PostHog vs. The Industry
Welcome to Cautious Optimism, a newsletter on tech, business, and power.
Due to a childcare emergency — every working parent’s favorite thing — the newsletter is both late today, and brief. I promise to make it up to you by writing even more than usual tomorrow. Hugs — Alex
📈 Trending Up: Gross … yuck … ew … inevitability … AI usage … self-driving … venture-backed weapons … escalation risk in Iran? …
📉 Trending Down: Millennial finances … shock … Intel’s human capital … consumer spending …
PostHog vs. The Industry
PostHog is a startup that is doing everything wrong, and getting it all right in the process. The latter claim is simple to back up: Stripe just led a $70 million round into the company at a $920 million valuation. Given that Stripe isn’t technically a venture capital firm, that’s signal.
PostHog builds digital tools for product work, including analytics, experimentation services, data management products, session replay, and more.
Apart from a somewhat exotic late-stage lead investor, why do we care about PostHog? Because it’s breaking all the normal rules of software startups. A peek at its Why page details a company taking a contrarian approach to slinging code:
Its products are (mostly) open-source; the company also offers free tiers of its hosted service. That means developers can use its services for free (self-hosted) or free (generous free hosted access) before needing to put up a credit card.
Protection for overages in case something you build “goes viral by accident.”
Price cuts.
No outbound sales, no goal of selling the company, and PostHog claims to not “screw” customers with “contract terms like auto-renewal.”
Wild. Pretty much all the SaaS tricks that private equity loves to enact once it gets its hands on a software company? What if you did the very opposite?
Elsewhere in software-land, Salesforce is enacting price increases for “Enterprise Editions and Unlimited Editions for Sales Cloud, Service Cloud, Field Service, and select Industries Clouds” of “an average of 6%.” SFDC couches the price increases under the banner of we made the products better, so we’re charging more for them.
Which, fair enough. Is anyone going to drop Salesforce over a single price bump? Probably not, but the move by Salesforce to juice its customer base for another year’s single-digit ARR growth is evidence of just how much room the market might have for more startups like PostHog. Whatever the majors are doing? What if you just did the opposite?
C.R.E.A.M
There’s a lot of money in flight today. Spotify’s Daniel Ek is leading a “€600mn funding round in Helsing, valuing the German defence tech group at €12bn and making it one of Europe’s most valuable start-ups” the FT reports. That’s an enormous sum for a European defencetech company; you can compare Helsing to Anduril in that both are richly-funded upstart defense companies with a focus on AI and smaller weapons systems.
It’s good that both the EU and the United States are seeing a renaiisance of weapons startups. No, war is not good, but losing wars by not prepping for the future of conflict is the way liberal democracy and free market systems get eaten. So, teeth are required.
Elsewhere in flying cash, Bloomberg has a notable set of datapoints on how much money xAI — now parent company of X, FKA Twitter — has spent to date. Documents viewed by the financial reporting powerhouse indicate that before its currently planned debt sale of around $5 billion Musk’s AI foundation model AI company had raised $14 billion. Of which it had $4 billion left.
We know that xAI has spent heavily to build out its compute capacity, and staff up to build competitive models. Perhaps it’s a little surprising that a full $10 billion has been put to work, but what did we expect, really from one of the most expensive startup propositions in the history of the world?
OpenAI lost $5 billion last year against $3.7 billion worth of top-line, for reference.
xAI’s cash needs are large enough that it might sell another $4.3 billion worth of shares along with its hoped-for debt raise. The two at full-fat would gross the upstart another $9.3 billion and rebuilds its coffers to the ceilings.
Sticking to the theme of winged currency, the days of cheap coding tools could be ending. Anysphere, the company behind the Cursor coding service is moving upmarket. Here’s TechCrunch:
Anysphere launched a new $200-a-month subscription plan for its popular AI coding tool, Cursor, the company announced in a blog post on Monday.
The new plan, Ultra, offers users 20x more usage on AI models from OpenAI, Anthropic, Google DeepMind, and xAI compared to the company’s $20-a-month subscription plan, Pro. Anysphere also says Cursor users on the Ultra plan will get priority access to new features.
20x usage for 10x the price is a gamble that folks won’t exhaust their total compute supply, else the economics are back-of-the-envelope wonky. Given the cost of building (training) and running (inference) cutting-edge AIO models, did we really think that developers were going to get away with paying $20 per month for leading solutions?
Probably not. Recall also that Cursor recently stormed the all-time growth charts by scaling to $500 million ARR in what could be record time. More expensive tiers are a good way to double that figure. So long as Windsurf and others don’t hoist the Jolly Roger and refuse better operating margins in favor of low prices.
Closing, the fine folks over at Ramp just raised a $200 million round led by Founders Fund at a $16 billion valuation; that’s up $3 billion in just a few months. When considering Ramp’s success to date, it’s worth recalling that when the startup got started it was gently mocked as a Brex also-ran. Well, now it’s worth more than Brex, and by some counts handles more payment volume than its rival. How about those enterprise spend apples?
Remember when companies would go public when worth $1 billion or less? Wild, yeah? I doubt that we’ll see Helsing, xAI, or Ramp go public before 2027. At which point they will be eligible for inclusion in the S&P500 (minimum market cap as of January 2025? $20.5 billion). I
The old bar of $100 million worth of trailing revenue and rising profits is as dated as backing a biplane factory. But I did not expect the unicorn era to hold most of its winners back on the private markets until they were ready to oneshot the S&P.