While AI funding has been soaring, a quieter revolution is taking place. The most innovative AI startups aren’t necessarily the ones landing these massive investments — they’re the ones outmaneuvering larger players by embracing data mobility.
At Backblaze, we bootstrapped until just months before our IPO on less than $3 million in outside funding, while competitors raised hundreds of millions. This experience has shown me that the amount raised or spent does not necessarily determine the winners. And in AI, the ability to move and use data freely matters more than enormous capital reserves. The next wave of AI breakthroughs won’t come from those with the deepest pockets but from those who have built the most flexible, open infrastructure.
The AI Competitive Landscape Has Changed
For emerging AI companies, the major cloud providers present a paradox. These hyperscalers offer vast resources but increasingly compete in the same AI space as their customers. This creates an uncomfortable reality in which startups must decide whether to build on platforms owned by their competitors.
What we’re seeing is that innovative AI companies don’t want to work with hyperscalers precisely because those same hyperscalers are their direct competition. They need independent data storage providers that won’t restrict their options or compete with their core business.
This isn’t just speculation — an increasing number of our customers are AI companies. The pattern is clear: Innovative AI firms want sovereignty over their technology stack. Companies like DeepSeek have demonstrated that remarkable innovation doesn’t require unlimited resources but instead benefits from the ability to innovate.
The Hidden Tax on AI Innovation
What many organizations don’t realize is that cloud egress fees amount to a hidden tax on innovation. These fees — which can represent 40% of total cloud costs for some customers — punish exactly the behavior essential to AI development: Experimentation, iteration and deployment across different environments.
The major cloud providers are spending billions on capital expenditures, buying up chips and then trying to lock customers into their ecosystems to access those chips. This lock-in strategy works against the needs of AI development. Training and fine-tuning models often require moving data between different compute environments as opportunities arise. When every data movement incurs substantial fees, companies become hesitant to experiment — directly undermining their innovation potential.
Data Mobility Creates Competitive Advantage
Today’s AI startups need two critical resources to survive. Many focus exclusively on securing GPU access — the processing power needed to train and run AI models. But what we’re learning from our AI customers is that there’s a second, equally vital element: Data mobility.
AI companies need their data to flow unimpeded to wherever the right computing resources are available at the right time. This mobility creates inherent agility. When a new, more efficient GPU model becomes available, companies with data portability can immediately take advantage without complicated migrations or prohibitive fees. When a specialized AI cloud offers better performance for specific workloads, it can seamlessly shift operations. This responsiveness becomes a strategic asset.
The Decades-Long AI Tailwind
We’re still in the earliest stages of the AI revolution. Drawing parallels to my experience during the early days of cloud computing, I see AI as a decades-long tailwind that will transform every industry.
Consider how data needs will evolve: When kids today use Snapchat, they’re exchanging photos and text. But soon, they’ll be typing in a few prompts to auto-generate videos to share with friends. A video is dramatically more data than an image, which is dramatically more than a text message. This level of data will continue to explode, creating tremendous amounts of information that need to be stored somewhere.
This exponential growth is not only in data volume, but also in data complexity, making today’s infrastructure decisions even more consequential. The companies that prepare now for open, portable data infrastructure will be positioned to capitalize on AI’s evolution without being constrained by yesterday’s architecture choices.
Preparing for the AI Future
The companies leading the next wave of AI innovation won’t necessarily be those with the largest data centers or the most GPUs. They’ll be the ones who’ve built flexible, open infrastructure that allows them to adapt quickly as the landscape evolves. Ask yourself: Is your organization paying an innovation tax through restrictive data policies? In the AI era, escaping legacy storage lock-in isn’t just a cost-saving measure — it’s the foundation of sustainable competitive advantage. The businesses that embrace open cloud principles today will be the ones defining AI’s future tomorrow.
