Savana Asset Management marks the first anniversary of its Savana US Small Caps Active ETF (ASX, SVNP), reporting a strong 13.62% annualised return that outpaced its benchmark despite a volatile US market. The firm also reveals insights from a novel AI stress test highlighting current limitations of large language models in investment decision-making.
- SVNP delivered 13.62% p.a. since inception, outperforming benchmark by 13.61%
- ETF showed resilience during a sharp ~20% US market drawdown
- Disciplined exit strategy realised gains but occasionally missed further upside
- AI stress test reveals large language models are sensitive to wording and sentiment
- Savana emphasises disciplined, valuation-driven active management over AI reliance
A Year of Strong Performance Amid Market Turbulence
November 2025 marked the one-year anniversary of Savana Asset Management’s Savana US Small Caps Active ETF (ASX – SVNP). Despite a challenging environment for US small caps, SVNP has delivered an impressive 13.62% annualised return since inception, outperforming the S&P 600 Total Return Index (AUD) by 13.61% per annum. This result stands out in a year when small caps broadly struggled, validating Savana’s active management approach.
The ETF’s resilience was tested during a sharp market drawdown of approximately 20% in April and May, triggered by concerns over US tariff policies. SVNP’s algorithms demonstrated their robustness by mitigating downside risk and capturing significant upside during the subsequent rebound. On average, SVNP outperformed the market by nearly 2% in rising months and held steady with a slight outperformance during down months.
The Discipline Behind the Gains – Exit Strategies in Focus
Savana’s letter highlights the importance of strict exit discipline in their investment process. The firm shared examples such as Canadian Solar, where a disciplined sale after a 180% gain avoided subsequent price retracements. However, not all exits captured the full upside; notable cases include Nebius Group NV and SanDisk Corp, where the ETF exited positions before substantial further gains occurred, driven by surging demand linked to AI infrastructure.
This trade-off underscores Savana’s philosophy – their valuation-driven, algorithmic approach prioritises consistency and avoids behavioural biases, even if it means occasionally leaving some upside on the table. This disciplined process is credited with underpinning the strategy’s long-term robustness.
AI in Investment Management – Promises and Pitfalls
In a forward-looking section, Savana presents a stress test of large language models (LLMs) like Claude Sonnet 4.5 and ChatGPT to evaluate their reliability in generating investment recommendations. While LLMs showed high stability when given identical inputs, they were surprisingly sensitive to neutral rewordings and sentiment framing. Negative framing, in particular, skewed recommendations heavily toward selling, despite unchanged factual content.
This sensitivity raises concerns about the current readiness of AI tools to independently drive live portfolio decisions. Savana concludes that while AI can accelerate research and streamline workflows, it lacks the consistency and abstraction needed for dependable investment management without careful human oversight.
Looking Ahead – SVNP’s Role in a Changing Market
Despite ongoing uncertainty in US markets, Savana remains optimistic about SVNP’s prospects. The ETF offers a differentiated exposure to US small-to-mid caps, providing a valuable complement to traditional core portfolios dominated by large-cap indices like the S&P 500. Whether markets improve or face continued headwinds, Savana believes active management and disciplined stock selection will be crucial to navigating the evolving landscape.
With a solid foundation established in its first year, SVNP’s journey will be closely watched by investors seeking both growth and resilience in US small caps.
Bottom Line?
SVNP’s disciplined approach shines in year one, but the evolving role of AI in investing remains a critical watchpoint.
Questions in the middle?
- How will SVNP navigate potential shifts in US small-cap market dynamics in 2026?
- Can Savana’s disciplined exit strategy balance capturing upside with avoiding behavioral biases over the long term?
- What advancements are needed for AI tools to become reliable decision-makers in active portfolio management?