We got the inside scoop on Sequoia's @Alfred_Lin & @gradypb era for Bloomberg Businessweek.
Bigger checks. Evolving fund structures. New pools of capital. And an intense focus on speed and rigor across the partnership as the next wave of tech becomes decidedly more physical.
Latest w/ @RebeccaTorrenc5 & @EdLudlowbloomberg.com/news/articles/…
Sources: Modal Labs is in funding talks at a ~$15B valuation, up from $4.65B in May; Baseten is in funding talks at a $26B valuation, up from $13B in June (Bloomberg)
(Visit Techmeme dot com for the link and full context!)
Source: 1789 Capital, where Donald Trump Jr. is a partner, is in talks to raise $3B for its second growth fund and has already raised $2B of that target amount (@rebeccatorrenc5 / Bloomberg)
(Visit Techmeme dot com for the link and full context!)
The AI inference rush is real & rich.
Scoop w/ @dinabass on Modal Labs in talks to raise a new round of financing at a $15 billion valuation, as Baseten also discusses a new round
bloomberg.com/news/articles/…
With its second growth fund, 1789 wants to make fewer, bigger bets, and lead more rounds. Case in point: it's leading Polymarket's new ~$1B round with a ~$300M check.
More here: bloomberg.com/news/articles/…
New venture funds from Bessemer. The firm is shifting to growth, with the vast majority of the new funds (even the one reserved traditionally for early-stage investments) allocated toward later-stage rounds
Scoop here:
bloomberg.com/news/articles/…
NEWS: Bessemer is raising $5.75 billion in fresh funding to support a broader shift towards more growth stage investments, reports @nmasc_@spdholakia joins me on Bloomberg Tech later today
bloomberg.com/news/articles/…
An adjacent observation: startup rivalries are emerging overnight and heat up super fast (which is a fun tension for journalists to dig into!)
If your competitor is raising, you probably are too. Customer bake offs are monthly. And takeover interest is constant/widespread
I’ve been in venture for ~6 years (incl 2021!) and I have never experienced this before. Every week I see 8-10 qualified opportunities that are absolutely ripping. And every week each one gets 2-3 verbals in a couple of days.
But unlike 2021, these companies seem to be working! People are no longer underwriting stars or community member growth or other dumb 2021 metrics. It’s actual cash paid by actual people at actual companies (in the 99% of non fraudulent companies).
Every startup is thus constantly competing against every other startup raising that week for who was most interesting enough to get diligence’d that week. Actually insane times.
So much of the AI debate treats the frontier like a guaranteed OpenAI/Anthropic duopoly. That's not necessarily the case.
Mirendil/SSI etc. are useful reminders that the next leap may come from a lab DC has barely heard of -- and that policy has to work regardless.
Pacing where? Latest scoop is about a young startup raising capital at a $5 billion valuation for its efforts to build self-improving AI.
Mirendil, which was launched by former Anthropic staffers and has hired over 20 employees, is set to launch its first frontier model by early next year.
w/ @shiringhaffary
Pacing where? Latest scoop is about a young startup raising capital at a $5 billion valuation for its efforts to build self-improving AI.
Mirendil, which was launched by former Anthropic staffers and has hired over 20 employees, is set to launch its first frontier model by early next year.
w/ @shiringhaffary
Some startups, like Harvey, Abridge, Ramp, and Rogo, are embracing open-weight models or training their own models to reduce expensive reliance on frontier labs (Bloomberg)
(Visit Techmeme dot com for the link and full context!)
Loving all the new coffee shops launches in San Francisco😍😍
- Soon & Soon
- Wildfox
- Les Mils
- Side Characters
- Origin Lab
- Kissaten Hifi
- Always Awake
Harvey’s AI costs got so high its gross margins plunged from 50% to -50% in 6 months.
Its response: build its own model using open-weight AI.
Startups from Abridge to Rogo are following suit to cut costs and reduce their reliance on OpenAI and Anthropic. My latest w/ @nmasc_👇
The number one topic coming up at Silicon Valley dinners lately: the sudden, serious rush among AI startups to build their own models. It's a push to cut costs, reduce dependence on the big model providers, and prioritize control.
My latest w/ @RebeccaTorrenc5bloomberg.com/news/articles/…
I've been reporting on a few startups doing this for a quite a while now. Cursor, for example, is an early example of a company building its own models to manage costs of routing to the expense frontier. The shift now is driven by the rise of cheap and effective open source models.
From Moonshot's growing forward-deployed engineers program to Harvey margins, there's more in our scoopy piece on @technologybloomberg.com/news/articles/…
What a reminder of how much venture math has changed: a $700 million seed round is underway in the UK.
I've been noticing less spinouts from the bigger firms this year and I wonder if this is one reason why: people need to work at firms with deeper pockets in order to participate in some of these extraordinary rounds. Of course early checks, first tranches, etc will still get you in, but competition and costs are steep & this is just the latest example of it
A UK artificial intelligence startup founded by former Google DeepMind researchers is in talks to raise $700 million in a seed round bloomberg.com/news/articles/…