SEB’s AI Fund Posts Strong First-Year Performance, Highlighting Growth in AI Infrastructure and Trends

SEB’s AI Fund Posts Strong First-Year Performance, Highlighting Growth in AI Infrastructure and Trends

(IN BRIEF) SEB’s newly launched AI fund has reported impressive first-year results, outperforming global benchmarks under the management of Alexander Winberg. The fund focuses on investing in companies advancing artificial intelligence, with its performance exceeding both world and benchmark indices. A significant trend has been the extensive deployment of AI infrastructure by major cloud providers like Amazon, Google, and Microsoft, which is crucial for AI development. As the infrastructure expands, future trends are expected to include a shift towards “Application Specific” models, which are tailored for specific tasks, and the emergence of “Edge AI,” which processes data directly on devices. In its first year, the fund achieved a 45.2% increase, with Nvidia emerging as a top investment performer, while Oracle and AbCellera faced challenges.

(PRESS RELEASE) STOCKHOLM, 2-Oct-2024 — /EuropaWire/ — SEB’s recently launched fund, dedicated to investing in artificial intelligence (AI) companies, has experienced a remarkable first year, outperforming global market benchmarks. Fund manager Alexander Winberg shares insights into the factors contributing to this success and discusses emerging trends in the AI landscape.

While past performance does not guarantee future results, and investors should be aware that fund values can fluctuate, the fund’s initial returns have been impressive, surpassing both world and benchmark indices. Interested parties can find detailed information about the fund’s performance, sustainability disclosures, and prospectuses at seb.se/fonder.

Surprising Developments in AI Infrastructure

One of the most notable surprises over the past year has been the extensive deployment of AI infrastructure. Major cloud providers, including Amazon, Google, and Microsoft, have significantly expanded their capacities, establishing themselves as comprehensive resources for AI technology.

The construction of AI infrastructure resembles that of data centers, albeit on a much grander scale. These AI data centers often require advanced technology, utilizing Graphics Processing Units (GPUs) that allow for parallel computations, vastly enhancing processing speeds compared to traditional Central Processing Units (CPUs). Notably, AI data centers can consume 5 to 10 times more energy than standard data centers.

Ongoing Expansion of AI Capabilities

The growth of AI infrastructure appears to be a continuing trend. While precise data on infrastructure expansion remains limited, estimates suggest that the first iteration of ChatGPT utilized around 10,000 GPUs, while Meta is expected to deploy a model requiring approximately 350,000 GPUs by year-end. Microsoft is rumored to be planning a data center with over 1 million GPUs by 2028. Importantly, improvements in semiconductor technology are expected to double GPU performance roughly every two years, indicating a significant potential for future growth.

Anticipating Future AI Trends

As the infrastructure for AI expands, a notable shift is anticipated towards “Deployment” models. Although numerous AI models exist today, their current application remains relatively low compared to the substantial investments in infrastructure. For instance, it has been reported that in 2023, spending on AI infrastructure exceeded software investments by a factor of more than 50.

While general-purpose models like ChatGPT have gained widespread recognition, the next wave of AI utilization is expected to focus on “Application Specific” models, which are tailored for specific tasks and trained on specialized datasets. These models not only perform better for particular applications but are also less expensive to train.

Another emerging trend is “Edge AI,” which involves processing data directly on devices rather than relying on cloud infrastructure. Companies like Apple are actively working to integrate AI capabilities into smartphones, and this shift is poised to play a crucial role in the automation of manufacturing and the development of self-driving vehicles.

Fund Composition and Performance

The SEB AI Fund comprises a range of innovative companies, including Adobe, which offers an AI-enhanced version of Photoshop known as Firefly, significantly reducing image editing time. Another example is RELX, whose Lexis+ AI model improves efficiency for legal professionals by saving 6-7 hours weekly while increasing accuracy by 30% compared to general models. In the coding sector, Salesforce claims to save approximately 20,000 hours monthly using AI tools.

In its inaugural year, the fund achieved a notable 45.2% increase, outperforming its benchmark index, which rose by 35% from June 30, 2023, to June 28, 2024.

Top and Bottom Investments

Among the fund’s standout investments, Nvidia has performed exceptionally well, soaring around 200% since the fund’s inception. As a critical provider of GPUs necessary for AI infrastructure, Nvidia holds a dominant position in the market. Other strong performers include cloud service providers such as Meta, Alphabet (Google), and Microsoft.

Conversely, Oracle has lagged behind its cloud competitors in database services, although it has shown signs of improvement. Additionally, AbCellera, which leverages AI to expedite drug discovery, remains in the early stages of development, impacting its current performance.

Media Contact:

Press hotline: +46 8 763 91 10
e-mail:press@seb.se.
SEB’s switchboard: +46 771 621 000

SOURCE: Skandinaviska Enskilda Banken AB (SEB)