Most companies are already sold on AI. The problem is they can not make it work. According to Gartner, up to 85 percent of enterprise AI projects never reach production. A startup with Marc Benioff’s backing is raising capital on the belief that deployment is the most expensive unsolved problem in tech right now, and that whoever cracks it will be worth more than most of the model builders getting all the press.
The Gap Nobody Is Talking About
Here is the situation. Companies are spending billions on AI tools, platforms, and pilots. According to IDC, global AI spending hit $235 billion in 2025 and is projected to reach $632 billion by 2028. But most of that money does not produce results. It sits in sandbox environments. It gets stuck in IT reviews. It dies in committee meetings.
McKinsey reported that only 8 percent of companies have managed to embed AI meaningfully across their core business operations. Eight percent. That means 92 percent of companies that bought AI still do not have anything running in production at scale.
That gap is not a small problem. That is a structural failure in how AI gets built and shipped. And now Benioff, the founder of Salesforce and one of the more consistent early checks in enterprise software, is backing a team that thinks they can fix it.
The Real Money Is in Making AI Actually Work
The startup is not trying to build a better model. They are not competing with OpenAI or Google. The play is different. They are building the layer between the AI tool and the working product, the scaffolding that most engineering teams rebuild from scratch every single time they try to ship something.
Think of it this way. If AI is electricity, most companies are still wiring their own houses. Every team starts from zero. They figure out how to connect the model, manage the prompts, handle edge cases, build monitoring, set up fallbacks, and get approval from legal. That process takes months. Sometimes years. And it breaks differently at every company.
This is what the rich versus poor mindset looks like in tech right now. Companies with 50 or more engineers can absorb that cost. They hire specialized teams. They build internal tooling. They ship in six months. Small companies and midmarket operators spend the same six months and never finish. They stay in pilot mode forever.
The startup is betting on compression. They want to take a six month deployment cycle and turn it into six weeks. If they do that reliably, they own the most valuable chokepoint in enterprise AI. Every company that wants to ship AI becomes a potential customer. According to PwC, AI could add $15.7 trillion to the global economy by 2030. The startup that controls deployment controls the on-ramp.
I have seen this play before. The money in tech rarely goes to whoever builds the best thing first. It goes to whoever makes the best thing usable. Stripe did not invent payments. They made payments easy. This startup is making the same bet on AI deployment.
Content teams dealing with this exact problem, stuck between having an AI tool and actually shipping output at scale, are increasingly turning to platforms like InVideo AI to bridge the gap between idea and finished product without rebuilding the workflow every single time.
What This Means for You
If you run a business and you are still in AI pilot mode, this news should bother you. You are not stuck because AI is hard. You are stuck because the deployment layer has not been productized yet. That is changing fast, and the companies moving now will have a two year head start on everyone waiting for perfect.
Here is what I would do right now.
Stop waiting for your IT team to figure it out in house. The companies winning with AI right now did not build everything from scratch. They found tools that did the heavy lifting and deployed fast. Velocity beats perfection in this market.
Audit where your AI projects actually die. Is it the model? Probably not. Is it the integration, the compliance review, the data pipeline, or the approval chain? Find the real chokepoint and either fix it or buy your way around it. Most deployment failures are not technical failures. They are process failures dressed up as technical ones.
Watch this space closely. A wave of deployment focused startups is coming. When Benioff puts money into something, other enterprise investors follow. Expect three to five major players to emerge in the next 18 months, all targeting the same problem from different angles. The winner will probably be the one that charges less per seat and integrates with what you already use.
Operators who want to move fast should also watch deal aggregators. Sites like AppSumo regularly surface early access to deployment and automation tools at one-time prices before they go subscription only. If the next wave of AI deployment tooling shows up there first, it pays to have an account and be watching.
The Bottom Line
The AI race is not over. It is entering a new phase. The builders who captured phase one were the model labs. The builders who capture phase two will solve deployment. Benioff has made a career of knowing which phase comes next. I would not bet against him on this one. The company that makes AI shippable for the average enterprise will be worth more than most of the AI companies getting all the headlines today.
Frequently Asked Questions
What is the AI deployment problem?
The AI deployment problem refers to the gap between building or buying an AI tool and getting it into production where it runs reliably at scale. According to Gartner, up to 85 percent of AI projects fail to reach this stage. The problem is usually not the AI itself but the integration, compliance, and infrastructure work required to actually ship it.
Why is Marc Benioff investing in AI deployment startups?
Benioff has a long track record of investing early in enterprise software infrastructure, most famously Salesforce itself. AI deployment sits squarely in that category. With the majority of enterprise AI spending failing to produce production results, the deployment layer represents one of the largest unmet needs in the current tech market.
How big is the AI deployment market?
Global AI spending is projected to reach $632 billion by 2028 according to IDC. A significant portion of that spend is wasted on projects that never ship. The startup that reliably compresses deployment timelines could capture a meaningful share of what is effectively a multitrillion-dollar bottleneck.
What can small businesses do about the AI deployment gap?
Small businesses should focus on prebuilt AI tools rather than custom builds from scratch. The deployment problem is most severe for companies trying to engineer everything in house. Using prebuilt platforms and staying alert to new tooling from the current wave of deployment focused startups gives smaller operators a real edge without a six month engineering cycle.
Will AI deployment startups replace internal engineering teams?
No. Internal teams will still be needed to configure, monitor, and customize AI systems for company-specific needs. What deployment startups replace is the months of undifferentiated infrastructure work every team currently rebuilds from scratch. They compress the grunt work so engineers can focus on the parts that actually require deep company-specific knowledge.
