Axios AI+

June 10, 2024
I'll be live from Apple's Worldwide Developer Conference starting at 10am PT today as the iPhone maker lays out its AI strategy. We'll be covering it all at Axios.com.
Today's AI+ is 753 words, a 3-minute read.
1 big thing: What Cisco routers tell us about Nvidia's fate


The fat profit margins that have fueled Nvidia's phenomenal stock price run-up stem from marketplace bottlenecks that are bound to ease.
Why it matters: Nvidia is the AI revolution's bellwether investment, and whenever it starts going down instead of up, the entire AI market is likely to retrench.
Driving the news: Nvidia split its stock 10-for-one Friday after the market closed.
- The value of holdings in the company didn't change, but the price of individual shares will now feel more affordable to smaller investors.
By the numbers: Nvidia's $2 trillion valuation looked bubbly about three months ago.
- The company took 25 years to hit the $1 trillion mark, nine months to hit $2 trillion and a little more than three months to hit $3 trillion.
- At this rate, $4 trillion could arrive in July.
Yes, but: Nvidia's climb is going to get harder and hit limits, whatever fun speculators have along the way.
- The company's strength lies in its chokehold over the most powerful chips needed to train and run today's advanced AI services.
- With both giants like Google and Microsoft and startups like OpenAI, Anthropic and xAI all stockpiling those processors and bidding up their prices, Nvidia can command enormous margins.
That won't last forever. The company's many competitors — including Intel, Qualcomm and Apple — have ramped up their efforts at AI chipmaking in ways that will inevitably fence in Nvidia's growth.
- The best case for Nvidia and its competitors is that AI demand will keep growing at a mad pace, and all these firms will get to slice up an ever-bigger pie. But there's no guarantee of that.
- If big AI providers don't solve the technology's many problems — from inaccuracies and "hallucinations" to uncertain consumer demand and long-term fears of runaway AIs — demand for these specialized chips could soften or vanish, and the AI frenzy would stop dead in its tracks.
Something like that happened 25 years ago. Cisco, which sold the routers every company needed to get online, was the Nvidia of the '90s internet boom. Its stock price chart in the late '90s looks very similar to Nvidia's today.
- From 1998 to 2000, Cisco quintupled in price — and then, from 2000 to 2002, it collapsed, when a hardware glut arrived just as a market downturn dampened demand.
The bottom line: The "sell pickaxes to miners" phase of every tech gold rush is real — but it never lasts that long.
2. Exclusive: Robinhood CEO backs AI math startup
For the past year, Robinhood CEO Vlad Tenev has quietly been helping build Harmonic, a startup building an AI system to solve some of the world's toughest math problems.
Why it matters: As good as many of the latest generative AI models are at language, they tend to be bad at math.
Driving the news: Tenev is leading the funding for Harmonic and serving as its executive chairman, though he doesn't have an operational role at the Palo Alto-based company. Harmonic's CEO is Tudor Achim, the co-founder of Helm.ai.
The big picture: While computers are fast at computations and solving equations, Tenev says they have yet to rival humans when it comes to solving and verifying mathematical proofs.
- The 15-person startup has already achieved an 83% success rate on a key mathematical benchmark, known as F2F, when aided with computer algebra systems.
- The company says it's working to create what it calls "mathematical superintelligence."
- Harmonic sees applications in areas such as aerospace, automotive and medicine, where guaranteed accuracy is essential.
- "As AI starts writing more and more code, we need systems to verify it," Achim tells Axios. "There won't be enough humans."
Fun fact: Tenev got his Stanford degree in math. "Even throughout growing Robinhood, math has always been a particular passion of mine," he says.
3. Training data
- Amid a flurry of concern from security and privacy experts, Microsoft said Friday it would make Windows' forthcoming Recall feature opt-in, and take other steps to better protect the automatic screenshots it takes. (Axios)
- Research from Google shows that AI could help make traffic signals more efficient to reduce the time we spend stopped at red lights. (The Wall Street Journal)
- Autonomy co-founder Michael Lynch was found not guilty in a trial on fraud charges surrounding the company's acquisition by Hewlett-Packard. (Axios)
- One reason that Apple's hydraulic-press "Crush!" ad misfired so badly, according to this writer, is that today's tech giants are "at a loss for something to say about new products." (Axios, The New York Times)
4. + This
Call me old-fashioned, but I'm not ready for an AI beauty pageant.
Thanks to Megan Morrone and Scott Rosenberg for editing this newsletter and to Caitlin Wolper for copy editing it.
Sign up for Axios AI+



/2024/06/10/1717984928093.gif?w=3840)