Uber is getting back into the self-driving game
And: Swedish AI unicorns, French foundation AI model concerns of global import, and rising AI investment overall? Progress.
Welcome to Cautious Optimism, a newsletter on tech, business, and power.
Good news! The United States economy is holding onto momentum despite trade turbulence. Retail sales rose 0.6% in June after dropping “almost a full percentage point in May.” That’s welcome news, given how important the American consumer is to the global economy. Adding to that good news, jobless claims fell for the fifth consecutive week, indicating that the labor market is hanging in there.
Durable economic data and inflation that’s creeping higher? Sounds like a recipe for holding interest rates steady. POTUS wants them cut by two-thirds. This situation is precisely why having an independent Fed leader matters. And if you are sympathetic to the current POTUS having more say, just ask yourself how much you would appreciate a MMT-favoring future president calling the shots at the central bank!
— Alex
📈 Trending Up: ICE arrests of non-criminals … tax breaks for investors … profits, at TSMC … cheating … automated mining (really!) … Via, which is going public … flash flooding … battles over education funding … Russian threats against the West … silly trade deals …
📉 Trending Down: OpenAI-Microsoft … Fed independence … venture returns in the blitzhiring era … public broadcasting … Axel Springer … house prices in Florida … housing demand in Austin … water access in the UK …
Over in AI-land, CO is keeping an eye on OpenRouter’s LLM usage data. The startup, which allows customers to hot-swap models, tracks overall token generation on its platform, giving us a great picture of the changing state of developer demand for one model or another. It’s the next thing to look at after parsing a new SOTA model’s benchmark results.
And what do we see in the wake of Grok 4 (USA, xAI) and the Kimi-K2 (China, Moonshot AI) model families coming to market? Today Kimi-K2 takes two spots (10th and 20th), while Grok 4 is in 12th. In aggregate, Moonshot has 2.5% of OpenRouter token share this week, while xAI has secured 2.3%. Mistral is above both at 3.3%. Speaking of which:
Can your AI speak French?
Over in Europe, funding to “AI-native” startups rose to €3.04 billion in the first half of 2025 — up 61% from the same period of 2024 according to Sifted. That’s a fat gain in percentage terms, even if the total Euro sum is modest.
Scanning for other European AI signals, Lovable is raising capital in Sweden like it’s a pre-product AI shop in the Valley with at least one founder whose name is on the Attention Is All You Need paper. Yes, the vibe-coding and app-building Lovable is now worth $1.8 billion after stuffing $200 million into its accounts.
With just 45 people and $75 million worth of ARR, I doubt that Lovable needs $200 million worth of cash. But, if you are going to raise, your valuation becomes anchored to your revenue and pace of revenue growth. From there, your lead investor will want to buy enough of your company to meaningfully participate in its upside.
So, if your ARR multiple is 24 — a number we’re seeing more and more in AI-land — and your revenue base is above the mid-point of eight-figures, well, you’re going to raise nine.
Elsewhere in European AI progress, Mistral’s latest AI model — Voxtral, which focuses on speech — is garnering good reviews. This particular shot in the arm of product momentum is likely welcome, given that France’s AI leader is scoping out a $1 billion fundraise.
Swedish AI unicorns, French foundation AI model concerns of global import, and rising AI investment overall? Progress.
Uber is getting back into the self-driving game
Ages ago when it seemed that self-driving was just around the corner, Uber spent heavily on its own technology. After both tragedy and a far longer march to market usefulness than many anticipated, Uber wound up selling its self-driving efforts to Aurora, taking an equity stake in the latter company as part of the transaction.
Since that 2020 divestment, Uber has partnered with a host of self-driving companies to bring the technology to its platform. The trade is simple at surface-level: Uber has demand, self-driving companies have supply, and if the two are brought together everyone can make money.
Uber teamed up with companies from around the world, eventually building up a list of friends that included most self-driving names you know offhand:
Despite its work to link hands with so many companies, Uber announced a new plan today to form a partnership with Lucid (EVs) and Nuro (self-driving technology) to build a “fully integrated robotaxi experience developed for comfort, safety, and scale.”
The new service is coming to a “major US city” next year, Uber told investors.
The plans are not small. Uber intends to “deploy 20,000 or more Lucid vehicles equipped with the Nuro Driver over six years,” some of which the company will own with others managed by “third-party fleet partners.” Naturally, the new robotaxi service will be exclusive to Uber’s platform.
Uber intends to execute “multi-hundred-million dollar investments in both Nuro and Lucid,” it added. Shares of Uber stock flat following the news. Shares of Lucid are up 26% in early trading. Shares of Nuro are unmoved, because it’s not public. But as Nuro raised fresh capital just a few months ago at a $6 billion valuation, you can expect that its backers are about to mark their shares higher.
Why would Uber go back to working on its own self-driving service instead of sticking to the partnership game? Because, I reckon, the robotaxi companies that are currently Uber partners want to own the customer in time, just as Uber doesn’t want to share its own customer base with a third-party that would expect to absorb some of the available margin.
Marriages of convenience don’t last forever, in other words.
And Uber is richer than it has ever been. It can afford to put hundreds of millions into its new partners and spend to get to get its own self-driving fleet into the market. For those of us still living in cities with organic drivers — fleshbots, meatsacks, call us what you will — seeing Uber jump into the robotaxi market is great news. More competition, more good, and with Tesla and Waymo already locked in slow-motion but aggressive competition, Uber joining the fray will only double the fun.
Maybe your neighbor shouldn’t be a venture capitalist?
Generally speaking I am in favor of investments that are available to some becoming available to all. Lower barriers to investing is good. I recall making my first stock trade as a tot, buying a few shares of Cray Supercomputer back when I was borrowing a parental laptop and logging on early so that I could catch the market open on the West Coast. I think the fee I paid to make that trade was a double-digit percentage of the total buy, because I was a broke kiddo, and eTrade or whomever I used at the time charged about one burrito to execute the transaction.
Since then, things have gotten better for regular folks. You can access zero-cost trading services, invest in zero-cost index funds, or explore an exploding market of niche ETFs that allow the retail crowd to make sectoral, leveraged, or inverse bets on the market. Again, good.
I am also in favor of equity crowdfunding when done ethically. I begin to get a little leery, however, when the most opaque and illiquid investment vehicles work to drum of retail interest, however. Why? Because opacity and illiquidity are hard to explain to many buyers until something goes wrong, and they find themselves stuck.
Onto the news: After opening the door during his first administration, POTUS is weighing an executive order in his second that would allow for regular folks to invest in private equity funds via their 401ks. Details are scarce, but the profit motive is pretty simple: PE firms want more investment, and 401ks hold trillions.
There’s good logic to the concept of allowing more people to invest in private companies, and firms that do the same: The number of public companies is falling, and breakout startups in today’s market often stay private until a far greater percentage of their future worth is captured by private backers before they do list — if they ever do. So, if you want to get in on the upside of the majority of corporate America, you need to look at companies that have yet to go public.
Cool. And as noted above, democratization of investing optionality is a good thing. Provided that 401k investments in PE vehicles is managed in some way that prevents the average person from accidentally paying insane fees to a firm that generates sub-market returns while holding their principal hostage, this could all work out.
But I worry. Investing in private companies is incredibly high risk — Joke: Want to have $1 million? Start with $5 million, and cut a bunch of angel checks — with fees far in excess of what you can get from index funds, barriers to exiting a position, and lots of other chaff like continuation funds, strip sales, and the like. Just as I would not recommend that my most middle-ground friend start trading options, I would worry if they started to play venture capitalist or private equity denizen with their retirement money. That’s not ‘money you can afford to lose’ or whatever other warning we tend to append to private-market dealmaking. It’s the opposite.
So long as the move is made with cautious; so long as we’re not about to transfer stored retirement wealth into the hands of PE giants, this could work out. Could, and I am stretching to say so.
What I think we need is something between startup and public-company status. Something that would exhale more operational results than we get from unicorns today, but not so much that they have to spend all the required coin to list. At that point, private-market opacity would drop, and I think I’d rest easier knowing that the same folks who may put down their post-tax earnings on scratch-offs are not about to get rekt by an industry that they do not understand.
More if and when we actually get the details, but tread carefully here.
I’m talking my book here: Most of our money is in zero-cost index funds, with a sliver in private companies and exotics (Crunchbase stock from my time there, a single angel check, and a dusting of bitcoin ETF). But really it’s just in the stock market writ large. So far, so good!