The AI Barbell Strategy: Reason Behind the Elimination of the Medium Level Participants
In the rapidly evolving AI market, a distinct trend has emerged, favouring the extremes over the middle ground. This phenomenon, known as the barbell strategy, has significant implications for companies, investors, and talents alike.
On one end of the spectrum, we find the left stack, composed of open source models, commodity hardware, and a lack of regulatory oversight. This side of the market, while small, boasts a degree of immunity due to its open-source nature, making it challenging for regulators to keep up.
Conversely, the right side of the AI market is characterised by proprietary models, custom silicon, and direct engagement with regulators. This side requires substantial resources for compliance and has the power to influence rules through lobbying.
Real-world evidence suggests that startups attempting to occupy the middle ground are struggling or disappearing, while those on the extremes are thriving. This trend is reflected in the venture capital (VC) landscape, where 90% of investments are guaranteed to fail, 9% return capital, and 1% need to return 1000x.
The middle of the AI market is grappling with the burden of compliance costs that are crushing companies, a lack of influence on rules, and being caught in regulatory nets. As a result, this sector is projected for complete extinction, with assets being absorbed by the extremes, talent being redistributed, and investors being educated.
The traditional bell curve distribution of companies, returns, and opportunities has been replaced by the barbell structure in the AI market. Companies must either be at the level of OpenAI or commodity providers, with no profitable middle ground.
The barbell strategy involves holding extremely safe assets and extremely risky bets, avoiding the middle entirely. The middle positions break under stress, have limited upside, are exposed to significant losses, and are harder to manage than extremes.
Geographically, AI hubs are either San Francisco, focusing on frontier AI, or open internet, focusing on commodity AI. Regional AI hubs are failing and losing talent to the extremes.
The left side of the talent barbell in AI consists of open source contributors, hobbyists, researchers, cost-conscious startups, and academic institutions. On the other hand, the right side includes companies like OpenAI, Anthropic, and DeepMind, offering compensation packages of $5M+ and unlimited resources for cutting-edge research.
For enterprises, the procurement barbell strategy involves commoditizing everything that doesn't matter and paying premium prices for critical solutions that create advantage, while eliminating middle vendors and solutions.
The right side of the AI market is projected to have 2-3 total winners with trillion-dollar valuations, platform monopolies, and attempts at AGI. Potential winners likely to establish a strong position include European players like Aleph Alpha and Mistral AI, supported by EU investments in AI Gigafactories and research centres.
For talent, the career barbell strategy suggests joining giants for resources and impact or going solo for freedom and upside, while avoiding the middle (companies that will die). For investors, the portfolio barbell strategy involves making many small bets on open source/community projects and a few huge bets on potential winners, with no investments in the middle (guaranteed losses).
For startups, the strategic implications suggest choosing between completely free, open source, community-driven options (left), raising $1B+ to compete for AGI (right), or avoiding the middle (death zone). The decision matrix for enterprises suggests using commodity options when the task is well-defined, the quality bar is low, volume is high, and cost sensitivity is extreme, and using frontier options when competitive advantage is needed, quality is crucial, innovation is required, and cost is less important.
Nassim Taleb's barbell approach, which rejects the bell curve middle, advocates for 90% ultra-safe and 10% ultra-risky positions, with no medium risk. The distribution strategy in the AI market also favours the extremes, with the left distribution being self-service, viral, and free, and the right distribution being direct sales, partnerships, and platforms.
In conclusion, the AI market is splitting into two extremes: massive winners and minimal survivors, with everything in between getting crushed. To survive and thrive in this landscape, companies, investors, and individuals must choose to embrace the extreme safety of commoditization or the extreme risk of frontier competition.
Read also:
- Linde Wins Major Engineering Design Contract for Equinor's Low Carbon Hydrogen Project at H2H Saltend, Progressing Towards a Greener Future
- Stock markets in Asia experience a surge following a record-breaking rally in U.S. stocks, fueled by optimism towards potential interest rate reductions.
- Transportation via roads plays a critical role in India's shift towards clean energy.
- EU Commission President von der Leyen emphasizes continuity, affirming that Europe's automotive future will be powered by electricity
