brought to you with
✦
Happy Friday, builders. 🕶️
May your weekend budget be bigger than your AI budget.
Echodyne raised about $200 million over 12 years, stayed profitable and just agreed to sell for up to $5.3 billion. Apple brought in $114.5 billion in one quarter, then started trimming how much AI some engineers can use each day. One is a 400-person radar maker and the other is Apple, but they share a habit: they know what everything costs. Neither waited for a bad month to start counting. So while the AI money is flowing, who at your company is reading the bill?
In today's newsletter, we'll get into:
The drone-radar maker that sold for up to $5.3B
Apple's engineers go on an AI allowance
A yogurt brand's AI-made twin
Humanoid robots for $8,500 a month
✦
The founder’s dashboard / Your quick roadmap
✦
FOUNDER BOARD
📬 Same email, two matched groups of ~2,300: SaaStr's challenger sender got 2.4x the opens and 2.9x the clicks. Run a matched split before you renew yours.
🤖 AI assistants bring SaaStr 0.5% of visitors, converting 1.5x better than search. Track that channel in GA4, then check if ChatGPT, Claude and Gemini rank you top 3.
🎯 Martin Tobias got 4,500 pitches last year and wrote 10 checks. His $20 tool ranks 20 investors by fit for your deck and drafts the outreach.
🎬 Google cut Gemini Omni 1.2 Flash to ~$0.07 per second of 720p video (from $0.10), so a 15-second clip costs ~$1.05. Reprice your ad-creative tests.
✍️ GoDaddy's CEO skips any document he can "tell pretty quickly" was AI-written. Write your investor updates yourself, starting with your original observation.
Startup Investor Finder
Slidebean’s Investor Finder helps startups find and track the right investors faster. Filter by stage, industry, and location, organize outreach, and manage your fundraising pipeline in one streamlined workspace.
✦
RUSHIN' ROULETTE
Six bullets of updates
📡 Defense tech company Shield AI agreed to buy Echodyne, a Seattle-area maker of drone-tracking radar, in a deal worth up to $5.3 billion. Echodyne is profitable, raised only about $200 million, and was valued at $485 million after its last priced round in 2022. That's roughly 10x its last private valuation, proof that raising little and staying profitable can still end big.
🤖 Agility Robotics, maker of the Digit humanoid robot, started trading on Nasdaq on Nov. 5, the first dedicated US humanoid company to go public. Its filings show $1.8 million in 2025 revenue against a $140 million operating loss. Its rental plan, about $25,000 to deploy plus $8,500 a month for software and maintenance, is a handy benchmark if you price robots as a service.
🚚 Electric truck startup Harbinger raised a $240 million Series D, co-led by Tiger Global, Capricorn and Ridgeline, as it expands into defense and battery storage. FedEx has ordered 2,000 trucks worth more than $300 million, and a FedEx executive now sits on Harbinger's board. It's a clean example of how one anchor customer can carry a raise.
⏳ IREN, a "neocloud" that rents out AI computing power, said more than $4 billion in annual recurring revenue it expected in 2026 has slipped to at least January. The customer is Microsoft, which asked IREN to upgrade its network, co-CEO Daniel Roberts said. If you rent GPU capacity, get your delivery dates in writing: even fully contracted compute slips when a bigger customer wants changes.
🔒 Goldman Sachs is limiting advanced AI coding tools, like Anthropic's Claude Code and Microsoft's GitHub, to engineers and "strats" (its quant specialists), Bloomberg reported. The bank cites risk and productivity, and traders who built models and charts with them are reportedly frustrated. Before you hand coding agents to everyone, decide who actually needs them and why.
🧑💼 Perplexity launched Teammates, always-on AI agents that each take a role in your workspace, available only on its $200-a-month Max plan. The Verge calls it "much more expensive" than similar agents from OpenAI (Dots) and Anthropic (Claude Duet), which do comparable jobs. Before you staff a team with bots, compare agent pricing across all three.
At $8,500 a month, robot or human?
✦
SPONSOR
Build your shortlist before the show
Planning your next trade show? Start with the companies you want to meet. Lensmor helps B2B founders find relevant events, research exhibitors, and shortlist accounts that fit their ideal customer profile.
Reveal available business contacts and export selected records for outreach. Build your target list before you book the trip, so your team arrives with a plan for who to contact and why.
✦
STARTUP NEWS
Echodyne raised about $200M, then sold for up to $5.3B
Twelve years after Bill Gates co-led its initial round, Echodyne has a buyer. San Diego defense tech company Shield AI, maker of the V-BAT drone, agreed this week to acquire the Kirkland, Washington, radar maker in a deal worth up to $5.3 billion.
Eye in the sky: Echodyne makes compact radar that spots and tracks drones. It's built into counter-drone systems in more than 40 countries and into gear that stadiums and airports use to catch stray drones.
The slow road, per GeekWire:
2014: Echodyne spins out of Intellectual Ventures, a Bellevue invention lab, with a $15 million initial round led by Gates and Madrona. Madrona puts in $5.5 million at a $24 million pre-money valuation (the company's price before the new cash).
2022: Its last priced round values it at $485 million, per PitchBook. In total, it has raised about $200 million.
July 2026: It opens a $40 million factory in Woodinville with 260 employees. Today it has about 400, and it makes a profit.
$4.8 billion up front: Shield pays that at closing, plus up to $500 million more if Echodyne hits performance targets. The guaranteed part alone is roughly 10x the 2022 valuation. To pay for it, Shield is raising $2.5 billion, including $1 billion at a $28.8 billion valuation (more than double its March price). If the deal dies, so does the financing.
Niche work if you can get it: Echodyne never raised a giant round. It picked one hard problem, built the hardware and made money, so when a buyer wanted drone radar, Echodyne was the obvious call. The lesson: owning a niche and staying profitable can still end at about 10 times your last private price.
Closing time: the deal should close in early 2027, pending regulatory approvals. CEO Eben Frankenberg stays, and no layoffs are planned. His old college roommate, Smartsheet co-founder Brent Frei, has a line for it: "Our running joke is that we keep building 10-to-12-year overnight successes."
✦
STARTUP TV
How to Calculate Customer Lifetime Value the RIGHT Way
In this episode, Caya breaks down how to actually measure CLTV (and CAC) like a pro; using real templates, realistic churn, and actual payback periods.
We’ll cover the most common mistakes in LTV math, how to adapt it for SaaS, e-commerce, and marketplaces, and what your numbers really mean for growth.
✦
BIG TECH NEWS
Apple made $114.5B last quarter. Now it's trimming engineers' AI
Apple is shrinking the daily AI budget for some of its engineers, Bloomberg reported. The news came days after Apple posted a record $114.5 billion September quarter, up 12% from a year ago.
Allowance day: AI tools bill in tokens, the small chunks of text or data a model reads and writes. Apple caps how many dollars' worth of them each employee can use per day. Per Bloomberg, via Fast Company:
Some engineers' daily cap has already dropped from $300 to $200.
Apple reportedly plans to cut it further, as low as $100 for some engineering teams.
Many other departments are already capped at $100, and Apple is considering $25 for certain teams.
Employees read the cuts two ways. Some thought the $300 limit was more than most people needed. Others saw a cost cut, and maybe layoffs. New CEO John Ternus is also reportedly pushing to remove layers of management.
Tokenmaxxing is out: Apple is still pressing everyone to use more AI. But many tech employers are finding, per Fast Company, that their big AI bills haven't bought the productivity jump they expected. So the new goal is making each dollar show up as work. (Goldman Sachs is rationing too, by job title. See Bullet 5.)
The meter runs per head: here's some illustrative math, not Apple's. A $200 daily cap, fully used across about 250 workdays, comes to roughly $50,000 per engineer per year. A 20-person team maxing out would spend about $1 million. Unlike a server bill, that cost grows every time you hire.
Cap and trade: if your team runs on AI tools, set a default budget per seat now, plus an easy way for people who need more to ask. Pick the number yourself, before your invoice picks it for you.
Looking ahead… the $25 cap is only under consideration so far. But if Apple is metering AI after its best fiscal year ever, expect your investors to ask what yours costs per seat.
✦
STARTUP EVENTS
Startup Events and Deadlines
Crash Course in Financial Modeling | Webinar
Startmate Accelerator | Nov 08 | Apply
Web Summit 2026 | Nov 09 | Attend
SaaStr AI Annual 2027 | May 11 | Attend
✦
Scale your growth: Join names like HubSpot, Ahrefs, and Deel. Put your product in front of 85,000+ startup founders. Let's talk.




