Sunsama +25%, OpenAI vs Anthropic, and tokens vs humans

This week: Sunsama's first price increase in five years, an IPO window that reopened without SaaS, and the question CFOs are now asking out loud — does the next dollar go to tokens or to a new hire?

Sunsama raises prices 25% after five years flat

On June 8, the SaaS Price Hub tracker flagged a price move at Sunsama: the daily planner's single Pro plan now costs $25 per month, or $20 per month billed annually — up from $20 and $16, per Sunsama's pricing page. That is a 25% increase on both billing cycles, and the company's first price change in five years.

Sunsama explained the move in its published pricing manifesto: after five years of holding price steady while absorbing rising costs, it repriced to stay profitable and sustainable. The structure itself didn't change — one paid plan, all features included, no free tier, a 14-day trial with no card required, and the 20% discount for paying annually survives intact, as Morgen's pricing breakdown confirms.

The pattern is worth noting: independent productivity tools tend to hold prices flat for years, then reset in one visible jump — the opposite of the 7-8% annual escalators common in enterprise renewals. Spread over five years, Sunsama's 25% rise works out to roughly 4.6% a year.

What it means for buyers: monthly subscribers now pay $300 a year instead of $240. Single-plan vendors leave no downgrade path — when the only plan repricing, the choice is absorb or churn. If you're on monthly billing, switching to annual at $240 a year now saves exactly the amount of the increase.

Anthropic files. OpenAI plans price cuts. SaaS still waits.

On June 1, Anthropic confidentially filed a draft S-1 with the SEC, the company confirmed, days after closing a $65bn Series H that valued it near $965bn, per CNBC. TechCrunch notes it is the first major AI lab to formally start the IPO process, ahead of OpenAI; Yahoo Finance reports an annualized revenue run-rate above $47bn.

The contrast on the other side of the software market is stark. Crunchbase News reports that while IPO activity overall is holding up in 2026, not a single venture-backed SaaS unicorn has filed this year. Figma trades down more than two-thirds from its peak, Navan has shed over half its value, and Blackstone-backed Liftoff withdrew its planned IPO this month. Large private names like Canva and Rippling face the same closed window.

Ten days after Anthropic's filing, the WSJ reported on June 11 that OpenAI is weighing "drastic" cuts to its token pricing — explicitly anticipating that Anthropic will cut first. OpenAI's consumer subscriptions sit at $8, $20, and $100+ per month; Anthropic's Claude Pro is $17 with an annual subscription and Max at $100+. Both companies already lose billions on compute. A price war between the two market leaders, immediately ahead of dueling IPOs, sets the ceiling for what every other AI vendor can charge through 2027 — and gives every CFO a defensible reason to push back on AI vendor renewals over the next two quarters.

The business-model read: public markets are no longer pricing "software" as one category. Usage-based AI infrastructure commands a near-trillion-dollar valuation while per-seat application software trades at a discount on the thesis that AI displaces it. For buyers, the practical consequence is that SaaS vendors locked out of public markets must squeeze growth from existing customers — expect firmer renewal terms, AI add-on meters, and repackaging across the incumbent tools in your stack before any of them ring a bell on the NYSE.

The CFO's new question: tokens or humans?

For the first time, enterprise technology budgets are starting to look like labor budgets — and CFOs are openly making the comparison. In a May 29 CNBC segment, anchor Deirdre Bosa pressed two CEOs at the center of that trade: Arvind Jain of Glean (just past $300M ARR) and Matan Grinberg of Factory AI.

Jain's argument: AI spend now competes head-to-head with headcount on the same expense line. The next incremental dollar inside Fortune 500 companies is increasingly going to AI capacity rather than a new hire, and annual AI budgets are being exhausted in weeks rather than quarters. That's the pattern the SaaS Price Hub tracker started flagging in mid-May with the per-resolution billing shift on Zendesk and the 65%+ of major SaaS vendors who've layered an AI consumption meter on top of seat pricing.

Grinberg's framing: companies are now routing the same workload across multiple AI models — picking by cost, not capability. Factory AI's product orchestrates this routing at the engineering-team level. The pricing implication: model choice is becoming a buy-side decision rather than a vendor lock-in. Margin-conscious buyers will increasingly demand multi-model support inside whatever SaaS they're paying for, and "powered by Anthropic exclusively" or "powered by OpenAI exclusively" turns into a downside in 2026 procurement.

What it means for buyers: when the AI line on your invoice starts rivaling a salary, the procurement frame has to change. You're not buying software anymore — you're allocating labor budget through a non-headcount vehicle. The questions to ask vendors: what does worst-case monthly cost look like at peak usage, what's the model-substitution path if a cheaper alternative emerges, and what's the cap before re-negotiation triggers.

Also this week

• Nvidia acquired predictive-AI startup Kumo AI for over $400M on June 3, folding enterprise churn and credit-risk prediction models into its full-stack push, per Fortune.
• Ramp raised a $750M Series F at a $44B valuation on June 4, led by Iconiq, GIC, and Ontario Teachers', per SaaSRise's weekly roundup.
• Supabase raised a $500M Series F at $10.5B on June 4, led by GIC, as AI app builders drive backend demand, per SaaSRise.
• AlphaSense raised $350M at a $7.5B valuation on June 3, led by Vitruvian Partners with Accenture Ventures and J.P. Morgan Asset Management, per SaaSRise.