Founders: Grok for Startups is accepting applications for product credits and support.
Credits depend on eligibility and usage. The program aims for six to twelve months of usage; it isn't a funding round.
Review the terms:
x.ai/startups
Too much inventory. Not enough buyers.
Offe Market raised $3M for commerce where discounting still has to protect a brand's introduction to a customer.
Rachel Gannon's operating assignment:
devcuration.com/articles/off…
A medical bill can tell two different financial stories. Somebody still has to negotiate.
TempusDirect secured Aldrich's backing for claims analysis and services.
Read our analysis:
devcuration.com/articles/tem…
Your U.S. company is incorporated. Who handles payroll next month?
@gotavi_co acquired Avatar Advisors to extend that operating relationship.
Read the acquisition story:
devcuration.com/articles/got…
Virent Capital's first platform investment is Kanahoma. Seth Odell stays CEO.
The bet reaches beyond advertising into enrollment execution.
Read the investment story:
devcuration.com/articles/kan…
A promising medicine still needs a route to patients.
Golden Age Health secured Kos Biotechnology Partners' backing for regional commercialization in Asia. Amount undisclosed.
Meet the people carrying that work:
devcuration.com/articles/gol…
A $40B post-money valuation.
@Kalshi raised $1B, co-led by @sequoia and Wellington.
Tarek Mansour and Luana Lopes Lara are building toward an exchange institutions depend on.
Our analysis:
devcuration.com/articles/kal…
A basketball has its own sound. A roaring arena makes it difficult to hear.
@official_esr_ raised another $2M to separate useful audio from the noise.
Valtteri Salomaki and Ethan Castro are building technology that connects microphones, tracking data and live production.
A strawberry shipment can lose its best buyer before anyone spots the quality problem.
Loamy raised $6M to connect food safety, inventory and sales decisions.
Charles Julius and Jason Pesek are building for the moment a warning can save a shipment from the wrong destination.
Park Road Partners closed its first fund at $695M.
Kyle Peterson originally targeted $375M.
LPs have backed his investment experience before the new firm has a realized fund record of its own.
The fundraising result is impressive. The portfolio now has to earn that confidence.
Getting six friends to agree on dinner can become a management position nobody applied for.
@ekuyda is building @wabi around the little jobs that appear when people try to make plans together.
The company raised $55M at a $1.25B valuation, led by IVP.
Wabi combines AI assistance with shared spaces and personalized tools for things like meal planning, restaurant lists and group activities.
The commercial question is whether the software earns a recurring place in the relationship.
A helpful suggestion gets used once. A shared plan has to survive somebody changing their mind, another person missing the message and everyone arriving at a different time.
Wabi has described the product direction, but active-user totals, retention and revenue remain undisclosed.
devcuration.com/articles/wab…
Bose stopped making Sleepbuds. The people who depended on them still needed to sleep.
Three former Bose colleagues built @OzloSleep around that unfinished customer relationship.
N.B. Patil, Brian Mulcahey and Charles Taylor acquired and licensed sleep-related technology and turned it into a new business.
Ozlo has announced a $22M Series B led by Modi Ventures, including previously issued convertible instruments.
The product has a demanding customer review process.
Somebody has to put the device in their ears, turn onto a pillow and decide whether the whole experience deserves another night.
Ozlo now wants to expand hardware, software, research and services around that relationship. Its customers will make the decision again every evening.
devcuration.com/articles/ozl…
The network goes down. The supply decision still has to be made.
That operating reality sits behind @salesforce's agreement to acquire Rune Technologies.
David Tuttle and Peter Goldsborough built Rune's military logistics software to keep critical information available locally and exchange prioritized updates when connections return.
Rune says its TyrOS product supports inventory, personnel, equipment and distribution decisions, including work with the U.S. Army and Marine Corps.
Salesforce intends to bring that capability into Missionforce.
The acquisition price is undisclosed and closing remains subject to conditions.
The interesting integration question is what happens when decisions made offline have to return to the enterprise record without losing the context that made them necessary.
devcuration.com/articles/sal…
Every large company has people carrying an absurd amount of unfinished business in their heads.
A decision from Tuesday. A dependency nobody documented. The reason one team promised another team something six months ago.
Alex Vogenthaler and Peter Terrill built Rocker around the problem of keeping that context available when somebody needs to act.
The company raised $12.85M with @greycroftvc leading.
Rocker describes assistants that connect project history, decisions and dependencies while keeping access subject to permissions. It also offers deployment inside a customer's own cloud.
That last detail belongs in the sales conversation.
An organization may want the benefit of shared institutional memory while remaining deeply protective of where that memory lives and who gets to use it.
devcuration.com/articles/roc…
An employee builds an AI agent, connects it to company systems and eventually leaves.
Who owns the agent on Monday morning?
@mintive has made that question part of its product architecture.
@JiquanNgiam, Vijay Vasudevan, Bozhi See and Phillip Bensaid have raised $7M for infrastructure that gives agents identities, delegated permissions, monitoring and transferable ownership.
The company describes a customer example at Modern Treasury where an agent researches support cases across documentation, code and operational data before a specialist responds.
Customer-facing actions still require human approval.
The commercial opportunity is larger than making an agent useful once. Enterprises need the assignment, permissions and accountability to survive changes in the people running the business.
devcuration.com/articles/min…
Dave Yuan (@davelyuan) of @tidemarkcap makes a practical case for testing new AI products separately from existing software pricing. Learn what customers use and value before changing the commercial terms for everyone already paying.
hgcapital.com/insights/orbit…
The interesting promise in @DoorDash's proposed $17B acquisition of @Instacart is directed at the grocers.
They would remain free to use competing delivery platforms.
Tony Xu and Chris Rogers are proposing a combination that could put delivery, digital storefront technology and advertising into a much broader relationship with retailers.
Instacart brings Marketplace, Enterprise and Ads. DoorDash brings its delivery network and an enormous consumer transaction base.
The proposed deal would also bring DashPass and Instacart+ under one membership fee.
But the grocer still has to decide how much of its business should depend on a single supplier.
The transaction is expected to close in late 2027 or the first half of 2028, subject to approvals. Until then, these remain separate companies and the proposed integration remains ahead.
devcuration.com/articles/doo…
Sebastian Thrun and @adrianaoun have raised $130M in seed financing for Dulo, a company pursuing foundation models for hardware design.
That is a substantial bet on a problem where a beautiful answer can become an expensive manufacturing headache.
A generated part still has to survive material constraints, production methods, tolerances and the people responsible for building it.
Dulo has not publicly supplied a customer list, performance benchmark, supported CAD integration or product release date in the record reviewed.
That makes the next demonstration unusually consequential.
The founders have serious experience and serious backing. The engineering market will want to see what the system can produce, how the work can be checked and whether it makes the next manufacturing decision easier.
devcuration.com/articles/dul…
The most expensive AI mistake may be the one an expert recognizes immediately but an engineering team never learns how to prevent.
Crux Labs is building around that handoff.
Julie Kheyfets and Krishna Kulkarni have raised more than $6M to connect domain-expert judgment with agent evaluations and proposed code changes.
The product aims to capture why a specialist rejects an answer, test that reasoning against production interactions and send proposed fixes toward the engineering repository.
There is a serious question underneath the technology.
When experts disagree, whose judgment becomes the standard?
Crux says its platform can evaluate 100% of production traffic. That is a coverage claim. Whether the evaluations are correct is a separate question customers will need to test.
devcuration.com/articles/cru…
A 1991 industrial controller was ready for retirement.
The safety chain wasn't.
Neuron Industries replaced the Siemens S5 at CureWood while preserving the hardwired safety chain and keeping the original controller available as a fallback.
Dennis Ren and Kenneth Rhee are building AI-assisted industrial controls around the realities of physical equipment.
The company reports roughly three hours of human engineering and an eight-hour cutover on a line with 189 hardwired inputs and outputs.
Those are different clocks, and they matter.
A factory owner has to know what happens when a new system fails at the worst possible moment. Neuron's $10M seed financing gives the team more room to build around that responsibility.
devcuration.com/articles/neu…
