Discover more from The Monthly Flex
The Monthly Flex - July '23
A collection of topics on tech and the startup industry brought to you by Flex Capital.
Our Latest Thinking.
Decoding Tech's Next Big Thing: VR vs. AI
The rising consensus in the tech world points to AI as the “new, new thing”. It’s similar to the advent of the Internet and mobile in terms of importance, but with a key difference – it’s a sustaining technology that primarily benefits established players. In stark contrast, VR stands as a disruptive technology, a new platform where users can spend time, thereby generating opportunities for startups.
Once too hailed as the “new, new thing”, the initial enthusiasm for VR has dwindled due to current technological limitations. However, the recent launch of Apple’s Vision Pro in June sparked renewed excitement around VR. In this piece, we lay out our framework for assessing whether it’s worth investing in the Vision Pro platform and compare its potential to that of AI’s prospects.
We deep dive into what Vision Pro could mean for companies and investors in our latest article:
What Matters Now.
SafeGraph CEO & Flex Capital GP Auren Hoffman hosted Immad Akhund, CEO & Founder of Mercury and former Partner at Y Combinator. Immad is also a serial founder and a Dual Threat CEO with over 300+ angel investments.
Tune in for a truly incredible episode 👇
🏦 SVB collapse and future of the banking system
🏧 How Mercury handled an influx of over $2B in deposits
💸 Fundraising from 5,000+ investors
💰 UI as a strategic advantage
Choose your AI Adventure: French Laundry vs. Taco Bell
Databricks' recent acquisition of MosaicML for $1.3B is certainly sparking conversation among AI enthusiasts, as it signals the emergence of two distinct schools of thought surrounding AI infrastructure.
If your aspiration is to become the world's best restaurant, you almost have no choice but to become a chef. With that lofty goal, you’d be unlikely to trust the cooking to anyone else. However, if your main objective is to launch a business as quickly as possible, you might find it more advantageous to go with a franchise model and bet on a proven formula that “just works” but might lack a special kick.
In the realm of AI, a similar principle applies. Companies can either commit to a vendor ecosystem, leveraging platforms like OpenAI as their core infrastructure, or they can deploy and fine-tune LLMs on top of their own data with solutions such as Mosaic.
Similar to the franchise approach, investing in a vendor ecosystem can accelerate market entry and simplify adoption, but given the fact that the same models are accessible to all players in the market, this approach could impede a company's ability to differentiate itself and establish AI-driven competitive edges.
On the other hand, companies that develop and train unique models based on their proprietary data can build strong defensibility, with open-source foundation models such as Mosaic’s MPT-30B serving as their core infrastructure.
As the usage of such AI-powered products grows, customers continually input their own data, augmenting the capability of these models by facilitating further fine-tuning and generating higher customer-facing value in return. This constant iterative improvement generates a cycle of accumulating advantages for companies, allowing them to distinguish themselves in the same way as a one-of-a-kind restaurant that offers an experience that can’t be replicated elsewhere.
There is no doubt AI will soon be a core component of every major software application. The Mosaic acquisition strategically positions Databricks at the forefront of this tidal wave and gives it the potential to serve as the platform of choice for enterprises looking to build proprietary data advantages while competitors such as OpenAI remain poised to take over the SMB market.
Few startups consider M&A as part of their strategy. Perhaps it seems like the domain of more established and often public companies, or perhaps it seems too expensive or complex for an early-stage company to pull off successfully.
On the other hand, founders that do consider M&A tend to be overoptimistic about the potential outcomes and lack a strategy for successfully navigating this path. More often than not, they’re also not aware that most of the work should actually happen before even considering an acquisition.
In his recent post, Datavant founder and former CEO Travis May explains why start-ups may want to engage in M&A and shares his playbook for how other start-up CEOs can effectively incorporate M&A into their strategies.
Travis is uniquely suited to share this advice as he has now led 15 transactions (3 on the sell-side and 12 on the buy-side) ranging from 6-figure to 10-figure deals. He also largely credits M&A for accelerating Datavant’s path by several years, establishing product-market fit and product scale much earlier.
Bonus reading: Want to take the process in reverse? Check out Auren Hoffman's guide on Mastering the Art of Spin-Outs.
See you next month 👋