets can realistically compete in.

VNG CEO Le Hong Minh admits even his firm's AI efforts carry "significant financial risk" despite having access to data from 150 million global users.

OpenAI’s costs doubled to $540 million last year during ChatGPT's development.

Innovative AI Hardware Update

Landing AI founder Andrew Ng notes specialized AI hardware represents the most expensive barrier to entry right now, with computing giants like AMD and Intel dominating the space.

However, he sees opportunities for Vietnamese firms to stand out with developer tools and practical applications rather than foundational algorithms.

Experts believe open-source codes - while not as powerful as tech titans' closed-source models - are key to fueling local innovation while keeping costs contained.

"We can reuse and adopt those open-source models into our smaller data sets for enterprise use. It’s much cheaper that way," says Nguyen Xuan Phong, FPT's chief AI officer.

With both homegrown models and strategic partnerships on the rise, Vietnam's AI ecosystem shows no signs of slowing its rapid advancement. Tech leaders here seem determined to prove AI leadership doesn't begin and end in Silicon Valley.

Reply

or to participate.