How to lose $7.5M a day for more than a year
Welcome to Cautious Optimism, a newsletter on tech, business, and power.
Happy Wednesday! Today’s mandatory reading is this piece from Sahil regarding his time at DOGE — and how it ended. Reading between the lines, he got canned for talking. Which is ironic from a “maximally transparent” organization, but what can you do. It’s 2025 and no one has a stable gig.
Which we’ll get to down below. In the meantime, today’s big event is Nvidia earnings after the bell. We’ll touch on Box and Okta earnings today, too. To work! — Alex
📈 Trending Up: Mistral’s agentic dreams … Joby Aviation, after Toyota money arrives as expected … burdens on app stores … corporate filmaganda … we’re number one! … government spending …
📉 Trending Down: Reasons for Southwest’s continued existence … not paying the vig … the price of war … Texas …
How to lose $7.5M a day for more than a year
Yesterday we took a look at the Salesforce-Informatica deal, poking at its economics and potential implications for SFDC’s newly expanded corporate arsenal.
Missing from our coverage was a reminder that this was round two. Last year, Salesforce took a run at Informatica for around $11 billion. However, the deal didn’t get done — perhaps because traders had bid Informatica shares above its potential exit price, thus making it a hard deal to consummate at a price that Salesforce was comfortable with.
Fast forward about 400 days, and Informatica exits to Salesforce for $8 billion. That works out to about $7.5 million worth of lost value per day since the deal fell apart. Thus the phrase pigs get fat, hogs get slaughtered.
I volunteer for the brain computer
Semafor reports that Musk’s Neuralink has raised $600 million at a $9 billion valuation, greatly expanding its worth since it was valued at around $3.5 billion in 2023. As far as funding rounds go, it’s a doozy.
Putting aside the company’s ownership structure, the results from its brain-computer interface in terms of helping those unable to move to communicate with digital machines are astonishing. I would volunteer, but as I’m fortunate enough to have full control of my limbs, others have need over my personal curiosity. Still, I intend to get my own brain-computer chip when they are commercially available.
I also want to go to space before I die, for whatever that’s worth.
I think it’s worth hitting pause for a moment and taking stock of technologies that were long science fiction, but are today quickly becoming a reality:
Self-driving cars, thanks to Waymo’s commercial growth and market expansion (and a host of Chinese companies like WeRide and Pony AI).
Implantable connections between human brains and computers (Neuralink).
Flying cars, thanks to improving eVTOL machines that are racing towards commercialization.
Synthetic intelligence, thanks to the genAI boom. We’re not yet at HAL, or Her levels, but we’re getting close — and with the pace of AI development staying pretty damn hot, it’s hard to not be pretty stoked about what we’re going to be able to do in the coming years.
Incremental improvements to nuclear power, and regular advances towards commercial fusion power.
A rapidly expanding and maturing in-orbit economy, a critical step to becoming a multi-planet species.
There’s a lot of bad stuff going on. Tariffs are bad. Hot wars are bad. Climate-related disasters are costing more and more. Insurance is a mess. You can add your own downbeat news to the list. But at the same time we’re still doing incredibly cool stuff as as species, and I take heart in it.
Box, Okta and diverging earnings fates
Shares of enterprise storage and productivity software company Box are sharply higher today after earnings, while shares of identity-focused Okta are down sharply. What happened? Let’s find out.
Box
After guiding to revenues between $274M and $275M, Box turned in $276.3M in total top line.
Current-quarter guidance indicates revenue of $290M to $291M, right on top or just ahead of analyst expectations.
Raised guidance of $1.165B to $1.170B for its current fiscal year, up from guidance of $1.155B to $1.160 billion stated in its preceding earnings cycle.
All that’s lovely, but Box shares are up nearly 18% today. What’s driving that gain? Box CEO Aaron Levie waxed about the company’s AI-driven sales momentum (transcript source):
RPO grew 21% year over year, and we saw strong outperformance on billings. [...] Following the launch of Enterprise Advanced late in Q4, in Q1, we saw strong momentum in customer adoption as enterprises look to Box to help them transform their AI-driven workflows around content. […]
For years, the value of [corporate data] has been limited to only what humans can do with this information. Which means most companies have never been able to truly tap into the full value of this information. But in a world of AI and AI agents, this finally becomes possible. […]
We unveiled all new capabilities to support AI agents that can do deep research, search, and enhance data extraction on content securely in Box. […] None of this would have been possible even a year ago. But with the cost of AI inference dropping, context windows expanding to support larger datasets, reasoning models handling much more complex tasks, and better understanding of designing agentic workflows this all of a sudden becomes possible. […]
Even after the large number of early adopters we saw in Q4, bringing on enterprise advanced, we were pleased to see the continued growth in deals in Q1. […] Additionally, [AI is driving] pricing improvements, in our target 20% to 40% increase range for these enterprise advanced deals.
To sum, after posting piddling 4-5% growth in its most recent quarter — sorry, Aaron — Box has some material AI tailwinds behind that are helping it chart a faster-growing future.
Box expects 8% growth in its current quarter, and 7% in the year. Those are better than its trailing quarter’s performance, and as companies tend to sandbag, we can infer that Box exiting the calendar year at around a 10% growth rate is not impossible. Rebound!
To wit, remaining performance obligations (RPO) at Box were up 27% year-over-year in its most recent quarter, while billings scaled 27% as well. If you were looking for a reason to bet on Box, the work that its staff have done to pivot the company from trad EFSS to something more AI-native is to be applauded.
Okta
After guiding for revenue of $678M to $680M for its most recent quarter, Okta turned in $688M worth of total revenue, good for growth of 12%.
RPO grew 21% to $4.084B, and Okta’s operating and free cash flow both came in at 35% of revenue for the quarter, a brilliant figure implying material cost control.
So, why did investors send its shares around 13% lower today? Guidance of growth that decelerates to 10% in the current quarter, and 9% to 10% for the full year. That’s a steady decline.
Now Okta states it is taking a “prudent approach to forward guidance” inclusive of “potential risks related to the uncertain economic environment for the remainder of FY26,” but that doesn’t mean that the street is happy to see Okta’s cash generation machine decelerate.
It’s 2025, and in an AI world investors want acceleration, and nothing less.
Pour one out for the kids
Continuing our coverage of the changing job market for early-career professionals, we have more bad news. Recall that CO recently took a look at the relationship between interest rates and Mag7 hiring. The gist was that as rates began to rise, employment growth at those companies cratered to effectively zero.
Layoffs have continued since. Now, data from SignalFire — a venture capital firm with a fresh, new fund — indicates that the job developer job market is not only becoming more competitive, but for less experienced SWEs it’s becoming lethal.
Observe:
Notably despite the massive decline in new-grad hiring from the largest tech shops, the Big Dogs remain slightly better places to land a gig if you are green. Perhaps not for long, however. From the same SignalFire dataset:
The above shifts are harming job prospects for new CS grads. One more chart from our friends at SignalFire:
It seems that not only are the largest companies cutting back on new grad developer hires, they are doing so with gusto. From 25% of top grads going to Mag7 gigs in 2022 to 11% and 12% in recent years.
That’s good news for startups, if they want to pick up junior talent on the cheap. But since it seems that every company wants more-senior talent, you have to wonder what the kids are going to do.
The idea that AI is taking all the jobs is overblown, at least for now. Despite the success of Harvey and other vertical AI startups targeting the legal profession, post-graduate employment per the American Bar Association actually improved last year compared to 2023. But technology eats its own cooking first, and other industries second. I remain concerned that the race for senior talent at the expense of earlier-career, and more automatable employment is going to be an issue in years to come.