Snowflake Added to the NDI-FutureTech Portfolio

Discussing the addition of Snowflake into the NDI-FutureTech portfolio, available on eToro.

At our Q3 rebalance, Snowflake ($SNOW) was added to the NDI-FutureTech portfolio. Earlier in the year, it was on our radar, but at 20-30x EV/S, investors were pricing SNOW to perfection, expecting flawless execution without any hiccups, for many quarters and years. To us, this represented a very unfavourable risk-reward, because at such valuation levels, it would be extremely hard for SNOW to surprise on the upside, but relatively easy to disappoint investors, which is why we kept on the sidelines.

The valuation has since more than halved, as SNOW investors have swung from peak optimism to substantial pessimism. This swing in sentiment has been triggered by larger-than-anticipated growth deceleration and subsequent doubts that the company will not be able to reach its FY28 revenue target of $10bn unless they reaccelerate growth. This darkening cloud over SNOW emanates from ongoing cloud cost optimizations and growing competition from Databricks in the cloud data warehouse market.

But digging deeper, it seems that investors believe SNOW's moat is weakening, in large part due to their decision to support and facilitate Iceberg tables. A significant part of SNOW's moat has been having its customers' data stored inside Snowflake. Hence, on the surface, allowing customers to store their data in an open storage format threatens SNOW's moat. We believe the alpha exists in the notion that most investors do not see that in practice, this could result in 100x more compute through SNOW, which is the higher margin aspect of their business.

Complicating SNOW's play regarding Iceberg tables, is Databricks recent decision to acquire Tabular, the commercial entity behind Iceberg. However, our differentiated view is that this move highlights Databricks' own weaknesses regarding its approach to developing and monetizing Delta Lake, an alternative open-source storage format. The acquisition could complicate matters for Databricks, leading to poor ROI. Overall, SNOW's approach to facilitating Iceberg is a more straightforward strategy with better ROI, despite its rival now owning the main contributor to Iceberg.

We see another source of alpha emanating from analysts' views on Frank Slootman retiring and Sridhar Ramaswamy becoming the new CEO. Slootman is a favourite among Wall Street, having led both ServiceNow ($NOW) and SNOW to huge successes. Wall Street doesn't like change while a company is performing strongly, which is generally the case for SNOW, despite recent growth declines. However, we see that Slootman wasn't the right person to take SNOW into the next era of GenAI. With Ramaswamy's pedigree in machine learning and GenAI, we're confident he is the right person to steer the Snowflake ship in this direction.

We feel that these sources of alpha, along with several products recently made GA (generally available) and more to become GA this year, will give SNOW a good opportunity to positively surprise the market. However, it's possible that SNOW could disappoint in the short term due to substantial recent changes. In the upcoming ER on 21st August, investors will look for hints of strong early traction among customers for these new products, though meaningful contributions should be expected in FY26.

SNOW's future success largely depends on their expansion into unstructured data domains. So far, Databricks seems to be having more success expanding from unstructured to structured data than vice versa. Databricks claims to have increased revenue from Photon, their data warehouse product, from $100m ARR in 2022 to $350m ARR in 2023. While this may be true, as a private company, their claims face less scrutiny, and we have found via our research that they are partial to making exaggerated claims.

Generally, SNOW still has a clear lead on the structured side, with superior governance, controls, and sharing capabilities. However, their unstructured data offering remains immature, and Databricks maintains a user experience advantage among developers and data professionals.

With Ramaswamy at the helm, SNOW's path to maturity should accelerate, potentially transferring its strong user experience from structured to unstructured data domains.

At a high level, the competition between SNOW and Databricks will likely be determined by who best balances open and closed systems and functionality. SNOW began with a closed system that is gradually opening up, while Databricks is moving in the opposite direction, with much of this competition manifesting in their approaches to Iceberg tables.