Does China’s focus on AI applications in robotics make China resistant to the AI mania of the West? While China's strategy is more grounded in practical, industrial AI application with less speed of capital influx into the sector, and more tangible return on the money, the West is all in on something super.
This assessment highlights a fundamental divergence in US and Chinese AI strategies that has intensified this year.
• China's Practical, Industrial Focus: China is heavily prioritizing AI integration into the real economy—specifically manufacturing, robotics, and logistics—aiming for tangible, measurable returns, or "AI+". This approach is often more "open" (open-source) to accelerate adoption, utilizing a "two-loop" strategy where industrial data feeds back into model improvements.
• The "Super" Focus of the West: The US and its allies are heavily focused on developing "frontier" models, striving for Artificial General Intelligence (AGI) and large-scale model performance.
• Capital Influx Disparity: While the West is investing heavily in this "super" pursuit—with US private AI investment in 2025 reaching $285.9 billion, over 23 times more than China's $12.4 billion.
The 2026 Landscape:
• The Performance Gap Closed: Despite the massive difference in private capital investment, the performance gap between the top U.S. and Chinese models has effectively closed.
• Industrial Dominance: China leads in AI-powered robotics, with nearly nine times the volume of industrial robot installations as the U.S.
• Open vs. Closed: As Western firms guard their models, Chinese labs are largely, though not entirely, embracing open-source, allowing for rapid deployment and refinement.
In essence, the West is gambling on a "superintelligence" breakthrough, while China is building on industrial, real-world utility.
The emotional picture in the West is somewhat uncomfortable:
“What a rally! Qualcomm is up 1,200% in the past year, Sandisk 915%, Nvidia 390%, Lam Research 460% and MicroStrategy 1,260%.
Hold on a minute, didn't they change their name to "Strategy?" Sorry, we were looking at a table of stock performance ending in March 2000, not today. As Jonathan Krinsky and his team at BTIG point out, many of today's marquee companies also were top performers back then.
Comparing today's action to the peak of the greatest American stock bubble ever is always going to invite pushback. It is also unoriginal: Pundits have overlaid charts showing recent gains for the Magnificent Seven with dot-coms, the Roaring '20s, railroad fever, Japanese property and tulip bulbs.
But echoes of the tech wreck are getting hard to ignore. In some ways the recent surge has pushed the AI boom past that episode's heyday.
Through the third week of March 2000, for example, the top-performing 10 stocks in the Nasdaq Composite were up an average of 622%. Through this past Tuesday, BTIG points out they were up 784%.
One difference that AI optimists point to is that price/earnings ratios for the hottest stocks back then were loftier than the likes of very profitable Nvidia, Alphabet or Microsoft today.
Yet the S&P 500's trailing and cyclically adjusted P/E multiples are almost as high. That's because AI-related stocks make up a larger share of the index than "TMT" (technology, media & telecoms) did 26 years ago. And since what most investors own is the S&P 500, or baskets of stocks that look similar to it, descending to a more typical valuation would be just as unpleasant.
Some are reassured that today's hand-wringing actually gives the rally legs because the market loves to "climb a wall of worry." Everyone on Wall Street supposedly was wearing rose-colored glasses back in early 2000.
But that isn't quite true. Look at this paper's coverage during the heady weeks before the peak and there was plenty of guessing about when and where it would stop. On the day the Nasdaq Composite peaked above 5000 points, one strategist cautioned that Fed tightening eventually would end the party.
"Will the index be at 6000 when that happens? 5200? It's impossible to say."
And it's still impossible. Valuation is useless as a timing tool and recent earnings from tech giants have mostly been good or great.
Manias often peak on good news, though, and they generally go out with a bang, not a whimper.
Technical indicators paint a picture that's similar to tech's last hurrah.
Even the names ring a bell.” [1]
1. It's Looking an Awful Lot Like the Tech Bubble. Jakab, Spencer. Wall Street Journal, Eastern edition; New York, N.Y.. 11 May 2026: B10.
Komentarų nėra:
Rašyti komentarą