Cursor's (maybe) sale

Cursor in Talks to Raise $2B…and Maybe Get Acquired

All eyes are on Cursor this week. The AI coding startup is in talks to raise a $2 billion funding round at a valuation exceeding $50 billion, according to CNBC. And just yesterday, SpaceX said it’s obtained rights to either acquire Cursor for $60 billion later this year or pay $10 billion for their work together.

The funding: 

• Andreessen Horowitz is expected to co-lead the round, with Nvidia and Thrive Capital also participating. All three are existing investors.

• This would mark Cursor's third major raise in under a year. The company closed a $2.3 billion round at a $29.3 billion valuation in November, which itself came after a $900 million round in June. Other backers include Accel, Coatue, DST Global, and Google.

Teaming with SpaceX: SpaceX claims that combining Cursor’s product with SpaceX’s supercomputer “will allow us to build the world’s most useful models.” The two companies have become more closely intertwined recently. Two senior engineers at Cursor left last month to join xAI (which SpaceX acquired), and xAI was reportedly planning to supply computing power from its data centers to Cursor.

Could Google Build New AI Chips With Marvell?

Google is in discussions with chip designer Marvell Technology to develop two new chips aimed at running AI inference workloads more efficiently, according to The Information.

What's being built: One chip would be a memory processing unit designed to work alongside Google's existing tensor processing units. The second would be a new TPU tailored specifically for inference. Google plans to complete these chip designs by next year before moving to test production.

The breakdown: No contract has been signed yet, but if or when there’s a deal, it would make Marvell Google's third chip design partner. The others are Broadcom, which handles high-performance TPU variants, and MediaTek, which builds versions of chips at 20–30% lower cost. 

44% of Songs Uploaded to Deezer Are AI-Generated

Streaming platform Deezer revealed this week that nearly half of all songs uploaded to its platform daily are now AI-generated—about 75,000 tracks per day, up from just 10,000 when it first launched its AI detection tool in January 2025.

The numbers:

• 44% of all daily uploads are AI-generated. Despite the volume, AI tracks account for only 1-3% of total streams on the platform.

• 85% of those AI-generated streams are flagged as fraudulent (likely bots gaming royalty payouts) and stripped of monetization entirely.

• Deezer has flagged more than 13.4 million AI tracks in its catalog so far this year.

The Backend Other Builders Forgot

You built an app with AI. Now you need a database, auth, hosting—and suddenly you're stitching together four services with no idea who else is on the same server.

OptiDev ships what others leave out:

• Your own database: managed, not shared, not bolted on

• Single-IP whitelisting: connect your databases securely

• Encrypted & isolated: your data never touches another user's

• Auth & edge functions: login, roles, serverless—all included

• Battle-tested: built by a team running 200,000+ screens

Other builders give you a frontend and wish you luck. OptiDev gives you an enterprise-grade backend your data actually deserves.

AI workspace for one-person companies

Floatboat learns your editing, decision, and execution habits then turns them into reusable "Combo Skills"—automations that act like a personal AI team. It connects to 3,500+ tools and includes built-in browser automation and desktop integrations.

How you can use it

• Turn your repeatable workflows into automations without stitching together multiple apps

• Drag context directly into an agent and invoke native tools like email and calendar

• Package your playbook so it runs while you're offline

• Scale your output without hiring

Pricing: Paid

AI-driven in-app onboarding

Frigade autonomously learns your product by using it, then delivers contextual onboarding flows, guided tours, and usage insights inside your UI—without manual documentation or engineering overhead.

How you can use it

• Reduce support tickets with personalized, in-context user guidance

• Launch feature tours that adapt to each user's behavior

• Boost feature adoption and retention with zero-setup flows

• Escalate complex cases to humans automatically when needed

Pricing: Paid