How Startup Teams Build Products That Actually Last
- Stephanie Roulic
- 4 hours ago
- 5 min read
At Startup Boston Week, the panel “From Cradle to Grave: Designing Products That Last (and Last)” explored one of the hardest transitions in startups: moving from a scrappy prototype to a real product that can survive customers, manufacturing, scale, and the real world.
Moderated by Sheri Palazzo, the conversation brought together leaders across engineering, manufacturing, batteries, and venture capital, including John Unima, Raimund Koerver and Praveen Sahay.
And while the title focused on durability, the discussion became something much bigger: a reality check on what it actually takes to build hardware companies in 2026.
Your Prototype Is Not Your Product
One of the clearest themes throughout the session was that early-stage founders often underestimate how different a prototype is from a scalable product.
“In the beginning, you’re willing to do anything and everything to get it working,” Palazzo explained. “A working prototype means you have something to show a customer, investor, or partner.”
But once a product enters the market, expectations change quickly. “You can hand-wave around things when you’re building one-offs,” Palazzo said. “But once it actually hits the market, you’re setting expectations based on competitors and customer standards.”
That transition from “it works” to “it works reliably, repeatedly, and at scale” is where many startups struggle.
Koerver described how quickly design flexibility disappears once manufacturing begins, “When you go from prototype to production, you lose a lot of design freedom,” he said. “And then people actually want to use it, not always how you intended.”
That means customer behavior itself becomes part of the engineering challenge.
“You haven’t even developed a spec or your spec isn’t comprehensive enough,” Koerver said. “And suddenly customers are doing things with the product you never expected.”
Hardware Doesn’t Give You Many Second Chances
Throughout the panel, the speakers repeatedly contrasted hardware startups with software companies. In software, founders can often release quickly, iterate constantly, and improve products over time. In hardware, mistakes are significantly more expensive.
“You don’t have the privilege of tweaking your product continuously,” Sahay said.
John Unima compared hardware design to building a house, “It’s really hard to change the foundation of a house after the house has been built,” he said.
That reality forces hardware teams to think further ahead earlier in the process. “You need to look way down the road,” Unima explained. “You need to hope you get the foundational parts right.”
The panelists emphasized that this doesn’t mean founders need to solve every possible problem immediately. Instead, they encouraged startups to focus on identifying and reducing the highest-risk areas first.
“If you try to worry about all the problems at once, you run out of time, people, and money,” Palazzo said.
Customers Will Always Use Your Product Wrong
Another recurring theme: customers will inevitably use products in ways founders never anticipated. And that becomes especially painful once manufacturing is already underway.
Koerver shared an example from the battery industry where a component supplier manufactured large quantities of parts that later created downstream problems.
“You already bought them,” he said. “Your warehouse is full of this material and the supplier designed it specifically for you. They’re not taking it back.”
That reality makes expectation-setting incredibly important.
“You always test something you’re almost sure you can deliver on,” Koerver said.
Rather than overpromising, the panel advocated for carefully scoped milestones and clearly negotiated customer agreements. Palazzo summarized the philosophy simply, “Under promise and over deliver.”
Startups Waste Too Much Time Solving Problems That Already Exist
One of the strongest practical takeaways from the session was the importance of borrowing from industries that have already solved similar problems.
“Don’t waste your minimal time, energy, and money solving problems other people have already solved,” Palazzo said. “It’s okay to copy them.”
Unima agreed, emphasizing the importance of hiring experienced operators who understand both startup chaos and production discipline, “You get penalized in school for copying, but you get promoted at work,” he joked.
The conversation repeatedly returned to the value of experienced teams. According to Unima, one of the biggest mistakes startups make is hiring people who only understand mature systems but cannot operate inside startup ambiguity.
“You need people who understand what steady-state production looks like,” he said, “but who can also operate in the ‘we need to get this done today’ environment.”
Hardware Is Hard, But That’s the Opportunity
Sahay took the conversation in a broader direction, arguing that the future competitive advantage for the U.S. lies in physical products and manufacturing innovation.
“The entire value proposition of software companies is collapsing,” he argued, pointing to the rise of AI-generated software and lower barriers to entry. By contrast, he sees hardware as having two enormous advantages:
stronger barriers to entry
massive market opportunities
“If you’re working with customers and building to their specs, you become deeply embedded,” Sahay explained. “Removing you becomes very difficult.”
He also emphasized that hardware companies can create long-term defensibility through manufacturing partnerships, supply chains, certifications, and technical know-how that competitors cannot easily replicate.
Even seemingly ordinary products still contain room for innovation.
Sahay highlighted a portfolio company that redesigned residential water heaters to reduce energy consumption by 40% while eliminating scaling and mineral buildup, “There is scope for innovation in everything,” he said.
Hardware Startups Need Customer Conversations Earlier Than They Think
One of the most repeated lessons from the panel was simple: talk to customers earlier and more often. Sahay shared advice once given to a founder in his portfolio, “Don’t spend any of this money on R&D. Spend all of it on plane tickets.”
The idea was straightforward: before building deeply, founders need to understand exactly what customers actually care about. Because many startups, according to Koerver, end up solving “PowerPoint problems.”
“They’re solving something that doesn’t actually exist,” he said. “They just haven’t had a customer tell them yet.”
The panel encouraged founders to spend less time assuming and more time validating, especially in hardware, where changes become exponentially more expensive later.
AI Is Already Changing Hardware
While the session focused heavily on physical products, AI still became part of the conversation.
The panelists argued that AI is increasingly shaping both product development and manufacturing workflows, “Physical AI is where the real wealth creation is going to happen,” Sahay said, pointing toward robotics, automation, and intelligent systems operating in the physical world.
Palazzo added that even hardware startups are increasingly expected to have an AI story when speaking with investors, “You’re going to get asked: how are you using AI?” she said.
For engineers, AI is also changing how systems are designed. Unima explained that hardware increasingly exists to collect clean, structured data that software and AI systems can interpret and act upon. “It’s changing how hardware engineers think about the systems they’re designing,” he said.
There Is No Single “Right” Way to Build Durable Products
Toward the end of the session, the panel tackled a founder question about whether startups should build “military-grade” durability into products or focus on cheaper, replaceable systems.
The answer: it depends entirely on the business model,“There is no right or wrong answer,” Palazzo said. “Unless your customer expects it to last 20 years.”
In fast-moving industries, overengineering can actually become a disadvantage, “A drone you build today could be outdated in two years,” Sahay noted.
The key, according to the panel, is aligning durability, speed, cost, and customer expectations intentionally rather than assuming every product needs maximum longevity.
The Real Challenge Isn’t Building the Product
By the end of the session, one message became clear: building the technology is only part of the challenge.
The harder part is balancing product development, manufacturing, customer expectations, distribution, partnerships, fundraising, and long-term scale all at once.
As Palazzo summarized, “The challenge of startups is having the tech, money, and business development all progressing equally at the same time.”
