The weekly top 10 for B2B tech operators · Every Friday
Done - check your inbox to confirm.
Burn Multiples tells you how efficiently your startup converts cash burn into ARR. It's the clearest signal of whether your growth is efficient - or just an expensive way to burn money. This breakdown covers benchmarks by stage and why 1.5x is the new magic number for survival.
The dreaded slide on everyone's pitch deck. What's the size of the market? TL;DR: If the market is not huge, same with your startup. Understanding the nuances of TAM, TOM, and SOM is a great way to get started understanding market sizes as it applies to your business and the business problem your startup solves.
A data center takes 12 to 18 months to build. And apparently connecting it to the grid takes 5 to 7 years! That gap is why nearly half of announced AI capacity is now planning its own on-site power and also why more than a quarter of the 110 data centers due online in 2025 are delayed.
This is a great guide (from a former designer at OpenAI) for us startups who want a more distinct and cohesive visual identity - and how to leverage visuals for your prompting with AI.
ICONIQ's third State of AI report is out - opening line is a banger: "_a year ago, boards wanted to hear the AI strategy. Now they want AI unit economics_". 66% of those surveyed put agentic capabilities in their top three product investments (so it's the clear #1), AI products went from 32% to a projected 42% of revenue, internal agent productivity gains sit under 30% (not suprised), Consumption pricing rose from 35% to 42% of AI builders in six months, outcome-based from 18% to 23%.
Another great report out this week - this one on how much to pay the Finance bigwigs: The 2026 Technology Finance Executive Compensation Benchmarks - CFO median base $285K, median target bonus $100K, median annual equity $1M. Cuts by revenue tier, geography and funding source. But a CFO at a $100M+ company earns about 55% more than one at a sub-$10M company.
Apple is suing OpenAI for trade secret theft, and the filing reads like they have all the receipts plus some. Apple audits company devices and server logs and alleges that one ex-engineer never returned his Apple laptop and then used it to pull confidential docs. They want discovery. They call this 'the tip of the iceberg.' Popcorn time.
We're beyond vibe coding, and our agent co-workers write code faster than us humans can review it. This article thinks that the answer isn't reviewing faster; it's reviewing less - their agent gives the final approval on roughly 1 in 3 PRs merged to main, 1.6K last month. It only touches PRs under 500 lines, with auth, secrets, and billing deny-listed.
Repeat after me - persona, problem, proposition, positioning, product, promotion - those are the six Product "Ps" s for product validation - find out before you build it - pitch it to 10-20 target customers, and come back with either a send it, iterate it, or kill it decision.
Most VC firms scale by adding partners or dollars. Theory Ventures scaled by adding engineers - three investors and a nine-person intelligence organization, thirteen people total, building agents that map markets and surface companies before they raise. The results: 2x the deals analyzed.
Get the top 10, every Friday
Curated SaaS and tech insight from around the web, repackaged for people to put to good use — free.
Done - check your inbox to confirm.
SPEND - 48% of equity-backed private SaaS companies are running at a loss, versus 17% bootstrapped. Crazy stat - VC-backed companies are spending 100% more on marketing and customer success, and 70% more on sales, to grow 25% a year against a bootstrappers' 20% average. The biggest single line is R&D at 22% of revenue, ahead of sales at 15% and marketing at 8%.
The person writing you a check is not necessarily the same person getting value out of your business. So take a read of this insightful article from HeavyBit on differentiating messaging based on this premise and the different profiles.
Sure, AI can make you fast - but can it make your company faster? This article argues that it's because we are handing off the slow stuff to everyone downstream of you.
The AI Corner watched a 90-minute Jeff Bezos interview so you don't have to, and also pulled 10 ideas founders and investors should/could steal. Here's a good one - the engineer's job is not typing code; it's identifying problems, and that's the layer that survives.
Venture is on a boom. Global venture funding reached an all-time record of $510B in H1 2026 (ALL of 2025 was $440B). Sounds crazy, but the majority of that money is concentrated into some insane deals (OpenAI and Anthropic account for $217 B of that figure all by themselves). IPOs and M&A are also back.
It's getting competitive - Model prices are collapsing. a16z clocked a 100x drop, from $60 to $0.60 per million tokens in three years. But AI bills are going the other way - because token burn is going way up.
Check out this study of 1M+ high-volume keywords - about a 30% decline in search demand. But SaaS is a winner: its search volume is up 48% year-over-year, one of only three verticals beating the AI-driven drop-off. FinTech got hit hardest; it's down 37.7%. Also - crazy new data point - 18% of consumers have bought something on an AI recommendation without running a separate search
Want to read a 40-page paper on "The Impact of Google AI Overviews on Publisher Traffic and User Experience"? Didn't think so, so here is the TL;DR - When a Google AI Overview appears (on about 40% of searches), clicks throughs drop 39.8% and zero-click searches jump 34.5%. What is interesting here is that Ad clicks don't move, so Google's revenue is just fine, and the publisher eats the loss. Users were no happier with the summaries than without - so it's diverting traffic without really helping the searcher.
Everyone's future problem - AI is eating the bottom of the Dev career ladder, not the top. US software developers aged 22 to 25 are down 19% from their late-2022 peak, while the 41-to-49 cohort is up 14%.
Coca-Cola cut their marketing-investment decision cycle from two weeks to an hour using 'Fuel Light 360', their proprietary AI scenario-modeling tool. It lets teams model marketing spend scenarios in real time, which also shifted meetings from arguing over whose right to actual decisions.
CAC PAYBACK: The 'payback' period is the nuance of why we measure CAC. How long until we break even? Benchmark-wise, the negative trough is way longer than you think, so take a seat! New B2B customers, on _average_, take 2 years and 2 months to become profitable. SaaS world changes that a bit - payback periods have accelerated in the - 5X increase in payback under 6 months (see report at #2 below for more on that) CAC payback still really highlights the deep dependency on access to capital to fund a Tech company's growth through these SaaS Cash Flow Troughs.
A genuinely meaty report from Mercury Fund - their 2026 Early-Stage GTM survey (125+ B2B leaders) covers funnel, unit economics (see CAC above), comp, org design, and AI.Win rates up from 29% to 33%, mid-funnel sliding (Lead-to-MQL 41% to 32%), 4:1 LTV:CAC from 0% to 14% of teams. ChatGPT adoption at 80%, Claude at 60%, Gemini at 50%, though that may have changed in the months since this survey was completed.
ChatGPT referral traffic to monitored sites jumped over 60% in May, and homepages went from 3.5% of those referrals to \~24%. The key takeaway: your homepage is now read by the LLM and clicked by the human, so it needs to serve both of 'em. Seven tactics in there from Clay, Rippling, Lovable, Ahrefs.
Agentic AI spending is gonna hit $201.9B in 2026, up 141%, and it overtakes chatbot/assistant spending in 2027 for the first time, according to Gartner's first dedicated AI forecast. Chatbots peak at $264.7B then decline; agentic grows at 119% CAGR to $752.7B by 2029. Talky Bots to Acty Bots.
The high-agency test - do you dare to try it? It's the 10x you. Picture a version of you ten times faster and braver, then do what they'd do today, the scary email, the project you'd close. The gap, apparently, isn't information, it's willingness to take discomfort now.
Check this report covering platform sales and revenue - The biggest gap between high and low performers in platform sales isn't how they sell, it's what happens after signature: on whether customers can see value post-go-live, top sellers scored 1.86 vs 3.23 (lower is better) - platform sale fails not because buyers aren't interested, but because sellers answer feature questions instead of the three buyers ask, can it be done, is it worth it, will we survive.
Nightmare VC story time - A Sequoia partner passed on Cloudflare early because he didn't think a woman could run a security infrastructure company (now worth \~$80B). A $5M VC commit was cut to $100K on closing day, A GP fell asleep during a Series A pitch, and nobody in the room woke him - way more cringe stuff in there!
If you even vaguely sell a wrapper, don't price AI like you are reselling electricity. Price it based on the work it gets done. If a customer brings their own API key and sees raw inference cost on their bill, will your pricing survive? Cost-plus margin breaks, value-based holds, because you're selling outcomes, not tokens. Good example in there - Sierra charges per resolved ticket (and $0 for failures).
Digging this one back from the archives - every feature you say yes to is maintenance, bugs, and onboarding cost you signed up for forever. So knowing when NOT to code is the skill most of us need to get better at or master.
Ascend COO Omar Ismail boosted ARR 38% in six months without hiring a growth team, using Claude Code to build and run the entire stack. Six months ago: $20M ARR, 95% from word-of-mouth, zero paid acquisition
Taking the week off next week - Happy 4th of July break y'all!
Time to value (TTV)- an old but re-emerging metric du jour (see why in #2 below): the gap between a customer signup and the first real product value, the Aha moment (like Dropbox's first shared file, or Slack's first channel message). There is also a corollary "trough of disillusionment" that your customers may need to navigate.
Is NPS dead? A panel of CS leaders from Content Square, LinkedIn, and Carta discussed this very question at SaaStr AI 2026 - their opinion? Yup! NPS has low response rates; non-responders are churners, and there is no correlation with gross revenue retention. One of them keeps it only because the board still asks for it. Replacement KPI: time to value - see #1 above.
Big report on AI pricing. 89% have exceeded their initial AI budget - 45% significantly, 44% moderately, only 9% on budget. Top reason: AI features drove more usage than expected (67%), usage scaled faster than expected (63%). Only 10% blamed vendors changing pricing post-sale. 67% name IT as the primary owner of AI cost risk, just 17% Finance. 55% find credit/token pricing harder to evaluate for AI than for SaaS.
Following on from #3 above, Gartner projects 40% of enterprise apps will ship task-specific AI agents by the end of 2026, up from under 5% in 2025. B2B consultant Jason Greenwood argues that embedding is the lock-in moat: switching providers means losing institutional context and breaking automations, not just swapping an API key.
This is a good one - I'm guilty of more than I care to admit - "The cemetery of failed startups doesn't give interviews." Giacomo Falcone has a rundown of 17 cognitive biases we can all admit to, covering survivorship bias, anchoring (whoever speaks first sets the negotiation ceiling), and the blind spot that the more sophisticated you are, the better you rationalize flawed thinking.
I haven't talked much about this in a while - and it comes with another quotable post - "Money lets founders scale noise," Oh and here is another quote (and also the title of the book referenced) "The Power of Broke" and I can certainly relate - A funded founder can spend months on brand, hires, and features before learning the offer is unclear, capital removes that pressure too early, letting founders buy activity before earning clarity.
Probably one of the most significant AI drama/stories yet. The US government ordered Anthropic to suspend Fable 5 and Mythos 5 over the weekend, citing "national security". The directive barred access by any foreign national, including Anthropic's own non-citizen staff, forcing Anthropic (a couple of hours later) to pull both models worldwide, including on AWS Bedrock. And then they had to go visit Trump in DC - coincidentally also on his birthday.
New working paper on AI productivity - Claude Code now signs over 5% of all public GitHub commits, Autocomplete raised developer commits 40%. Sync agents took the cumulative effect to 140%, async agents to 180%. New iOS apps roughly doubled to \~100K/month since agentic coding arrived, but total usage of new apps is flat across the App Store
That bonkers SpaceX $75B IPO last week (biggest in history) also paid a crazy $500M in underwriting fees, also the biggest ever - which sounds like a massive amount but is only 0.67% of the overall $$. Most IPOs pay 4-5%. Goldman Sachs and Morgan Stanley each took 20% of SpaceX's fee allocation, which is about $100M apiece.
The OG Cloud bro, Salesforce, reaccelerated subscription growth from 9% to 12% at $45B+ ARR. The catch in the new deck: seat-based core apps grew just 7%; the reacceleration came from a data/platform layer up 23%.
Get the top 10, every Friday
Curated SaaS and tech insight from around the web, repackaged for people to put to good use — free.