Gobi Partners, a leading Asian venture capital firm, has announced an undisclosed strategic investment in ArmourZero Holdings, a cloud-based cybersecurity platform based in Malaysia.

The investment, made through the Gobi Dana Impak Ventures (GDIV) fund, is backed by Khazanah Nasional and aligns with Khazanah’s Dana Impak mandate.

According to Tho Kit Hoong, CEO of ArmourZero, the investment will accelerate innovation.

Co-founded in 2022 by cybersecurity expert Hoong and tech innovator Chong Wai Lun, ArmourZero aims to address the cybersecurity needs of software developers and small and medium enterprises (SMEs).

Also Read: Embracing unity: A celebration of diversity and inclusion at ArmourZero

The platform tackles issues such as high cyber threat incidence, inadequate threat containment, prohibitive costs, and limited access to integrated security systems.

Key solutions:

ShieldOne: A unified threat monitoring, management, and response system. It integrates endpoint security, email protection, and patch management into a single platform. ShieldOne provides real-time threat protection, 24×7 Managed Detection and Response (MDR), and partners with industry leaders such as CrowdStrike and Checkpoint.

Managed Detection and Response (MDR): A core feature of ShieldOne, it offers real-time threat detection, proactive incident management, and rapid response through a dedicated team of cybersecurity analysts.

ScoutTwo: An AI-powered application security system that secures web and mobile applications from development to deployment. It provides instant vulnerability detection, risk prioritisation, and AI-powered remediation recommendations.

ArmourZero aims to bridge this gap by helping SMEs mitigate risks, reduce costs, and strengthen their digital defences. In Malaysia, over 28,000 cyberattacks were recorded in 2022, with incidents between 2017 and 2021 resulting in RM2.23 billion (US$490 million) in financial losses.

ArmourZero has subsidiaries in Malaysia, Singapore, and Indonesia. The company’s core activities are based in Malaysia.

The cybersecurity market in Southeast Asia is projected to grow from US$35 billion in 2023 to US$84 billion by 2028. SMEs, which make up 99 per cent of Malaysian businesses, are particularly vulnerable due to limited resources and awareness.

The post Gobi invests in ArmourZero to bolster SME cybersecurity defences appeared first on e27.

As artificial intelligence (AI) continues to redefine industries and workplaces, concerns about job displacement persist. However, rather than rendering human workers obsolete, AI is expected to complement human capabilities—emphasising the need for uniquely human skills. A recent Workday study reveals that the tech will be a driving force behind a global skills revolution, making human-centric skills more valuable than ever.

The report elaborates how AI is playing an increasingly pivotal role in skill development by alleviating workers from routine processes and enabling them to focus on higher-order tasks. By automating repetitive activities, the tech allows individuals to channel their creativity and problem-solving abilities into more strategic and imaginative work.

Additionally, AI-driven skills assessment and gap analysis improve productivity by ensuring that employees receive targeted learning opportunities, making professional development more efficient and data-driven.

Beyond productivity gains, AI fosters adaptability and resilience—critical skills in an era of rapid technological change. By offering interactive learning experiences and processing vast amounts of data to provide insights and decision support, AI enhances employee engagement and professional growth. This empowerment extends beyond individuals, as the tech facilitates the exchange of information, making skills data actionable at scale and enabling businesses and governments to expand workforce opportunities.

AI excels at processing vast amounts of data, automating repetitive tasks, and enhancing efficiency. Yet, it lacks the nuanced understanding, empathy, and ethical reasoning that define human interactions. This is why several human-centric skills continue to remain relevant even in the age of AI, according to the report.

Also Read: Atome taps BlackRock, InnoVen for expanded US$80M credit facility

As organisations integrate the tech into their operations, the ability to navigate complex social dynamics, make ethical decisions, and lead with emotional intelligence will become essential. Employers are increasingly prioritising soft skills such as adaptability, collaboration, and critical thinking. These competencies enable individuals to work effectively with AI-driven tools, fostering innovation, enhancing teamwork, and maintaining a workplace culture built on trust and transparency.

Strategies for developing human-centric skills

To prepare for an AI-enhanced future, organisations and individuals must focus on skill development in key areas. Workday’s research highlights several strategies for strengthening human-centric capabilities:

1. Prioritising upskilling and reskilling

The evolving job market demands continuous learning. Businesses should invest in training programmes that enhance AI-related skills while reinforcing human strengths such as problem-solving, leadership, and adaptability. Employees who embrace lifelong learning will remain competitive in a shifting landscape.

2. Promoting human-machine collaboration

AI should be seen as a tool that enhances human capabilities rather than a replacement for human workers. By leveraging the tech for data-heavy tasks, employees can focus on strategic decision-making, creativity, and interpersonal relationships—areas where human intelligence is irreplaceable.

3. Strengthening communication and teamwork

AI can streamline workflows and facilitate collaboration, but strong interpersonal skills remain critical. Organisations should foster environments that encourage relationship-building, diverse perspectives, and collective problem-solving.

4. Cultivating human-centric leadership

Leadership in the AI age requires a shift toward empathy, emotional intelligence, and people-focused management. Effective leaders must balance AI-driven insights with human judgement, ensuring that employees feel valued, supported, and motivated.

5. Addressing skills gaps

A skills-first approach to talent development is essential. Organisations should identify gaps in human-centric competencies—such as ethical decision-making, cultural awareness, and resilience—and integrate these into training initiatives.

Also Read: Atome taps BlackRock, InnoVen for expanded US$80M credit facility

6. Building a culture of trust and transparency

For AI adoption to succeed, organisations must ensure transparency in AI-driven processes. Employees should have access to explainable AI systems and understand how technology impacts decision-making. Trust fosters a more inclusive and engaged workforce.

7. Encouraging ethical AI development

AI systems should align with ethical and organisational values. Businesses must equip employees with the skills to assess AI-driven decisions critically, ensuring fairness, accountability, and responsible technology use.

8. Strengthening critical thinking and problem-solving

AI can enhance analytical capabilities, but human judgment remains crucial. Training should emphasise creative reasoning, adaptability, and decision-making to ensure employees can interpret AI-generated insights effectively.

—

Image Credit: Annie Spratt on Unsplash

The post Human-centric skills in the age of AI: How to never lose touch with humanity in the workplace appeared first on e27.

The artificial intelligence (AI) landscape is evolving rapidly, and DeepSeek’s emergence could have significant implications for AI adoption and market dynamics. Bain & Company’s latest analysis presents a range of potential scenarios, offering key takeaways for businesses and industry leaders navigating this transformation.

DeepSeek’s impact will likely unfold across three possible scenarios, ranging from bullish to bearish, depending on how AI infrastructure costs and investments develop.

Bullish scenario: Expanding AI adoption
In an optimistic outlook, ongoing efficiency improvements lead to cheaper inference costs, accelerating AI adoption in a phenomenon known as Jevons’ paradox.

As AI becomes more accessible, demand for high-end training and advanced models will remain strong, encouraging sustained investment in cutting-edge AI capabilities. This scenario envisions a future where businesses increasingly integrate AI into their operations, leading to a broader and more dynamic AI ecosystem.

Moderate scenario: Infrastructure cost reduction
A more measured scenario predicts that while AI training costs remain stable, spending on AI inference infrastructure could decline by 30 to 50 per cent. This shift would prompt cloud providers to scale back their capital expenditures from an estimated US$80 billion to US$100 billion annually to a range between US$65 billion and US$85 billion per provider.

Despite this reduction, the expenditure would still represent an increase over 2023 levels, suggesting continued growth but at a more controlled pace.

Also Read: Gobi invests in ArmourZero to bolster SME cybersecurity defences

Bearish scenario: Constrained investment
In the most cautious outlook, AI training budgets shrink significantly, and spending on inference infrastructure declines sharply. Cloud providers’ capital expenditures could drop to between US$40 billion and US$60 billion, a level that, while still higher than in 2023, signals a slowdown in AI infrastructure expansion.

If realised, this scenario could indicate a temporary cooling-off period in AI investment, potentially leading to more selective AI deployments and a focus on cost efficiency rather than aggressive expansion.

Strategic considerations for CEOs

Given the uncertainty surrounding DeepSeek’s impact, Bain & Company’s report offers strategic advice for CEOs who want to effectively navigate the evolving AI landscape.

Prepare for cost disruption
Businesses should anticipate a future where AI inference becomes significantly cheaper, creating new competitive dynamics. Companies that proactively adjust their strategies to leverage more cost-effective AI solutions will be better positioned to capitalise on the changes.

This includes reassessing existing AI budgets and exploring new AI-driven opportunities beyond cost reduction.

Monitor market signals closely
CEOs must stay attuned to industry trends, particularly capital expenditure patterns, GPU demand, and AI adoption rates. A slowdown in infrastructure spending may indicate that efficiency improvements are reshaping AI economics. Understanding these shifts can help businesses adapt their AI strategies accordingly.

Also Read: How upcoming CPI data could influence fed policy and cryptocurrency prices

Key market signals to watch include:
– Sustained enterprise demand for high-performance AI models.
– Increasing restrictions on AI model access and distillation controls by leading AI labs.
– Validation of cost-saving projections and the emergence of previously unaccounted-for expenses.
– Evidence that DeepSeek was trained on existing models, potentially influencing AI development strategies.
– Continued prioritisation of advanced, frontier AI models for training purposes.
– The rapid proliferation of derivative models and new competitors.
– The growing popularity of low-cost open-source models, which may accelerate AI adoption in diverse sectors.

Think beyond productivity
While many companies initially adopt AI to improve operational efficiency, Bain & Company advises businesses to go further by leveraging AI to redefine their core offerings. The most successful firms will be those that move beyond automation and embrace AI-driven innovation.

This could take the form of personalised customer experiences, AI-enhanced product development, or entirely new services that leverage AI capabilities.

Broader implications for the AI market

DeepSeek’s emergence is part of a larger trend in AI development, where open-source and cost-efficient models are gaining traction. If AI inference costs continue to decrease, it could democratise AI access, enabling smaller businesses and startups to integrate AI solutions that were previously cost-prohibitive.

This shift could lead to a more competitive and dynamic AI ecosystem, where innovation is driven not only by major tech firms but also by emerging players leveraging new AI models.

At the same time, concerns over model security, data integrity, and ethical AI development are likely to remain at the forefront. Companies must balance cost considerations with responsible AI implementation, ensuring that AI models are not only efficient but also aligned with regulatory and ethical standards.

DeepSeek represents a pivotal development in AI, with the potential to reshape how businesses approach AI adoption and investment. Whether the market follows a bullish, moderate, or bearish trajectory, companies must remain agile, closely monitoring cost trends and market signals while actively seeking ways to innovate beyond mere efficiency improvements.

—

Image Credit: Mimi Thian on Unsplash

The post The 3 ways DeepSeek will impact industries and what business leaders can do about it appeared first on e27.

Across Southeast Asia, from Singapore’s fintech hubs to Indonesia’s e-commerce powerhouses, startup founders are evaluating DeepSeek’s promise of democratising advanced AI capabilities at a fraction of traditional costs.

With performance levels reportedly matching industry giants like GPT-4 and Claude-3.5, DeepSeek’s emergence marks a potential turning point for regional startups traditionally priced out of cutting-edge AI development.

But as regional tech leaders and investors weigh its implications, a crucial question emerges: Could this be the catalyst that propels Southeast Asia’s startup ecosystem into AI’s next frontier?

We spoke with several VCs, AI experts, founders and top executives at AI startups to learn how DeepSeek will impact the local startup ecosystem.

Below are the insights they shared:

Mauro Sauco, CTO of Transparently.ai, an accounting fraud detection firm

DeepSeek has generated buzz for a good reason. Its innovative approach to AI model development is already influencing how large language models are built and could fundamentally reshape our future in AI.

That said, I tend to be cautious when technology is surrounded by so much hype. While DeepSeek’s approach is innovative and impressive, I’m still on the fence about who will truly benefit from it in the long run.

Also Read: DeepSeek: The smart disruptor in the AI race

Here are some of my thoughts:

The impact on the ecosystem:

DeepSeek’s advancements have advanced the state of the art in LLM development. However, major frontier model developers are likely to quickly adopt these innovations, which might diminish DeepSeek’s lasting competitive edge.

Who stands to benefit:

Frontier model developers are likely to gain the most from incorporating this innovative approach to AI model development, which will lower costs and drive further innovation.

a) Large enterprises: Companies with specialised requirements might fine-tune DeepSeek to better suit their needs.

b) Regulated industries: Organisations that need to operate models within a controlled, private infrastructure could find value in DeepSeek.

c) Startups: For most startups, the direct benefits appear limited. However, any potential benefits might come indirectly through cost reductions passed down from the savings that mainstream frontier model providers eventually deliver.

Practical considerations for startups:

I’ve personally tested DeepSeek on metal (referring to the model itself, not the service) and had my team run some preliminary tests. For startups like ours, the practical benefits—especially in terms of cost savings and performance improvements—don’t seem substantial:

a) Infrastructure demands: Running DeepSeek locally (or on VMs) requires significant memory and GPU resources to achieve acceptable latency.

b) Production complexity: Setting up a production-grade system means managing redundancy, availability, and global distribution.

c) Operational costs: The overall costs and operational burdens can add up quickly.

d) Cloud provider offerings: Although some cloud providers are now offering DeepSeek via hosted endpoints, the advantages are minimal. DeepSeek R1 might be on par with, or only slightly superior to, existing reasoning models. Given that mainstream frontier models are likely to integrate these advancements in the near term, the effort and cost of refactoring software—including prompts and evaluations—to accommodate DeepSeek don’t seem justified.

Advantages of mainstream frontier models:

One major advantage of using established frontier model providers is their continuous improvement. These companies invest heavily in enhancing their models, meaning that as users, we benefit from ongoing enhancements with minimal effort on our part. DeepSeek might struggle to match this, especially when pitted against the giants in the field.

In conclusion, DeepSeek marks a significant advancement in the LLM ecosystem with its innovative approach to AI model development. While it’s clear that DeepSeek has reshaped the way we will approach AI model development in the future, its broader impact is still unfolding. For startups, in particular, sticking with established API services remains the more practical and cost-effective choice given the continuous improvements and financial backing of major frontier model providers.

Alvin Toh, co-founder of Straits Interactive, a data protection startup

DeepSeek’s affordability and accessibility democratise artificial intelligence (AI), making it attractive to ASEAN startups. Its low-cost, high-performing models enable developers in companies to rapidly prototype and integrate it into their technology stack and allow startups to innovate in various fields without prohibitive infrastructural investments.

However, there are concerns surrounding DeepSeek’s higher level of hallucination compared to other models in the market (notably, in following rules, writing, creativity, and persuasiveness) and certain biases from China in its outputs.

Additionally, DeepSeek’s ambiguous privacy policies and lack of robust compliance certifications bring pause. Its privacy policy lacks clarity on data usage, causing unease about data leakage and misuse. There is also no explicit mention of adherence to key international standards like GDPR, making it risky for startups operating in regulated industries.

Moreover, recent attempts by security researchers at Cisco and the University of Pennsylvania to jailbreak Deepseek’s model with adversarial prompts revealed that it failed to block all 50 attempts, indicating that there’s still a security gap in deployment.

Startups handling regulated, sensitive, or proprietary data must, therefore, carefully evaluate DeepSeek’s scalability and adherence to privacy laws, policies, and service-level agreements (SLAs). Compliance with local regulations, like the PDPA for Singapore-based outfits, needs to be ensured before integrating DeepSeek into critical workflows.

Until DeepSeek provides stronger assurances of data privacy and compliance, startups should presently avoid using it with sensitive data and consider alternatives with a privacy-first approach.

Rei Murakami, Venture Partner at Kadan Capital

DeepSeek has delivered a few important AI milestones: In December, version 3 was released with Mixture-of-Experts architecture, faster inference, and longer context windows, followed by the R1 model in January. But let’s be clear: the real story isn’t about one company.

LLMs are on a path to commoditisation, with open source gaining momentum. I believe the recent negative public market reaction is misplaced. In the mid-term, AI adoption is about to dramatically increase. History proves it; back in 1943, IBM’s Thomas Watson infamously predicted a world market for not more than five computers. When costs decreased, demand skyrocketed. The same is now happening with AI.

Also Read: DeepSeeking the future: The ripple effect on tech, crypto, and global markets

Nowhere is this shift more significant than in Southeast Asia. In emerging economies, SaaS solutions struggled with monetisation due to low labour costs. But with token costs set to drop, autonomous agents will become viable even in price-sensitive markets. The AI revolution isn’t slowing—it’s just getting started.

We continue to be excited about application layer AI, particularly in vertical SaaS, where AI can unlock entirely new business models and drive real operational impact.

Dr Sze Tiam Lin, Senior Licensing Advisor, SMU Institute of Innovation & Entrepreneurship

DeepSeek enables startups in Southeast Asia to harness cutting-edge generative AI technologies without the need for massive budgets and push the boundaries of what’s possible, offering a more dynamic and competitive landscape for innovation. It also allows smaller players to adopt efficient inference systems and benefit from significantly reduced training costs, making it easier to develop sophisticated AI solutions.

DeepSeek stands out in its efficient pre-training that accelerates the learning process and shortens the time required to deploy customised and powerful AI models. With performance levels on par with the best versions of GPT-4 and Claude-3.5, startups can harness high-level capabilities at a fraction of the cost.

Additionally, DeepSeek’s unique voting technique offers self-feedback on open-ended questions, enhancing the effectiveness and robustness of the alignment process, a critical factor in refining AI systems.

The pace of adoption across Southeast Asia will depend on the governance frameworks in place regarding the use of open-source models and data privacy in commercial deployments. Different jurisdictions may have varying regulatory environments, and this could affect innovation and AI adoption.

Matt Spriegel, CEO and founder of Atiom, an AI-powered gamified platform

Southeast Asia has been a hub for rapid digital adoption, and DeepSeek is lowering the barrier for startups to leverage AI early in their journey. From customer service automation to deep-learning analytics, its entry democratises access to AI, making advanced technology more accessible to early-stage companies.

Startups will have more access to AI without heavy R&D costs, which will also accelerate innovation. In particular, AI-first startups, especially in fintech, e-commerce and healthtech will likely experience accelerated growth.

Scalable AI solutions will also help startups to compete globally. This will impact the ecosystem, which will experience a surge in demand for more AI engineers, data scientists and machine learning (ML) specialists.

Furthermore, DeepSeek’s launch signals a shift in AI development beyond Western dominance. As more models are built and trained locally, we’ll see AI solutions tailored to Southeast Asia’s unique market needs, driving industry-specific innovation across the region.

Simon Davis, founder and CEO of GOAT Gaming, an AI-powered network of games on Telegram

DeepSeek’s emergence represents a transformative moment for Southeast Asia’s technology landscape, fundamentally altering the economics of AI deployment in the region. By dramatically reducing infrastructure costs for AI implementation, it levels the playing field for smaller companies that previously couldn’t compete due to resource constraints.

The ability to run these models locally is particularly significant for Southeast Asian startups. Not only does it slash operational costs, but it also addresses crucial data sovereignty concerns that have historically complicated AI adoption in the region. Companies can now process sensitive data on their own servers, ensuring compliance with local privacy regulations without compromising AI capabilities.

This cost-effective, local-first approach opens up exciting possibilities for market-specific AI solutions. In Indonesia, for instance, companies can now realistically develop specialised models tailored to the unique cultural nuances of the market. This localisation potential extends across Southeast Asia’s diverse markets, where one-size-fits-all solutions often fall short.

The implications are profound. We might see a surge of innovative AI applications emerging from previously underserved markets. This democratisation of AI technology could catalyse a new wave of regional innovation, powered by companies that understand their local markets intimately.

Nitin Vyas, Sr. VP (Product & Data) at RedDoorz, a budget hotel network

As a hospitality technology company, we view the emergence of DeepSeek as part of the industry’s ongoing evolution and also making AI a level-playing field for everyone.

DeepSeek, as an entity focused on AI and advanced technology solutions, represents an opportunity to enhance operational efficiency, personalise services, and elevate customer experiences in the hospitality sector. A healthy competition among the AI giants will be beneficial for this industry as it gives a more balanced global view of how AI will shape the course of our civilisation.

Also Read: AI, personalisation, and 5 marketing activities you should be doing

This competition may enable corporations to implement AI more cost-effectively. At the same time, we could see an even faster pace of innovation and more advanced capabilities in the large language model (LLM).

Remi Choong, Elev8.vc, a VC firm

Cost-effective AI infrastructure: Increasingly cost-effective models, such as DeepSeek’s, will lower API costs. This enables AI startups to deliver cheaper and better solutions, similar to the disruptions we’ve seen in the solar panel industry.

Stronger demand for AI: With more affordable options, global AI adoption and spending will likely surge. Southeast Asian startups must quickly adapt to agile go-to-market strategies to capture the growing demand, particularly among SMEs.

Competition: Given lower AI costs, it is increasingly important for companies to adopt AI to maintain a competitive advantage. We can also expect to see an uptick in adoption across a broad spectrum of industries.

Balancing AI and hard tech: Investors need to ensure a balance between funding cutting-edge AI ventures and the infrastructure that supports them, especially in countries like Singapore with strong hard tech capabilities.

Jeff Pan, Belli.ai, which builds air cargo software for airlines

Speaking broadly, it’s unlikely that you’ll see the same velocity of change in Southeast Asia as you do in the EU/US, primarily due to the types of problems that we are solving.

Founders in the EU/US typically deal with high-capability customers who already have high internal capabilities and large IT budgets. In contrast, founders in Southeast Asia typically deal with early digitisation problems (helping low-capability customers with small IT budgets transition from paper and spreadsheets to basic CRUD applications), which will largely be unaffected by the cutting-edge progress you see being made by DeepSeek.

Much of the impact you will see (which is already happening) is that teams can generate 10x more impact with smaller teams, which you will see play out in (a) less hiring demands from startups, who don’t need as much headcount and (b) VC funds skewing away more towards top 10 per cent founders who can seize on these advantages rather than a broader base of portfolio companies.

Nofi Bayu Darmawan, founder and CEO of Komerce, an e-commerce enabler

The promise of AI in simplifying tasks and creating value, especially in the context of cost-effectiveness is huge.

When it comes to e-commerce, the impact of generative AI on customer engagement, especially its role in social media interaction, e-commerce, and automating customer service, is massive. DeepSeek can consolidate brand knowledge and communication across various platforms, including Google reviews and online chat on websites, enhancing the commerce ecosystem.

Warren Leow, CEO of Designs.ai

Deepseek has sparked a lot of interest because its progress has lowered service costs. This spurs more innovation and adoption among users as barriers to adoption are reduced.

In addition, having more choices across different vendors and technology providers, especially from open source, will only spur further improvements in the quality of applications being produced.

 

The post The DeepSeek debate: Opportunity or overhype for startups in ASEAN? appeared first on e27.

The Asia Pacific was once primarily recognised for its kaleidoscope of cultures, but today, it’s increasingly synonymous with groundbreaking tech developments. A gradual shift in the global order has led many to reconsider the vast opportunities that APAC, with its blend of mixed economies, traditions, and lifestyles, has to offer.

While the US—particularly Silicon Valley—has long been seen as the epicentre of tech and innovation, the momentum is now steadily shifting eastward. The reasons are clear: maturing economies, improved mobility, and greater accessibility have positioned APAC as one of the most fertile grounds to test ideas and tap into emerging markets.

Take Indonesia’s rural areas, for example, where efforts to democratise financial services stand in stark contrast to the sophisticated fintech landscape of Singapore, a city-state with a financially savvy populace. Yet Singapore, with its reliance on imports and limited land for waste, faces unique sustainability challenges that are driving a shift toward a circular economy. Meanwhile, in Japan, an aging population is accelerating the development of health tech and medtech, spurred by a staunch focus on well-being.

These examples are just the tip of the iceberg. Across the region, a surge of innovation is not just changing industries but reshaping everyday life in profound ways. The scope and scale of what’s unfolding are unparalleled, making APAC a pivotal space for the future of global innovation.

To further shed light, here are eight tech trends defining APAC in 2024:

Artificial intelligence goes vertical

The initial buzz around AI revolved around its potential for general-purpose tasks, like chatbots—think OpenAI’s ChatGPT, Anthropic’s Claude, and Baidu’s Ernie Bot. However, these broad applications often fall short in meeting the intricate demands of specific industries. This gap is giving rise to vertical AI, where models are tailored to specific needs.

• Adopt industry-specific AI: Businesses should evaluate the benefits of customising AI models to address the unique challenges of their industry, ensuring precision and relevance.
• Leverage local expertise: Engaging with regional AI ecosystems can provide the cultural and market-specific insights needed to refine these solutions effectively.
• Invest in tools: To fully capitalise on vertical AI, invest in technologies that enhance data organisation and accessibility—crucial for AI model training.

Self-driving revolution

The thought of cruising through the streets of Jakarta or Bangkok without the frustration of gridlock seems far-fetched now, but autonomous vehicles could soon make it a reality. Beyond reducing traffic, self-driving technology promises to enhance road safety and mobility, especially for the elderly and disabled.

Also Read: ‘AIR’ review: 3 lessons for dealmaking and entrepreneurship

• Initiate pilot projects: Companies in logistics and urban planning should collaborate to test the viability of autonomous vehicles in high-traffic cities.
• Expand and integrate services: Explore how self-driving technology can complement existing mobility services, enhancing efficiency and accessibility.
• Keep compliant: Companies venturing into this space should stay aligned with evolving regulations to maintain public trust and ensure safe implementation.

Electrification is the future

As the global push for sustainability accelerates, the shift to electric vehicles is becoming inevitable. With rising energy demands and international commitments like those from COP28, the transition from fossil fuels to renewable energy sources is gaining momentum.

• Be the first mover: Proactively invest in technologies that facilitate the transition to renewable energy, positioning your company as a leader in the field.
• Scale up infrastructure: Prioritise the development of EV charging networks to support widespread adoption and operational efficiency.
• Collaborate beyond borders: Engage in cross-border partnerships to access a wider array of sustainable innovations that fit your operational needs.

Humanoid robots near commercialisation

Robotics is no longer confined to assembly lines or science fiction. Humanoid robots, capable of replicating human movements and decisions, are on the verge of widespread adoption, bringing unprecedented efficiency and adaptability to various industries.

• Identify use cases: Businesses should assess which areas would benefit most from humanoid robots, starting with controlled environments like manufacturing.
• Partner with innovators: Collaborate with companies to ease the integration of robotics into your operations, even if your business has no prior experience.
• Prepare for workforce transition: Develop comprehensive training programs to ensure your workforce adapts smoothly to robotic assistance.

Easier, more convenient payments

Despite advancements in financial literacy and services, many still struggle with access—particularly in regions where traditional financial systems exclude those without formal credit histories or who live in remote areas. Overcoming these barriers is essential for economic equality and broader participation in the financial ecosystem.

• Broaden financial access: Financial companies should consider developing inclusive products, such as buy now, pay later (BNPL) services, to cater to underserved populations.
• Streamline transactions: Focus on enhancing cross-border payment processes through strategic partnerships and technology adoption.

Cradle-to-cradle

The depletion of natural resources is a critical challenge that threatens both the environment and future economic stability. Shifting from a linear economy to a cradle-to-cradle approach—where materials are continuously reused—is no longer optional, but necessary.

Also Read: Exploring the rise of finance-as-a-service in APAC

• Embrace circular practices: Integrate recycling and reuse into your production processes to align with cradle-to-cradle principles.
• Engage in regional initiatives: Participate in frameworks like ASEAN’s circular economy initiatives to align with broader sustainability goals.
• Invest in sustainable innovation: Prioritise R&D in technologies that transform waste into valuable resources, ensuring long-term environmental and economic resilience.

Syncing data on the edge

As digitalisation sweeps across industries, the need for high-speed, low-latency networks has never been greater. The explosion in internet and mobile usage, combined with the integration of AI and IoT, is pushing traditional networks to their limits.

• Adopt edge computing: Reduce latency and improve real-time data processing by integrating edge computing into critical applications.
• Accelerate 5G rollout: Companies should expedite the deployment of 5G networks to meet the growing demands of connectivity.
• Explore alternative technologies: Consider innovative solutions like laser-based communication to complement traditional network infrastructures.

Taking a quantum leap

Nvidia’s rise highlights the critical role of semiconductors, but the limitations of traditional computing, dictated by Moore’s law, are becoming evident. Quantum computing offers a way forward, promising to solve complex problems exponentially faster than classical systems.

• Invest in quantum research: Stay competitive by keeping abreast of the latest developments in quantum computing, which will impact most industries.
• Develop quantum-ready solutions: Begin bridging the gap between classical and quantum computing to ensure a smoother transition as the technology matures.

The seismic shifts in APAC’s tech landscape are more than just emerging trends—they are redefining the future of global innovation. From AI specialisation to the rise of electric mobility and quantum breakthroughs, these developments present both challenges and unparalleled opportunities.

To explore these trends in greater depth and understand their far-reaching impacts, join us at GITEX Asia from April 23–25, 2025, at Marina Bay Sands, Singapore. This event is where the future takes shape, and where you can engage directly with the minds and technologies driving these changes. Be there to witness, contribute, and lead.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. Share your opinion by submitting an article, video, podcast, or infographic.

Join us on Instagram, Facebook, X, and LinkedIn to stay connected.

Image courtesy of the author.

This article was first published on November 5, 2024

The post APAC’s tech revolution: 8 trends shaping the future of global innovation appeared first on e27.

Key highlights:

• Trump’s tariff decision: A 25 per cent tariff on steel and aluminium imports, including from Mexico and Canada, raises trade war concerns
• Market reactions: US equities remain resilient due to strong corporate earnings, while gold and the US dollar rise amid uncertainty
• Bond and energy markets: Treasury yields show mixed signals; Brent crude oil prices rise due to geopolitical tensions
• Asia’s response: Mixed market reactions, with optimism driven by DeepSeek’s AI model despite trade concerns
• Japan’s crypto regulation shift: Possible reclassification of cryptocurrencies as securities, impacting investments and tax policies
• Bitcoin accumulation by strategy: Strategy (formerly MicroStrategy) purchases 7,633 BTC, reinforcing institutional confidence in crypto

The financial landscape is navigating an ever-shifting environment, with risk sentiment holding steady despite significant macroeconomic developments on 11 February 2025.

One of the most notable events in recent days has been President Donald Trump’s decision to impose a 25 per cent tariff on all steel and aluminium imports, a move that includes key trading partners like Mexico and Canada without any exemptions. This policy, enacted under Section 232 of the Trade Expansion Act, has sent shockwaves through global markets, raising fears of potential trade conflicts and their broader economic fallout.

Trump has also hinted at the possibility of further increasing these tariffs and suggested the introduction of reciprocal tariffs, which could be announced as early as today or Wednesday. These developments have heightened market uncertainty as investors and analysts closely monitor whether these threats will materialise and how they might reshape global trade dynamics.

At the same time, the US corporate earnings season has provided a stabilising force, with strong performances from American companies reinforcing confidence in the economy’s underlying health.

However, the interplay between these macroeconomic and microeconomic factors, alongside other global trends such as Japan’s potential reclassification of cryptocurrencies and significant Bitcoin acquisitions by firms like Strategy (formerly MicroStrategy), paints a multifaceted picture of the current financial environment.

In this article, I will explore these developments in detail, analyse their potential impacts, and offer my perspective on how they shape the global risk sentiment.

Tariffs and market reactions

Let’s start with the tariff announcement, which has dominated financial news and market discussions in recent days. President Trump’s decision to impose a 25 per cent tariff on steel and aluminium imports under Section 232—a provision that allows the president to restrict imports deemed a threat to national security—marks a significant escalation in US trade policy.

Unlike previous tariff actions, which often included exemptions for key allies, this move explicitly excludes Mexico and Canada, two of the United States’ largest trading partners. This lack of exemptions has raised concerns, as it signals a more aggressive and unilateral approach to trade policy. Trump’s comments over the weekend and his warning that tariffs could “go higher” have added to the uncertainty, with market participants now bracing for the possibility of reciprocal tariffs.

Reciprocal tariffs, if implemented, would involve matching the tariff rates of other countries on US exports, potentially triggering retaliatory measures from affected nations. The timing of these potential announcements—possibly today or Wednesday—has kept markets on edge, as investors weigh the risks of a broader trade conflict.

Also Read: AI, personalisation, and 5 marketing activities you should be doing

From a market perspective, the immediate reaction to the tariff news has been varied. US equity indices, as measured by the MSCI US Index, rose by 0.7 per cent on Monday, with strong performances in the energy sector (+2.2 per cent) and information technology (+1.5 per cent). This resilience suggests that, for now, investors are focusing on the positive fundamentals of American companies rather than the potential negative impacts of tariffs.

The US earnings season has been particularly strong, with many companies surpassing expectations despite what analysts had considered a high bar. This strength in corporate fundamentals has provided a buffer against the macro uncertainties, supporting risk sentiment in the short term.

However, the longer-term implications of tariffs cannot be ignored. Tariffs on steel and aluminium could increase input costs for industries such as manufacturing, construction, and automotive, potentially squeezing profit margins and stoking inflation. If reciprocal tariffs are introduced, US exporters could face higher costs in foreign markets, further complicating the economic outlook.

Turning to the bond market, US Treasury yields ended Monday’s session with mixed results. Shorter-term yields, such as the two year and seven year, edged lower, reflecting some caution among investors about the near-term economic impact of tariffs.

Conversely, longer-term yields, including the 10-year (+0.2 basis points to 4.497 per cent) and 30-year (+1.4 basis points to 4.707 per cent), inched higher, suggesting that investors expect inflationary pressures from tariffs to persist over the longer term. This divergence in yield movements highlights the uncertainty surrounding the Federal Reserve’s next moves. Tariffs, by increasing costs and potentially delaying rate cuts, could complicate the Fed’s efforts to balance inflation and growth.

The US Dollar Index, meanwhile, rose by 0.3 per cent, reflecting safe-haven demand amid the tariff-related uncertainty. Gold, a traditional safe-haven asset, surged by 1.7 per cent to a fresh record high, underscoring investor concerns about geopolitical and economic risks. In the energy market, Brent crude oil prices rose by 1.6 per cent, supported by signs of a tighter market and geopolitical tensions, including Russia’s failure to meet its OPEC+ quota and rising natural gas prices in Europe.

Asian markets and crypto regulations

In Asia, the HSCEI index rose by 2.1 per cent for the third consecutive day, driven by optimism surrounding DeepSeek’s AI model and a perception that tariff tensions might be less severe than feared. However, early trading sessions on Tuesday showed mixed results for Asian equity indices, with US equity futures pointing to a lower open. This divergence highlights the uneven impact of tariff-related developments across regions.

While US markets have been buoyed by strong earnings, Asian markets remain more exposed to trade risks, given their reliance on exports. The resilience of risk sentiment in Asia, particularly in China, can also be attributed to positive developments in the AI sector, with companies like DeepSeek demonstrating resilience despite trade tensions. However, the broader implications of tariffs on global supply chains and economic growth remain a concern, particularly for export-dependent economies.

Also Read: How marketers can connect with APAC’s 450 million young gamers

Shifting focus to other global developments, Japan’s Financial Services Agency (FSA) is considering a significant regulatory change that could reclassify cryptocurrencies as securities. This potential shift, which could take effect by 2026, would have far-reaching implications for retail investors and the broader financial ecosystem. By classifying crypto as securities, Japan aims to strengthen investor protections, lower taxes on crypto investments, and enable domestic funds to invest in tokens.

This move could also pave the way for the approval of crypto exchange-traded funds (ETFs), including spot Bitcoin ETFs, which would attract institutional capital and boost market liquidity. Posts on X have highlighted the FSA’s plans, with some users speculating on the potential for tax cuts and ETF approvals.

However, these reports remain inconclusive, and the FSA’s final decision will depend on a comprehensive review of existing regulations. If implemented, this reclassification could position Japan as a leader in the global crypto market, potentially offsetting some of the negative sentiment surrounding tariffs.

Another notable development in the crypto space is the recent acquisition by Strategy (formerly MicroStrategy) of 7,633 Bitcoin for US$742 million between February 3 and February 9, at an average price of US$97,255 per Bitcoin. The firm now holds 478,740 Bitcoin, worth over US$46 billion, with an average purchase price of US$65,033 per Bitcoin.

This acquisition, representing 2.2 per cent of Bitcoin’s total supply, underscores the growing institutional interest in cryptocurrencies as a store of value and hedge against inflation. Strategy’s aggressive Bitcoin strategy has been closely watched by investors, with some viewing it as a bullish signal for the crypto market.

However, the timing of this acquisition, amid tariff-related uncertainty and rising gold prices, raises questions about the firm’s risk management approach. While Bitcoin has historically been seen as a safe-haven asset, its volatility and correlation with risk assets like equities could complicate its role in a tariff-driven market environment.

Balancing risk and optimism

From my perspective, the current global risk sentiment is a delicate balance between optimism and caution. On one hand, the strength of US corporate earnings and positive developments in sectors like AI and crypto provide a foundation for resilience. The MSCI US Index’s gains, driven by energy and tech, reflect confidence in the underlying fundamentals of the economy.

Similarly, Japan’s potential reclassification of crypto and Strategy’s Bitcoin acquisition signal growing institutional acceptance of digital assets, which could support risk sentiment in the longer term. On the other hand, the tariff announcement and the threat of reciprocal tariffs introduce significant uncertainty.

Also Read: Embracing global entrepreneurship: Redefining startup success beyond Silicon Valley

Tariffs, by increasing costs and disrupting supply chains, could stoke inflation and weigh on economic growth. The mixed performance of US Treasury yields, the surge in gold prices, and the rise in Brent crude oil all point to heightened concerns about the macroeconomic outlook.

In my view, the key question for markets is whether the positive microeconomic factors—such as strong earnings and innovation in AI and crypto—can continue to offset the negative macroeconomic risks posed by tariffs. While US markets have shown resilience so far, the potential for retaliatory measures from trading partners like China, Mexico, and Canada could escalate tensions and disrupt global trade.

For Asia, the optimism surrounding DeepSeek’s AI model and less severe tariff fears may provide temporary relief, but the region’s exposure to trade risks remains a concern. Japan’s potential crypto reclassification, if implemented, could be a game-changer, attracting capital and boosting sentiment. However, the success of this move will depend on the FSA’s ability to balance investor protections with market growth. Strategy’s Bitcoin acquisition, while bullish for crypto, also highlights the challenges of navigating a volatile market environment.

In conclusion, the global risk sentiment is supported by a combination of strong corporate fundamentals and positive developments in AI and crypto, but it remains vulnerable to tariff-related uncertainties. President Trump’s tariff announcement, under Section 232, has introduced significant risks, with the potential for reciprocal tariffs adding to the complexity. While US markets have been buoyed by earnings, the longer-term implications of tariffs on inflation, growth, and trade dynamics cannot be ignored.

In Asia, optimism surrounding AI and crypto provides a counterbalance, but the region’s exposure to trade risks remains a concern. Japan’s potential crypto reclassification and Strategy’s Bitcoin acquisition are positive signals, but their impact will depend on broader market conditions. As markets navigate this busy macro news backdrop, the interplay between microeconomic resilience and macroeconomic risks will shape the trajectory of global risk sentiment in the coming weeks and months.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. Share your opinion by submitting an article, video, podcast, or infographic.

Join us on Instagram, Facebook, X, and LinkedIn to stay connected.

Image courtesy of the author.

The post Global risk sentiment holds steady amid tariffs, AI optimism, and crypto shifts appeared first on e27.

B Capital, a global multi-stage investment firm, has taken a significant step to strengthen its foothold in tech and AI investments by appointing Yan-David (“Yanda”) Erlich as General Partner.

Announced in January, this strategic move underscores the firm’s commitment to backing transformative AI-driven companies across early-stage and growth investments. With a track record of scaling tech ventures and investing in high-impact AI innovations, Erlich’s leadership is expected to further elevate B Capital’s presence in this rapidly evolving sector.

A seasoned entrepreneur, operator, and investor, Erlich brings a wealth of experience to his new role. Prior to joining B Capital, he served as COO and CRO at Weights & Biases, a leading AI developer platform, and held investment roles at Coatue Management.

His entrepreneurial background includes founding and scaling multiple venture-backed startups, including ChoiceVendor, which LinkedIn acquired. With this blend of hands-on operational experience and deep investment acumen, Erlich is poised to drive B Capital’s AI strategy forward.

Speaking about his vision for the firm’s AI investments, Erlich highlights B Capital’s strong foundation of entrepreneurs, operators, and investors as a key advantage.

Also Read: Global risk sentiment holds steady amid tariffs, AI optimism, and crypto shifts

“B Capital already brings together an experienced team; it was a key reason I chose to join. Our strategic partnership with BCG is also a differentiating competitive advantage, as they have a unique vantage point on the AI transformation,” he shares.

In an email interview, Erlich shares his insights about AI and how B Capital is approaching investment in the space.

The following is an edited excerpt of the conversation.

AI is rapidly evolving. What criteria or signals do you prioritise when evaluating potential investments in AI startups? Are there any red flags you watch out for in this space?

When things change quickly, I find it useful to go back to basics. What is true for (most) successful businesses: they are started by founders who care deeply about their customer base and market.

They have a compelling product that is deeply loved and actively used by their customer base. Especially at first, it is better to have fewer, more avid customers than many lukewarm ones.

They move fast: build, ship, learn from the market, iterate. One of my most durable insights has been, “A startup is not a company: it is an experiment to see if a company deserves to exist. Until product-market fit, the speed of hypothesis testing trumps everything. Startups die when they spend faster than they learn.”

They are hyper-focused: startups are better suited to solving a small set of very hard problems than many semi-hard ones.

Also Read: AI agents redefine art: Unlocking boundless creative possibilities in a new digital era

They are talent magnets: great people love to work with other great people.

They think about the whole problem: not just what product to build but also how to take it to market, price it, and market it.

The red flags I watch out for are when folks claim that it is “because of AI”. These axioms of what makes a high-quality startup no longer apply.

As an investor in both early-stage startups and growth-stage companies, how do you balance the inherent risks of nascent technologies with the need for scalability and long-term impact?

The quick answer is that it is not easy. On one hand, I am a huge believer in technology’s positive transformational impact: improving lives and human productivity and creating new category-defining winners. When betting on a company, I always ask myself why now: what new disruption permits this company to win?

On the other hand, business fundamentals apply. You need a great product with high engagement, operate in a large (or, ideally, quickly growing TAM), constantly build and defend your moats, maintain quality as you grow your teams, and more.

Balancing building for the future and “solving the problems of today, today; the problems of tomorrow, tomorrow” is part of the difficulty (and the fun).

It is important to be diligent in what we can: at the earliest stages, that is, the founders and technology, and whether they solve an acute pain in a large or growing market. Later, operational execution and market positioning matter. It is always important to see if the company is a great magnet for top talent across both tech and GTM.

It is also key to help where we can: through my personal network, the whole B Capital network, BCG, and many more. Making the investment is the beginning of the adventure.

Also Read: What we can tell about AI investment in SEA this year

From your perspective, what are the most promising trends in AI today? Conversely, what do you see as the biggest challenges for AI startups in gaining traction or achieving product-market fit?

AI is going to reshape every aspect of the economy. I am particularly excited to invest in startups helping to usher in the era of AI co-workers: intelligent agents working alongside humans in all aspects of work, from functional roles including marketing (B Capital is an investor in Writer), engineering (Poolside), and legal (EvenUp) to verticals across robotics (Apptronik), climate (Overstory), and healthcare (Atomwise).

To achieve this, we will also see more advances at the foundation model layer, including in advanced reasoning, personalisation, context, memory, and the ability to take and act on feedback. We will also see new infrastructure solutions, new security models, and ways to “onboard” agents into your organisation and have them collaborate with humans. Each of these allows for one or more very interesting companies.

The challenge that is top-of-mind right now is how to bridge the gap between the high-quality AI demo and the AI application or agent the organisation feels safe to deploy to production. I am confident this will be solved, but it will not be a silver bullet. It will be achieved through a combination of model advancements, developer tooling, continuous evaluations, guardrails and other safety mechanisms, and novel tech and UX, including better feedback mechanisms, at the application layer.

Are there specific industries or verticals where you see the most transformative potential for AI technologies, and how does B Capital aim to support innovation in these areas?

I will use the opportunity to reinforce that I’m grateful to be part of a team that has both a global purview and where I get to work alongside climate and healthcare investment experts. Our combined and global subject-matter expertise feels uniquely competitive.

—

Image Credit: B Capital

The post B Capital General Partner Yanda Erlich on the red flags he notices when investing in AI space appeared first on e27.

Life Science Incubator (LSI) has unveiled its largest co-working laboratory in Singapore at Elementum, located in JTC’s one-north business park.

As per a press statement, LSI offers flexible lab and office solutions to meet the diverse needs of life science companies. The facility includes open lab spaces and private suites with dedicated tissue culture rooms.

Tenants can also customise their spaces, with options to design bespoke wet lab and office areas.

The co-working lab offers flexible contract arrangements and full-service support, catering to startups and established companies. The facility also fosters a vibrant and innovative community, facilitating tenant collaboration.

Also Read: The future of medtech in Singapore: Innovation amid regulatory challenges

With the expansion, Life Science Incubator has tripled its laboratory space. LSI also plans to expand into the broader Asia Pacific region, with Australia as its next key market. LSI has been actively engaged in the Australian life sciences ecosystem for three years and is in discussions to potentially launch its first Australian location later this year.

LSI provides fully equipped, agile, and reliable lab spaces for biotech, medtech, and foodtech companies. It works closely with local accelerators, universities, and polytechnics to stimulate entrepreneurship and provide external resources to support spin-offs and new ventures.

“Our mission at LSI is to remove barriers for life sciences startups and innovators by providing the critical infrastructure they need to accelerate breakthroughs,” said Zeïna Henni, Director of Life Science Incubator.

Singapore’s life sciences sector has seen strong investor confidence, with 14 companies raising SGD92.4 (US$68) million in 2024, making it the top destination in Southeast Asia for biotech investment. The country’s commitment to the sector is reinforced by significant investments under the Research, Innovation and Enterprise 2025 plan (RIE2025), which allocated approximately SGD28 (US$21) billion in 2024 to key sectors, including life sciences.

Furthermore, Singapore’s regulatory framework and intellectual property protections attract leading biotech and medtech firms.

The post Life Science Incubator expands in Singapore with new co-working lab appeared first on e27.

This is how eFishery went from agritech hotshot to a total PR disaster.

Once the poster child of Indonesia’s agritech industry, eFishery’s reputation is officially in crisis mode. The narrative is being written for them, fuelled by media coverage and rumours (or media coverage written based on rumours). Without an authoritative source, proactive engagement, and transparency, the company is spiralling further into crisis instead of taking control of their story.

How the story unfolded

It all started when DSA broke the story from a whistleblower tip that eFishery had been falsifying transactions, inflating financial reporting, and creating shell companies. The company reported a US$16 million profit, but investigations uncovered a US$35.4 million loss. This was then followed by extensive coverage from Bloomberg News, The Jakarta Post, and Tech in Asia.

And eFishery? Instead of owning the narrative, they issued the following response:

“We are fully aware of the gravity of the market speculation and we take this matter with the utmost seriousness,” eFishery said in an emailed statement. “We remain dedicated to upholding the highest standards of corporate governance and ethics in all of eFishery’s operations.”

This is the equivalent of watching a house burn down and saying, “We are aware of the fire and take fires very seriously.” This is saying something without saying anything.

By most accounts, eFishery’s technology and product are legitimate. And yet, one of the brightest moments in eFishery’s comms wasn’t even from the company itself – it came from a product manager’s LinkedIn post.

A LinkedIn post from an eFishery product manager helped restore some credibility by publicly confirming the authenticity of its technology, including IoT-powered feeders and water sensors. Yet, the company’s leadership should have been at the forefront of the response.

Also Read: Ecosystem Roundup: eFishery faces fraud allegations | Indonesia’s tech funding hits a 3-year low | iMotorbike raises US$10M

The real people behind this scandal

Outside of the 50+ page investigation reports and financial audits, this crisis involves real people – over 1,000 employees, investors, partners, and customers who deserve clear, honest communication. In most situations and also likely in this case – 99.9 per cent of the 1,000+ employees were not responsible and have probably spent countless hours post-leak trying to fix the issue.

At this stage, from a communications perspective, it is important for leadership to articulate a clear path forward and a strategy to rebuild trust.

How eFishery can salvage at this point and takeaways for founders

• Issue a formal, transparent statement: Avoid the vague, generic PR responses and provide real answers.
• Put a human face to the response: Start using a credible, human face to address concerns head on.
• Proactively engage with stakeholders: Internal memos, town halls, direct investor communications. Silence breeds speculation.
• Commit to real governance reforms: Appoint independent auditors, strengthen compliance, and communicate changes regularly.
• Create a long-term communications plan, and execute it.

eFishery can still be salvaged but only if they are proactive in their communications. At this inflection point, leadership has two choices: take control of the narrative and move forward or continue letting others define their story.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. Share your opinion by submitting an article, video, podcast, or infographic.

Join us on Instagram, Facebook, X, and LinkedIn to stay connected.

The post How eFishery lost control of its narrative appeared first on e27.

For years, the AI landscape was dominated by a few tech giants—Google, OpenAI, Meta—driven by the belief that training Large Language Models (LLMs) was an exclusive domain of companies with billions of dollars in computing resources. DeepSeek has just shattered that notion.

The historical parallel: When AI was considered untrainable

Neural networks were once considered impractical due to their training complexity. In the early days, before multi-layer perceptrons gained traction, single-layer networks struggled with even basic tasks. The introduction of backpropagation in the 1980s, pioneered by Geoffrey Hinton and his colleagues, was a turning point—it showed that deep networks could be trained effectively, giving birth to what we now call deep learning. Suddenly, what was once considered an untrainable, niche research field became the dominant paradigm in AI.

A similar shift is happening today. Previously, the assumption was that only trillion-dollar companies could afford to train LLMs. But DeepSeek has proven that with the right architectural optimisations and efficiency techniques, smaller players can break through.

The DeepSeek disruption: Smart moves over raw power

DeepSeek has demonstrated that AI isn’t just about brute-force computation—it’s about architectural intelligence. By leveraging techniques like:

• Mixture of Experts (MoE): Only activating necessary parts of the model to reduce computational overhead.
• Distillation: Training smaller models to mimic the performance of larger ones.
• Smarter resource utilisation: Running LLMs at a fraction of the cost of GPT-4.

DeepSeek has built a competitive model with just US$6 million in training costs, compared to OpenAI’s rumoured US$100 million for GPT-4. This is a fundamental shift, proving that AI dominance is not solely a function of computational power, but also of innovation in model design.

The market reaction: Nvidia’s dip and the reality of AI economics

Following DeepSeek’s announcement, Nvidia’s stock saw a temporary decline—an overreaction by the market, mistaking this development as a sign of declining GPU demand. In reality, it’s part of AI’s natural evolution: as architectures become more efficient, the focus shifts from sheer hardware reliance to algorithmic ingenuity.

This is a reminder that AI is not just a hardware race—it is an intellectual one. The best AI systems will not necessarily come from those who spend the most money but from those who think the most creatively.

Also Read: DeepSeeking the future: The ripple effect on tech, crypto, and global markets

China’s play: From hard work to smart work

China has long been perceived as a country that achieves success through relentless execution. DeepSeek challenges that stereotype, showing that Chinese AI research is not just about scaling hardware, but about making smart strategic moves. By proving that efficient models can rival state-of-the-art LLMs, DeepSeek has redefined the AI playing field, making it clear that the next AI breakthroughs may come from outside the traditional Silicon Valley elite.

The future: A democratised AI ecosystem

DeepSeek’s approach signals a new era—where smaller companies, startups, and research institutions can meaningfully compete in LLM development. This democratisation of AI will lead to:

• More innovation in architectures and training methodologies.
• Cost-efficient models that can be deployed widely.
• Increased competition, driving AI forward at an even faster pace.

In the same way that backpropagation unlocked deep learning’s potential, DeepSeek’s cost-efficient breakthroughs are making high-performance LLMs accessible beyond the corporate elite. The AI revolution is no longer about who has the most GPUs—it’s about who has the smartest approach.

DeepSeek has sent a clear message: The AI race is far from over, and the winners will be those who innovate, not just those who spend.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. Share your opinion by submitting an article, video, podcast, or infographic.

Join us on Instagram, Facebook, X, and LinkedIn to stay connected.

Image courtesy of the author. 

The post DeepSeek: The smart disruptor in the AI race appeared first on e27.