- Cadence Design Systems as a proxy for the semiconductor stack;
- SAP, Intuit and Workday as our risk-off picks, backed by a combination of enterprise-heavy client bases and resilient, non-discretionary revenue streams;
- Datadog, Twilio and ServiceNow as risk-on plays, owing to their consumption/ usage-based pricing models that lead to poorer earnings visibility, especially in conditions of economic volatility; and
- Alibaba Group Holding, Horizon Robotics and RoboSense Technology Co for exposure to China — a market too important to ignore when it comes to AI and robotics innovation.
Here are brief overviews for each company, outlining the rationale behind their inclusion in our thematic AI Portfolio, which represents not just a bet on the future of AI but also a bet on the future itself.
Cadence Design
See also: What’s agentic AI, and how will it affect you and your investments?
Semiconductors — CPUs (central processing units) and GPUs (graphics processing units) and AI accelerators — are foundational to the AI revolution, and the semiconductor value chain begins with the EDA (electronic design automation) industry, a near-duopoly dominated by Synopsys and Cadence.
In brief, EDA provides the software tools for engineers to design, test and validate chips prior to the manufacturing process. Thus, the EDA industry may be considered a good proxy to the wider semiconductor sector — without the same capital intensity, cyclicality and geopolitical risks of chip manufacturers and foundries.
We prefer Cadence over Synopsys because of its superior portfolio of back-end tools, which will arguably be increasingly prioritised as the industry shifts towards chiplet-based architectures. Between the two companies, Cadence Design also has higher gross margins (FY2024: 86% versus 80%) and marginally lower exposure to China (12% versus 16%).
SAP
SAP is the incumbent leader in ERP (enterprise resource planning) software. In recent years, the company has made notable progress in transitioning from on-premise licensing to cloud-based SaaS (Software-as-a-Service) business model via its RISE (existing customers) and GROW (new customers) cloud migration initiatives. Cloud revenue now represents half of total sales. With fewer than one-third of its existing customers fully migrated, this positions SAP with a substantial runway for driving additional cloud revenue growth and realising greater economies of scale within its cloud infrastructure. As with peers in the enterprise software space, SAP cloud strategy increasingly incorporates AI integration — including tools such as Joule (generative AI [GenAI] agents) and Knowledge Graph, which enhances decision-making in complex operations such as supply chain and procurement. Margin expansion and a clear AI road map make SAP a convincing inclusion to our portfolio.
Intuit
Intuit is a fintech SaaS platform best known for QuickBooks and TurboTax — industry standards in tax preparation and accounting for US small and medium businesses (SMBs). The acquisitions of Credit Karma and Mailchimp, in 2020 and 2021 respectively, expanded Intuit into a full-stack platform for small business growth — covering customer acquisition, credit-scoring, lending and financial management. While this broader product suite has been modestly dilutive to gross margins, it enhances customer stickiness and creates cross-sell potential. Intuit’s product ecosystem is collectively unified under its proprietary GenAI operating system, GenOS. In FY2024, Intuit’s online ecosystem revenue grew 20% year on year, while ARPC (average revenue per customer) rose 11%. On a broader level, Intuit is a direct beneficiary of workplace digital transformation and cloud migration.
Workday
Workday offers cloud-native ERP with a focus on HCM (human capital management) — an area it leads alongside Oracle, according to Gartner’s 2024 Magic Quadrant. With the latter increasingly focused on scaling its cloud infrastructure and data centre offerings, however, Workday enjoys a relatively unconstrained runway to extend its leadership. Over the past year, Workday has accelerated its AI-driven product strategy, notably through the launch of the Illuminate suite of AI agents. Adoption momentum appears strong thus far: AI SKU (stock keeping unit) expansions grew 30% y-o-y in the last quarter of its financial year ended January 2025. Further updates, including a System of Record platform that enables users to monitor and manage agentic workflows, as well as further expansions to the Illuminate suite are slated for later this year. For FY2025, subscription-based revenue accounted for 91% of total revenue, while gross revenue retention — which tracks the percentage of recurring revenue retained from existing customers before any upsells — stood at a healthy 98%.
Datadog
For more stories about where money flows, click here for Capital Section
As digital complexity rises, so does the need for observability. Datadog, a market leader in observability and security SaaS, enables enterprises to monitor servers, databases and applications with high granularity, thus providing critical visibility into digital operations. In particular, its LLM Observability suite addresses the challenges of monitoring AI workloads — where high compute costs make it critical to detect security vulnerabilities, system failures and costly downtimes early. The company’s acquisitions of Metaplane and Eppo in April and May 2025 respectively further solidify its edge in the end-to-end AI observability niche: Metaplane monitoring AI training data quality and Eppo enabling rigorous testing of AI model performance during development.
Twilio
Twilio operates in the CPaaS (Communications Platform-as-a-Service) sector, providing cloud-based APIs (Application Programming Interfaces) that enable developers to embed voice, video, messaging and email capabilities directly into customer-facing applications. For instance, its fintech-oriented Verify API allows fintech customers to incorporate OTP and two-factor authentication functionalities via SMS/WhatsApp to their platforms. With GenAI driving renewed interest in voice-based interfaces, Twilio is increasingly focused on AI voice intelligence. Conversation Relay, launched in December 2024, combines STT (speech-to-text) and TTS (text-to-speech) technologies for real-time, human-like interactions ideal for AI-powered call centres and customer support. Separately, Twilio’s customer data platform (called Segment) provides another compelling growth narrative of its own. Segment aggregates and unifies customer data — both from interactions flowing through Twilio’s platform and from third-party sources — into a centralised database. With proprietary data the new currency in the age of GenAI, Segment provides the kind of clean, high-volume data that is increasingly essential for analysis and training.
ServiceNow
As a leading cloud-based IT service management provider, ServiceNow specialises in digital workflow automation through its Now Platform — integrating disparate business functions such as IT services, human resources (HR) and customer relationship management (CRM) (where enterprises might otherwise rely on separate vendors such as Microsoft or Salesforce) into a unified management system.
AI has become a core growth driver for the company in recent quarters. In 1Q2025, management attributed strong results to the continued uptake of its Now Assist GenAI suite. Notably, the company reported 30% y-o-y growth in its US public sector segment — underscoring the clear cost-efficiency and operational benefits driving adoption even in traditionally budget-conscious sectors. Building on the momentum from 2H2024, when the company rolled out a major platform update (called Xanadu) that introduced integrated AI capabilities, ServiceNow has continued its AI expansion with new agentic rollouts across CRM, HR and IT functions, and announced the acquisition of Moveworks, a leading agentic AI start-up.
Alibaba
While the tech spotlight is focused mainly on Silicon Valley, China’s AI and cloud sector is rapidly evolving — with Alibaba at the forefront.
As China’s largest cloud-service provider, Alibaba leads in market share in both IaaS (Infrastructure-as-a-Service) (28.5%) and PaaS (Platform-as-a-Service) (26%) markets, dwarfing the next biggest contenders Huawei (13.3%) and Tencent Holdings (11.8%). IaaS refers to providing on-demand computing power, such as servers, storage and network; and PaaS refers to providing the development environment to build, test and deploy applications. Together with SaaS, these layers form the backbone of cloud computing.
Separately, Alibaba has also been investing aggressively in foundational AI models — for both general-purpose and specialised applications alike. Notable examples include the March 2025 release of Qwen2.5-Omni, which excels at edge inferencing, as well as Qwen2.5-VL in April, which is optimised for vision-based workloads such as object recognition and tracking.
On the whole, the capital-intensive nature of hyperscalers — driven by competition in data centre expansion and R&D — makes stable revenue streams essential for sustaining infrastructure growth and AI development. In this respect, Alibaba benefits from being the leading e-commerce player in China, with a 42% market share via Taobao and Tmall. With the Chinese government’s policy shift towards stimulating domestic consumption — the most recent list of stimulus measures having been announced on May 7 — Alibaba is well positioned to capitalise on the expected resulting uptick in consumer spending.
Horizon Robotics
Horizon Robotics is a key enabler of smart driving in China. It is the third-largest provider of ADAS (advanced driver assistance systems) — typically, lower-level automation — and fourth largest in autonomous driving (AD) solutions globally. Its Journey 6 chips power L2+ systems in mass-market vehicles, from BYD’s “God’s Eye C” to Li Auto’s “AD Pro”.
Horizon Robotics’ value proposition lies in its flexibility and cost efficiency: Unlike fullstack rivals such as Huawei, Horizon Robotics adopts a modular playbook — offering chips, intelligent driving operating systems and related hardware à la carte to automakers.
With Chinese automakers ramping up adoption of autonomous features, Horizon’s midtier pricing focus gives it a long runway for growth. Recent regulatory scrutiny, prompted by a fatal accident involving an ADAS vehicle, has centred on marketing practices (such as requiring accurate labelling of the autonomous level and banning remote software updates) and is unlikely to materially affect the industry’s development.
Horizon Robotics targets a four-year revenue CAGR (compound annual growth rate) of over 50% from 2024 to 2028, driven by both increased demand in ADAS/AD technology and higher prices from product upgrades. The company plans to launch, in the second half of the year, its Journey 6P AD chip, which rivals Huawei’s Ascend and Nvidia’s Orin series chips, for Level 2/3 autonomous driving but at a fraction of their costs.
RoboSense
RoboSense is a global leader in LiDAR (light detection and ranging) technology for automotives — where its long-term growth potential is closely tied to the eventual mainstream adoption of Level 3+ autonomous driving.
In this context, LiDAR remains the most precise solution for real-time 3D spatial awareness — critical for safety and advanced driving scenarios. As such, while some automakers — including Xpeng, the pioneer in mass-produced LiDAR-based smart cars with the P5 in 2021 — have shifted towards lower-cost, vision-based solutions for existing Levels 1 and 2 ADAS systems, LiDAR is expected to remain indispensable as the industry moves towards higher levels of autonomy. Consequently, this positions RoboSense for a long-term structural growth opportunity as smart driving technology continues to evolve.
Looking ahead, RoboSense is also expanding into non-automotive use cases — particularly in robotics and industrial automation. As at 2024, robotics accounted for just 12% of revenue, but management has expressedly flagged this sector as a strategic growth priority, given the overlap in sensor requirements across both industries.
In March 2025, RoboSense launched its Active Camera (AC1) series, which integrates LiDAR, cameras and motion sensors into a unified spatiotemporal sensing module — ideal for advanced robotic motion control and navigation. Currently, the company is a supplier for home-grown robotics giant Unitree Robotics.
As previously mentioned, Tong’s AI Portfolio started with a capital of US$100,000, fully invested. For the week ended May 14, the portfolio gained 1% lifting total portfolio value to US$101,031. The top gainers were Datadog (+4.9%), Twilio (+2.3%) and Cadence Design (+2.2%), while the big losers were Intuit (-1.2%), RoboSense (-0.6%) and Horizon Robotics (-0.5%).
The Malaysian Portfolio, on the other hand, fell 0.3% last week, paring total returns since inception to 185.4%. Kim Loong Resources (+0.9%) closed higher, while United Plantations (-1.8%) and Insas Bhd – Warrants C (-12.5%) finished in the red. This portfolio continues to outperform the benchmark FBM KLCI, which is down 13.5% over the same period, by a long, long way.
The Absolute Returns Portfolio was up 1.3% for the week, lifting total returns since inception to 23.8%. US and global stocks continued to recover from the recent selloff. The top three gaining stocks in our portfolio were Goldman Sachs (+10.8%), JPMorgan Chase (+6.5%) and Tencent (+5.7%). SPDR Gold (-5.7%) and Berkshire Hathaway (-2.9%) were the only two losers last week. We added Alibaba to the portfolio. The stock has since gained 1% from our initial investment cost. The acquisition pared cash holdings to 32% of total portfolio value.
Disclaimer: This is a personal portfolio for information purposes only and does not constitute a recommendation or solicitation or expression of views to influence readers to buy/sell stocks, including the particular stocks mentioned herein. It does not take into account an individual investor's particular financial situation, investment objectives, investment horizon, risk profile and/ or risk preference. Our shareholders, directors and employees may have positions in or may be materially interested in any of the stocks. We may also have or have had dealings with or may provide or have provided content services to the companies mentioned in the reports.
