And: OpenAI's delayed IPO, worker discontent, and Chinese AI! #
- Welcome to *. Cautious Optimism, a newsletter on tech, business, and power. Modestly upbeat
Friday. Rest in peace, Om Malik, the journalist and investor who I first met in the pages of Business 2.0 (a magazine that helped spark my interest in startups). I only got to know Om a little, but his reputation for kindness and intelligence speaks for itself. A real loss. Love to his family.
Stocks fell overnight around the world, especially in South Korea, China and Japan. The Strait of Hormuz is both open and closed, and oil prices are easing somewhat as tensions (again) slowly ratchet lower. And, Gina Raimundo, the Pride of Rhode Island, is working with a host of tech companies to “support workers as the AI boom reshapes the US economy.” We’ll keep an eye on the effort, which has attracted support from several AI labs.
Today, we’re talking about the California billionaire tax, the trench of distance between workers and managers that AI is creating, and U.S.-China relations as it relates to alleged theft of technology. To work! — Alex
P.S.:* Yesterday,*, writtenCO published a look atDatabricks’ growth and gross margin pressures as it accelerates growthin concert withFastForward. Enjoy!
📈 Trending Up #
Public corruption…the economic viability of AI…touch-screen Macs? …interest rates…chip mergers…AI-native Gusto competitors? …ass-covering…Google’s monopoly power in search…bridge idiocy
Taxing the rich: After collecting enough signatures to make the ballot, backers of California’s one-time, 5% tax on billionaires declined to retract the initiative by the Thursday evening deadline. As a result, and to the chagrin of the 0.1%, California will vote on the matter this November. The state’s governor is publicly opposed to the tax; the group proposing the tax even offered a compromise in the form of 2% levy that Gavin Newsom rejected.
Polling from May indicates that 54% of Californians are in favor of the 5% tax: 76% of Democrats, 53% of independents, and 18% of Republicans. Younger voters support the fee two-to-one, while 71% of renters back the concept. Efforts to delay or block the tax have thus far failed; now, it’s up to the ballot box.
Notably, Governor Newsom wrote a post on Substack explaining why he’s voting no. He brings up what I consider to be a critical point in the argument that the tax is poor way to generate revenue for the state:
[T]he fight to make the wealthiest Americans pay more in taxes is not one we should be fighting state by state. You may not be able to pick up and move to Texas or Florida to shelter your income from taxation, but I promise you that billionaires can, and do. Wealth is movable, and it shops for the state with the lowest taxes. The fight belongs at the federal level, where this broken system was created in the first place.
Regular folk don’t understand how mobile the ultra-wealthy are. These people already have homes in several countries, let alone states, and can move their tax-paying domiciles with moderate ease. California’s voters may think that they are set to give local billionaires a haircut, but they will likely wind up with far fewer heads to shave than they thought. The billionaires will simply fly into California when needed, and conserve their wealth and tax dollars to other states they find friendlier.
Newsom argues that only a national wealth tax can resolve budget issues, because smaller states with lower revenue needs can offer minimum tax rates, which effectively limits other states from raising taxes. Our views on the idea aside, a national wealth tax is a non-starter in Washington today. So, we get to watch California voters decide what they want to do. Some billionaires have already relocated, and more will follow. Call it a natural experiment, albeit one I think we can see the results already taking form.
Government-approved AI: OpenAI intends to roll out its forthcoming GPT-5.6 model to select partners first, marking a sea change in how AI is created and deployed, per Politico:
Details about the latest model, GPT-5.6, have not yet been released publicly. The company did not initially plan to restrict the release of the general-use model, but changed course at the White House’s request and in consultation with the Office of Science and Technology Policy and the Office of the National Cyber Director, the two people said, granted anonymity to describe internal discussions about OpenAI’s latest product release.
Oof. Instead of clear rules passed by Congress, we have the executive branch taking control of AI releases. It’s among the worst outcomes possible: obfuscated rules, backroom discussions, and, I presume, political favoritism deciding the future of key technology.
We’ve entered a new era. Anthropic’s latest and greatest models were borked by the federal government. OpenAI is releasing its cybersecurity-focused GPT-5.5-Cyber to select partners. And now the government is taking charge of how more general models are released.
Politico adds that OpenAI’s CEO Sam Altman said “the controlled release was not [his preferred] method for announcing its latest AI model, the company was hopeful it could work with the government to develop a more ‘sustainable’ approach to unveiling its iterations of models going forward.”
- This is the bad timeline.
[📉](https://finance.yahoo.com/news/servicenow-pledges-1-5bn-investment-110000403.html) Trending Down
[📉](https://finance.yahoo.com/news/servicenow-pledges-1-5bn-investment-110000403.html)
Blocking teens from social media…Chinese EV profits? …Airwallex(despite this!) …SoftBank’s stock
OpenAI’s 2026 IPO dreams: Back when unicorns were springing out of thin air, startups would grant their investors special privileges to ensure they got a $1 billion valuation in their next funding round. Being called a “unicorn” an imprimatur of success — this startup just made it. Just how going public is a way to tell potential customers and investors that a company’s here to stay, a billion-dollar private market valuation once cast a similar glow on startups that cleared the threshold.
One billion dollars is old hat. Today, the game is one *trillion *dollars or bust. I’m not being wry: the New York Times reports that OpenAI is considering delaying its IPO to 2027 as it chases a listing at a valuation of at least $1 trillion.
Why? Because people love round numbers. More importantly, the most important tech companies are worth a trilly or more. If OpenAI goes public at $900 billion, doesn’t that make it a pipsqueak?
Almost. There are more than 5,000 public companies worth at least a billion dollars in the world, while CompaniesMarketCap counts a mere 16 worth $1 trillion or more. Taiwan has one (TSMC), South Korea has two (Samsung, SK Hynix), and Saudi Arabia has one as well (Saudi Aramco). The United States houses the rest.
What’s worse, OpenAI’s rival AI labs are already worth $1 trillion (SpaceX, which acquired xAI), or set to be (Anthropic). Sam doesn’t want bronze.
- There’s a more prosaic reason for OpenAI to want the $1,000 billion price tag: Rewarding its investors with greater gains. The dollar figure may be artificial, but the higher the price OpenAI can command at IPO, the more money its most recent backers will make.
But wait, why might Anthropic clear that threshold and OpenAI struggle to match it? Anthropic’s run-rate revenue is reportedly about twice its rival’s. It’s easier to get a revenue multiple across the private-public transom if you have a larger revenue base.
Still, I didn’t think unicorn-era financial engineering was this pertinent today. Surely $900 billion is 90% as good as $1 trillion?
Workers’ faith in their leaders: A recent Glassdoor report shows a rapid decline in workers’ views of their leadership.
Here’s the chart you need:
Tech workers give their senior leadership a similar ~3.45 rating, down from ~4 in 2020. What’s driving discontent? Glassdoor notes that mentions of layoffs and inflation in worker reviews grew 25% and 34% compared to a year ago. Layoffs are being talked about more than ever, while inflation mentions are off a 2022 peak.
Similarly, “burnout” and “AI” were mentioned 43% and 240% more than last year, respectively, and worker views of artificial analysis are “turning negative:”
Workers are concerned by leaders demanding workers use AI while simultaneously touting AI as a reason for layoffs and reduced hiring. As recently as last year, the majority of discussions of AI on Glassdoor were positive (55% positive compared to 41% negative). That sentiment has taken a turn for the worse and flipped in 2026—the majority of discussions about AI on Glassdoor are now negative (53% of reviews are negative compared to 43% positive).
It’s not surprising. Companies are asking more from staff despite ongoing layoffs and AI deployments. Concern about being replaced or burnout from rising productivity expectations appear to be driving a wave of discontent. Depending on your politics, the fact that corporate profits are worth a record portion of GDP may be pertinent, or not.
- Why do we care? Apart from the fact that people’s happiness matters, AI is becoming a real ballot-box issue. If employees are angry with AI being rolled out everywhere — fairly or not — it could impact how they vote, what regulations are put into effect, and more. - Yelling at people to have more kids doesn’t raise fertility rates. Similarly, the endless river of social media posts attacking those unhappy with the economy for being ungrateful won’t encourage people to vote for the technology causing them trouble.
Here’s a headline for you: Elastic stretches workforce 7% thinner as AI does more of the heavy lifting
Can China compete in the AI race without cheating? #
Subscribe to Cautious Optimism to unlock the rest.
Become a paying subscriber of Cautious Optimism to get access to this post and other subscriber-only content.