Real AI Conversations: Bull and Bear Market Trends, Slowing Growth, Future Insights
NVIDIA is the third stock ever to reach a $3T market cap which has accelerated the S&P and Nasdaq to reach new record highs. Microsoft (MSFT), NVIDIA (NVDA) and Apple (APPL) make up 20% of the S&P Index. The collective market cap across all three companies totals $9.2Bn, an almost $3Bn increase, from a year ago. Concentration risk fuels concern if a broader downturn in big tech names stall as recent growth slowdown in the sector, has been cited and overhype AI concerns have increased.
Recent bullish outlooks related to AI, include Goldman who expected a 15% increase in US labor productivity and GDP from generative AI, while Bank of America believes that due to a lack of material rise in tech volatility, AI is not a bubble yet. Additionally, Morgan-Stanley notes that historically the S&P 500 produces above-average returns, in periods of rising concentration.
Outside of equity related AI bull and bear standpoints, general hesitations and discussions are being examined as considerations have come into play. The pace of AI improvement is slowing, and fewer applications exist for even the most advanced AIs. Additionally, building and running AI is extremely expensive. Although new, competing AI models frequently emerge, they take time to impact how most people work. These factors raise concerns about AI’s potential to become commoditized, its ability to generate revenue and profits, and whether it will truly create a new economy.
As it relates to the slowing pace of AI, most advancements in large language models like OpenAI’s ChatGPT and Google’s Gemini result from feeding them increasing amounts of data. However, companies have almost exhausted available internet data, with no additional vast sources of human-generated content left to utilize.
In terms of commoditization, a mature technology is one that everyone knows how to build, and without significant breakthroughs, no one has a performance edge. Companies then focus on cost-cutting. This scenario, seen in the electric vehicle industry, now appears to be happening with AI.
The concerns associated with the cost to run AI, reflect a significant often cited figure, “industry spend on NVIDIA chips to train AI totaled $50Bn in 2023 but generated only $3Bn, in revenue.” For popular services relying on generative AI, operating costs far exceed the already high training costs.
AI adoption remains uncertain as OpenAI reported $2Bn in 2023 revenue, that’s expected to double, by 2025. However, this is far from the revenue needed to justify its nearly $90Bn valuation. Recent user interest from new features has been high, but it’s unclear if users will remain long-term. Changing mindsets and habits will be one of the biggest barriers to AI adoption, a consistent pattern in the rollout of all new technologies.
There is no argument that AI will not continue to transform industries and jobs in the long run. However, previous scenarios related to new emerging technologies should temper expectations about AI’s continued rapid improvement and quick adoption, as the evolution of AI may require more time than presumed.
