Philippines Faces High Stakes in Open-Source AI Shift

By Francis Allan L. Angelo

The accelerating shift toward open-source artificial intelligence (AI) models presents both a golden opportunity and a cautionary tale for Southeast Asia, especially the Philippines.

As Chinese firms like DeepSeek and Qwen release open-weight large language models (LLMs) into the global tech ecosystem, developing countries are scrambling to tap their potential without stumbling into the hidden costs of adoption.

According to Ferdie Saputil, Country Director for the Philippines at global tech consultancy Searce, open-source AI models are transforming the accessibility and localization of AI across Southeast Asia — but success demands strategic groundwork, especially for nations still bridging the digital divide.

“Southeast Asia is poised to greatly benefit from the expanding open-source AI ecosystem,” Saputil told Daily Guardian.

He said China’s AI breakthroughs, particularly open-weight models, give the region an advantage through affordability, adaptability to local languages, and reduced reliance on foreign proprietary technologies.

“The availability of open-weight LLMs from Chinese companies like DeepSeek offers governments and businesses in the region more affordable and accessible AI solutions,” he added.

“This levels the playing field, enabling local startups and smaller businesses to compete with larger enterprises by leveraging the same core AI models, fostering innovation and economic growth.”

Localized AI for Diverse Societies

The rise of open-source AI goes beyond cost savings.

For multi-lingual, culturally diverse regions like Southeast Asia, especially the Philippines, open-source models enable better alignment with local values and language nuances.

Saputil cited projects like SEA-LION, a multilingual LLM initiative focusing on Southeast Asian languages, as evidence of how local customization is critical.

“Open-source models slash the cost and technical headaches of getting into AI, which is huge for developing economies here,” he said.

“They make it super easy to tweak AI for local cultures and languages… AI apps can actually make sense and be ethical for all the diverse folks living here.”

Saputil acknowledged that the Philippines has made “significant strides” in preparing for the AI revolution.

Government policies like the Department of Trade and Industry’s National AI Strategy Roadmap 2.0 and the Department of Science and Technology’s (DOST) original AI Roadmap highlight a commitment to ethics, governance, and cross-sector adoption.

These efforts are supported by major upskilling programs such as AI PINAS and SPARTA, which have trained over 49,000 Filipinos in data science and AI.

“Filipinos are super keen on AI,” Saputil said.

He pointed to a joint study by Microsoft and LinkedIn that revealed 86 percent of knowledge workers in the Philippines already use AI tools—well above the global average.

And the momentum is supported by economic projections.

The Philippine AI market is forecast to reach PHP 57.4 billion (USD 1.025 billion) by 2025, according to recent industry reports.

From 2025 to 2030, the local AI sector is expected to grow at a compound annual growth rate (CAGR) of 27.75 percent.

The growth will be driven by applications in finance, healthcare, customer service, and education, as well as the adoption of both proprietary and open-source AI solutions.

Yet Saputil noted that momentum is not the same as maturity.

“The ‘technology pillar’ in the AI Readiness Index still scores relatively low,” he said, referring to Oxford Insights’ global benchmarking of government preparedness for AI.

That weakness signals broader problems: outdated infrastructure, uneven regional deployment, and a persistent talent gap in AI engineering and deployment.

“AI integration in businesses, particularly among SMEs outside urban centers, remains limited,” Saputil said.

“To fully capitalize… the Philippines must address these critical areas.”

‘Free’ Models, Expensive Surprises

Open-source AI may be “free,” but the real costs begin when companies try to operationalize it.

“Grabbing free AI models from China sounds like a good deal, but the hidden costs will hit you hard if you’re not ready,” Saputil warned.

He said the computing power needed to run open-weight models often requires costly GPU hardware, high energy consumption, and major upgrades to network infrastructure.

“Most companies in Southeast Asia discover too late that they need to invest in expensive GPUs, deal with sky-high energy bills, and upgrade their entire network infrastructure,” he said.

“Those ‘free’ models suddenly don’t feel so free anymore.”

Talent is another issue.

The region lacks professionals with advanced AI deployment experience, and those with the right skill sets are being aggressively recruited by large tech players.

“When you do find them, they’re expensive and constantly being poached by bigger companies,” Saputil added.

“There’s also the compliance headaches.”

Southeast Asian countries vary in how they regulate AI, privacy, and cross-border data flows.

Using Chinese open-source models also introduces geopolitical and legal uncertainties.

“You might build something today that becomes non-compliant tomorrow,” Saputil said.

Even open-source licenses carry legal complexities that can lead to retroactive liabilities.

“Integrating AI into your existing systems is messy and expensive,” he added.

“Models need constant monitoring and retraining as performance drifts.”

Without proper oversight, what starts as a cost-effective project can spiral into a delayed, over-budget experiment.

Proprietary vs. Open-Source

Saputil said Philippine companies need to strategically evaluate whether to use proprietary AI models, open-source ones, or a hybrid of both.

“Deciding between a proprietary or open-source AI model is a big puzzle,” he said.

“Proprietary models like those from OpenAI or Google are like high-performance cars – they’re quick, reliable, and come with roadside assistance. But they’re pricey.”

Open-source models, like Meta’s LLaMA, offer flexibility and major cost savings but require in-house technical muscle and internal support structures.

“For a medium-sized company in the Philippines, a hybrid approach often hits the sweet spot,” he said.

“Use proprietary AI for general tasks like customer service chatbots or marketing support.”

“For sensitive data, like financial records or internal business processes, open-source models give you control.”

He added that companies can locally host open-source models to comply with the Philippine Data Privacy Act while customizing them for local dialects and culture.

“Think of a financial company tuning an open-source model to spot fraud using their specific, confidential transaction history,” Saputil explained.

“It’s all about finding that right balance.”

Building Blocks for Competitive AI

Saputil outlined four “non-negotiable” investments that companies must make to stay relevant in the AI-driven future.

First is talent development—not just hiring AI engineers but reskilling workers across departments.

“Open-source AI makes powerful tools accessible, but you need skilled people to truly leverage them,” he said.

Second is data infrastructure and governance.

“Data fuels AI,” Saputil stressed, adding that scalable storage and strong security protocols are essential.

Third is embedding ethics into AI.

“Mitigating biases and ensuring transparency” is no longer optional, especially as AI moves into critical operations.

Lastly, he urged a shift toward a culture of experimentation.

“The open-source AI world moves incredibly fast,” he said.

“Companies must be willing to adapt quickly, explore new models, and embrace agile development.”

PHL: Narrowing the Gap or Falling Behind?

Saputil believes the Philippines is “not falling further behind,” but cautions that progress remains uneven.

“We’re seeing positive signs: strong government commitment with AI roadmaps, high public interest and usage, and a promising projected market growth,” he said.

“But a notable ‘implementation gap’ exists.”

Many businesses still shy away from full-scale AI adoption, and there’s a long way to go in terms of building advanced infrastructure and cultivating specialized AI talent.

He emphasized that sustained effort—not just policy documents—will be the deciding factor.

“To truly close [the AI maturity gap], we need consistent execution backed by accelerating infrastructure, targeted AI education, robust public-private partnerships, and adaptive regulatory frameworks.”

“The rise of open-source AI is a significant equalizer,” Saputil added.

“But success hinges on focused, sustained effort.”

The Philippines is at a critical juncture: caught between promise and pressure, capability and constraint.

With tens of billions of pesos in projected market value and a digitally curious workforce, the country has the ingredients for AI-driven innovation.

But without serious investments in infrastructure, compliance, and education, the road to AI maturity could remain just out of reach.

As Saputil warned: “The open-source AI revolution isn’t about getting technology for free – it’s about building the supporting ecosystem needed to make that technology deliver real value.”